Authors: Warat Kongkitkul, Sompote Youwai, Siwipa Khamsoy, Manaswee Feungfung
Abstract: This study employs explainable artificial intelligence (XAI) techniques to analyze the behavior of asphalt concrete with varying aggregate gradations, focusing on resilience modulus (MR) and dynamic stability (DS) as measured by wheel track tests. The research utilizes a deep learning model with a multi-layer perceptron architecture to predict MR and DS based on aggregate gradation parameters derived from Bailey's Method, including coarse aggregate ratio (CA), fine aggregate coarse ratio (FAc), and other mix design variables. The model's performance was validated using k-fold cross-validation, demonstrating superior accuracy compared to alternative machine learning approaches. SHAP (SHapley Additive exPlanations) values were applied to interpret the model's predictions, providing insights into the relative importance and impact of different gradation characteristics on asphalt concrete performance. Key findings include the identification of critical aggregate size thresholds, particularly the 0.6 mm sieve size, which significantly influences both MR and DS. The study revealed size-dependent performance of aggregates, with coarse aggregates primarily affecting rutting resistance and medium-fine aggregates influencing stiffness. The research also highlighted the importance of aggregate lithology in determining rutting resistance. To facilitate practical application, web-based interfaces were developed for predicting MR and DS, incorporating explainable features to enhance transparency and interpretation of results. This research contributes a data-driven approach to understanding the complex relationships between aggregate gradation and asphalt concrete performance, potentially informing more efficient and performance-oriented mix design processes in the future.
Authors: Xin Wang, Yifan Zhang, Xiaojing Zhang, Longhui Yu, Xinna Lin, Jindong Jiang, Bin Ma, Kaicheng Yu
Abstract: Pharmaceutical patents play a vital role in biochemical industries, especially in drug discovery, providing researchers with unique early access to data, experimental results, and research insights. With the advancement of machine learning, patent analysis has evolved from manual labor to tasks assisted by automatic tools. However, there still lacks an unified agent that assists every aspect of patent analysis, from patent reading to core chemical identification. Leveraging the capabilities of Large Language Models (LLMs) to understand requests and follow instructions, we introduce the $\textbf{first}$ intelligent agent in this domain, $\texttt{PatentAgent}$, poised to advance and potentially revolutionize the landscape of pharmaceutical research. $\texttt{PatentAgent}$ comprises three key end-to-end modules -- $\textit{PA-QA}$, $\textit{PA-Img2Mol}$, and $\textit{PA-CoreId}$ -- that respectively perform (1) patent question-answering, (2) image-to-molecular-structure conversion, and (3) core chemical structure identification, addressing the essential needs of scientists and practitioners in pharmaceutical patent analysis. Each module of $\texttt{PatentAgent}$ demonstrates significant effectiveness with the updated algorithm and the synergistic design of $\texttt{PatentAgent}$ framework. $\textit{PA-Img2Mol}$ outperforms existing methods across CLEF, JPO, UOB, and USPTO patent benchmarks with an accuracy gain between 2.46% and 8.37% while $\textit{PA-CoreId}$ realizes accuracy improvement ranging from 7.15% to 7.62% on PatentNetML benchmark. Our code and dataset will be publicly available.
Authors: Avinash Maurya, Jie Ye, M. Mustafa Rafique, Franck Cappello, Bogdan Nicolae
Abstract: Transformers and large language models~(LLMs) have seen rapid adoption in all domains. Their sizes have exploded to hundreds of billions of parameters and keep increasing. Under these circumstances, the training of transformers is very expensive and often hits a ``memory wall'', i.e., even when using 3D parallelism (pipeline, tensor, data) and aggregating the memory of many GPUs, it is still not enough to hold the necessary data structures (model parameters, optimizer state, gradients, activations) in GPU memory. To compensate, state-of-the-art approaches offload the optimizer state, at least partially, to the host memory and perform hybrid CPU-GPU computations. However, the management of the combined host-GPU memory is often suboptimal and results in poor overlapping between data movements and computations. This leads to missed opportunities to simultaneously leverage the interconnect bandwidth and computational capabilities of CPUs and GPUs. In this paper, we leverage a key observation that the interleaving of the forward, backward and update phases generate fluctuations in the GPU memory utilization, which can be exploited to dynamically move a part of the optimizer state between the host and the GPU memory at each iteration. To this end, we design and implement \proj, a novel technique to split the LLM into subgroups, whose update phase is scheduled on either the CPU or the GPU based on our proposed performance model that addresses the trade-off between data movement cost, acceleration on the GPUs vs the CPUs, and competition for shared resources. We integrate our approach with DeepSpeed and demonstrate 2.5$\times$ faster iterations over state-of-the-art approaches using extensive experiments.
Authors: Farnoush Baghestani, Mahdi Pirayesh Shirazi Nejad, Youngsun Kong, Ki H. Chon
Abstract: Continuous monitoring of non-invasive skin sympathetic nerve activity (SKNA) holds promise for understanding the sympathetic nervous system (SNS) dynamics in various physiological and pathological conditions. However, muscle noise artifacts present a challenge in accurate SKNA analysis, particularly in real-life scenarios. This study proposes a deep convolutional neural network (CNN) approach to detect and remove muscle noise from SKNA recordings obtained via ECG electrodes. Twelve healthy participants underwent controlled experimental protocols involving cognitive stress induction and voluntary muscle movements, while collecting SKNA data. Power spectral analysis revealed significant muscle noise interference within the SKNA frequency band (500-1000 Hz). A 2D CNN model was trained on the spectrograms of the data segments to classify them into baseline, stress-induced SKNA, and muscle noise-contaminated periods, achieving an average accuracy of 89.85% across all subjects. Our findings underscore the importance of addressing muscle noise for accurate SKNA monitoring, advancing towards wearable SKNA sensors for real-world applications.
Authors: Ruyi Zhang, Hongzuo Xu, Songlei Jian, Yusong Tan, Haifang Zhou, Rulin Xu
Abstract: Training in unsupervised time series anomaly detection is constantly plagued by the discrimination between harmful `anomaly contaminations' and beneficial `hard normal samples'. These two samples exhibit analogous loss behavior that conventional loss-based methodologies struggle to differentiate. To tackle this problem, we propose a novel approach that supplements traditional loss behavior with `parameter behavior', enabling a more granular characterization of anomalous patterns. Parameter behavior is formalized by measuring the parametric response to minute perturbations in input samples. Leveraging the complementary nature of parameter and loss behaviors, we further propose a dual Parameter-Loss Data Augmentation method (termed PLDA), implemented within the reinforcement learning paradigm. During the training phase of anomaly detection, PLDA dynamically augments the training data through an iterative process that simultaneously mitigates anomaly contaminations while amplifying informative hard normal samples. PLDA demonstrates remarkable versatility, which can serve as an additional component that seamlessly integrated with existing anomaly detectors to enhance their detection performance. Extensive experiments on ten datasets show that PLDA significantly improves the performance of four distinct detectors by up to 8\%, outperforming three state-of-the-art data augmentation methods.
Authors: Haoxin Liu
Abstract: Link prediction is a fundamental task in graph analysis. Despite the success of various graph-based machine learning models for link prediction, there lacks a general understanding of different models. In this paper, we propose a unified framework for link prediction that covers matrix factorization and representative network embedding and graph neural network methods. Our preliminary methodological and empirical analyses further reveal several key design factors based on our unified framework. We believe our results could deepen our understanding and inspire novel designs for link prediction methods.
Authors: Shovito Barua Soumma, Kartik Mangipudi, Daniel Peterson, Shyamal Mehta, Hassan Ghasemzadeh
Abstract: Parkinson's disease (PD) is a progressive neurological disorder that impacts the quality of life significantly, making in-home monitoring of motor symptoms such as Freezing of Gait (FoG) critical. However, existing symptom monitoring technologies are power-hungry, rely on extensive amounts of labeled data, and operate in controlled settings. These shortcomings limit real-world deployment of the technology. This work presents LIFT-PD, a computationally-efficient self-supervised learning framework for real-time FoG detection. Our method combines self-supervised pre-training on unlabeled data with a novel differential hopping windowing technique to learn from limited labeled instances. An opportunistic model activation module further minimizes power consumption by selectively activating the deep learning module only during active periods. Extensive experimental results show that LIFT-PD achieves a 7.25% increase in precision and 4.4% improvement in accuracy compared to supervised models while using as low as 40% of the labeled training data used for supervised learning. Additionally, the model activation module reduces inference time by up to 67% compared to continuous inference. LIFT-PD paves the way for practical, energy-efficient, and unobtrusive in-home monitoring of PD patients with minimal labeling requirements.
Authors: Wentao Gao, Feiyu Yang, Mengze Hong, Xiaojing Du, Zechen Hu, Xiongren Chen, Ziqi Xu
Abstract: Time series forecasting is a critical task in various domains, where accurate predictions can drive informed decision-making. Traditional forecasting methods often rely on current observations of variables to predict future outcomes, typically overlooking the influence of latent confounders, unobserved variables that simultaneously affect both the predictors and the target outcomes. This oversight can introduce bias and degrade the performance of predictive models. In this study, we address this challenge by proposing an enhanced forecasting approach that incorporates representations of latent confounders derived from historical data. By integrating these confounders into the predictive process, our method aims to improve the accuracy and robustness of time series forecasts. The proposed approach is demonstrated through its application to climate science data, showing significant improvements over traditional methods that do not account for confounders.
Authors: Qi Zhang, Yifei Wang, Jingyi Cui, Xiang Pan, Qi Lei, Stefanie Jegelka, Yisen Wang
Abstract: Deep learning models often suffer from a lack of interpretability due to polysemanticity, where individual neurons are activated by multiple unrelated semantics, resulting in unclear attributions of model behavior. Recent advances in monosemanticity, where neurons correspond to consistent and distinct semantics, have significantly improved interpretability but are commonly believed to compromise accuracy. In this work, we challenge the prevailing belief of the accuracy-interpretability tradeoff, showing that monosemantic features not only enhance interpretability but also bring concrete gains in model performance. Across multiple robust learning scenarios-including input and label noise, few-shot learning, and out-of-domain generalization-our results show that models leveraging monosemantic features significantly outperform those relying on polysemantic features. Furthermore, we provide empirical and theoretical understandings on the robustness gains of feature monosemanticity. Our preliminary analysis suggests that monosemanticity, by promoting better separation of feature representations, leads to more robust decision boundaries. This diverse evidence highlights the generality of monosemanticity in improving model robustness. As a first step in this new direction, we embark on exploring the learning benefits of monosemanticity beyond interpretability, supporting the long-standing hypothesis of linking interpretability and robustness. Code is available at \url{https://github.com/PKU-ML/Beyond_Interpretability}.
Authors: Shuchen Wu, Mirko Thalmann, Peter Dayan, Zeynep Akata, Eric Schulz
Abstract: Humans excel at learning abstract patterns across different sequences, filtering out irrelevant details, and transferring these generalized concepts to new sequences. In contrast, many sequence learning models lack the ability to abstract, which leads to memory inefficiency and poor transfer. We introduce a non-parametric hierarchical variable learning model (HVM) that learns chunks from sequences and abstracts contextually similar chunks as variables. HVM efficiently organizes memory while uncovering abstractions, leading to compact sequence representations. When learning on language datasets such as babyLM, HVM learns a more efficient dictionary than standard compression algorithms such as Lempel-Ziv. In a sequence recall task requiring the acquisition and transfer of variables embedded in sequences, we demonstrate HVM's sequence likelihood correlates with human recall times. In contrast, large language models (LLMs) struggle to transfer abstract variables as effectively as humans. From HVM's adjustable layer of abstraction, we demonstrate that the model realizes a precise trade-off between compression and generalization. Our work offers a cognitive model that captures the learning and transfer of abstract representations in human cognition and differentiates itself from the behavior of large language models.
Authors: Ryan Liu, Jiayi Geng, Addison J. Wu, Ilia Sucholutsky, Tania Lombrozo, Thomas L. Griffiths
Abstract: Chain-of-thought (CoT) prompting has become a widely used strategy for working with large language and multimodal models. While CoT has been shown to improve performance across many tasks, determining the settings in which it is effective remains an ongoing effort. In particular, it is still an open question in what settings CoT systematically reduces model performance. In this paper, we seek to identify the characteristics of tasks where CoT reduces performance by drawing inspiration from cognitive psychology, looking at cases where (i) verbal thinking or deliberation hurts performance in humans, and (ii) the constraints governing human performance generalize to language models. Three such cases are implicit statistical learning, visual recognition, and classifying with patterns containing exceptions. In extensive experiments across all three settings, we find that a diverse collection of state-of-the-art models exhibit significant drop-offs in performance (e.g., up to 36.3% absolute accuracy for OpenAI o1-preview compared to GPT-4o) when using inference-time reasoning compared to zero-shot counterparts. We also identify three tasks that satisfy condition (i) but not (ii), and find that while verbal thinking reduces human performance in these tasks, CoT retains or increases model performance. Overall, our results show that while there is not an exact parallel between the cognitive processes of models and those of humans, considering cases where thinking has negative consequences for human performance can help us identify settings where it negatively impacts models. By connecting the literature on human deliberation with evaluations of CoT, we offer a new tool that can be used in understanding the impact of prompt choices and inference-time reasoning.
Authors: Po-Yu Liang, Jun Bai
Abstract: Biologists frequently desire protein inhibitors for a variety of reasons, including use as research tools for understanding biological processes and application to societal problems in agriculture, healthcare, etc. Immunotherapy, for instance, relies on immune checkpoint inhibitors to block checkpoint proteins, preventing their binding with partner proteins and boosting immune cell function against abnormal cells. Inhibitor discovery has long been a tedious process, which in recent years has been accelerated by computational approaches. Advances in artificial intelligence now provide an opportunity to make inhibitor discovery smarter than ever before. While extensive research has been conducted on computer-aided inhibitor discovery, it has mainly focused on either sequence-to-structure mapping, reverse mapping, or bio-activity prediction, making it unrealistic for biologists to utilize such tools. Instead, our work proposes a new method of computer-assisted inhibitor discovery: de novo pocket-aware peptide structure and sequence generation network. Our approach consists of two sequential diffusion models for end-to-end structure generation and sequence prediction. By leveraging angle and dihedral relationships between backbone atoms, we ensure an E(3)-invariant representation of peptide structures. Our results demonstrate that our method achieves comparable performance to state-of-the-art models, highlighting its potential in pocket-aware peptide design. This work offers a new approach for precise drug discovery using receptor-specific peptide generation.
Authors: Yuzhe Yang, Yipeng Du, Ahmad Farhan, Claudio Angione, Yue Zhao, Harry Yang, Fielding Johnston, James Buban, Patrick Colangelo
Abstract: The deployment of large-scale models, such as large language models (LLMs) and sophisticated image generation systems, incurs substantial costs due to their computational demands. To mitigate these costs and address challenges related to scalability and data security, there is a growing shift towards decentralized systems for deploying such models. In these decentralized environments, efficient inference acceleration becomes crucial to manage computational resources effectively and enhance system responsiveness. In this work, we address the challenge of selecting optimal acceleration methods in decentralized systems by introducing a meta-learning-based framework. This framework automates the selection process by learning from historical performance data of various acceleration techniques across different tasks. Unlike traditional methods that rely on random selection or expert intuition, our approach systematically identifies the best acceleration strategies based on the specific characteristics of each task. We demonstrate that our meta-learning framework not only streamlines the decision-making process but also consistently outperforms conventional methods in terms of efficiency and performance. Our results highlight the potential of meta-learning to revolutionize inference acceleration in decentralized AI systems, offering a path towards more democratic and economically feasible artificial intelligence solutions.
Authors: Heewoong Noh, Namkyeong Lee, Gyoung S. Na, Chanyoung Park
Abstract: While inorganic retrosynthesis planning is essential in the field of chemical science, the application of machine learning in this area has been notably less explored compared to organic retrosynthesis planning. In this paper, we propose Retrieval-Retro for inorganic retrosynthesis planning, which implicitly extracts the precursor information of reference materials that are retrieved from the knowledge base regarding domain expertise in the field. Specifically, instead of directly employing the precursor information of reference materials, we propose implicitly extracting it with various attention layers, which enables the model to learn novel synthesis recipes more effectively. Moreover, during retrieval, we consider the thermodynamic relationship between target material and precursors, which is essential domain expertise in identifying the most probable precursor set among various options. Extensive experiments demonstrate the superiority of Retrieval-Retro in retrosynthesis planning, especially in discovering novel synthesis recipes, which is crucial for materials discovery. The source code for Retrieval-Retro is available at https://github.com/HeewoongNoh/Retrieval-Retro.
Authors: Firas Bayram, Bestoun S. Ahmed
Abstract: Artificial intelligence (AI), and especially its sub-field of Machine Learning (ML), are impacting the daily lives of everyone with their ubiquitous applications. In recent years, AI researchers and practitioners have introduced principles and guidelines to build systems that make reliable and trustworthy decisions. From a practical perspective, conventional ML systems process historical data to extract the features that are consequently used to train ML models that perform the desired task. However, in practice, a fundamental challenge arises when the system needs to be operationalized and deployed to evolve and operate in real-life environments continuously. To address this challenge, Machine Learning Operations (MLOps) have emerged as a potential recipe for standardizing ML solutions in deployment. Although MLOps demonstrated great success in streamlining ML processes, thoroughly defining the specifications of robust MLOps approaches remains of great interest to researchers and practitioners. In this paper, we provide a comprehensive overview of the trustworthiness property of MLOps systems. Specifically, we highlight technical practices to achieve robust MLOps systems. In addition, we survey the existing research approaches that address the robustness aspects of ML systems in production. We also review the tools and software available to build MLOps systems and summarize their support to handle the robustness aspects. Finally, we present the open challenges and propose possible future directions and opportunities within this emerging field. The aim of this paper is to provide researchers and practitioners working on practical AI applications with a comprehensive view to adopt robust ML solutions in production environments.
Authors: Zeyuan Li, Yangfan He, Lewei He, Jianhui Wang, Tianyu Shi, Bin Lei, Yuchen Li, Qiuwu Chen
Abstract: Recently, large language models (LLMs) have achieved significant progress in automated code generation. Despite their strong instruction-following capabilities, these models frequently struggled to align with user intent in coding scenarios. In particular, they were hampered by datasets that lacked diversity and failed to address specialized tasks or edge cases. Furthermore, challenges in supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF) led to failures in generating precise, human-intent-aligned code. To tackle these challenges and improve the code generation performance for automated programming systems, we propose Feedback-driven Adaptive Long/short-term memory reinforced Coding Optimization (i.e., FALCON). FALCON is structured into two hierarchical levels. From the global level, long-term memory improves code quality by retaining and applying learned knowledge. At the local level, short-term memory allows for the incorporation of immediate feedback from compilers and AI systems. Additionally, we introduce meta-reinforcement learning with feedback rewards to solve the global-local bi-level optimization problem and enhance the model's adaptability across diverse code generation tasks. Extensive experiments demonstrate that our technique achieves state-of-the-art performance, leading other reinforcement learning methods by more than 4.5 percentage points on the MBPP benchmark and 6.1 percentage points on the Humaneval benchmark. The open-sourced code is publicly available at https://github.com/titurte/FALCON.
Authors: Yanliang Jin, Yifan Wu, Yuan Gao, Shunqing Zhang, Shugong Xu, Cheng-Xiang Wang
Abstract: The emergence of 6th generation (6G) mobile networks brings new challenges in supporting high-mobility communications, particularly in addressing the issue of channel aging. While existing channel prediction methods offer improved accuracy at the expense of increased computational complexity, limiting their practical application in mobile networks. To address these challenges, we present LinFormer, an innovative channel prediction framework based on a scalable, all-linear, encoder-only Transformer model. Our approach, inspired by natural language processing (NLP) models such as BERT, adapts an encoder-only architecture specifically for channel prediction tasks. We propose replacing the computationally intensive attention mechanism commonly used in Transformers with a time-aware multi-layer perceptron (TMLP), significantly reducing computational demands. The inherent time awareness of TMLP module makes it particularly suitable for channel prediction tasks. We enhance LinFormer's training process by employing a weighted mean squared error loss (WMSELoss) function and data augmentation techniques, leveraging larger, readily available communication datasets. Our approach achieves a substantial reduction in computational complexity while maintaining high prediction accuracy, making it more suitable for deployment in cost-effective base stations (BS). Comprehensive experiments using both simulated and measured data demonstrate that LinFormer outperforms existing methods across various mobility scenarios, offering a promising solution for future wireless communication systems.
Authors: Amr Gomaa, Bilal Mahdy
Abstract: Integration of human feedback plays a key role in improving the learning capabilities of intelligent systems. This comparative study delves into the performance, robustness, and limitations of imitation learning compared to traditional reinforcement learning methods within these systems. Recognizing the value of human-in-the-loop feedback, we investigate the influence of expert guidance and suboptimal demonstrations on the learning process. Through extensive experimentation and evaluations conducted in a pre-existing simulation environment using the Unity platform, we meticulously analyze the effectiveness and limitations of these learning approaches. The insights gained from this study contribute to the advancement of human-centered artificial intelligence by highlighting the benefits and challenges associated with the incorporation of human feedback into the learning process. Ultimately, this research promotes the development of models that can effectively address complex real-world problems.
Authors: Arpan Dasgupta, Gagan Jain, Arun Suggala, Karthikeyan Shanmugam, Milind Tambe, Aparna Taneja
Abstract: Mobile health (mHealth) programs face a critical challenge in optimizing the timing of automated health information calls to beneficiaries. This challenge has been formulated as a collaborative multi-armed bandit problem, requiring online learning of a low-rank reward matrix. Existing solutions often rely on heuristic combinations of offline matrix completion and exploration strategies. In this work, we propose a principled Bayesian approach using Thompson Sampling for this collaborative bandit problem. Our method leverages prior information through efficient Gibbs sampling for posterior inference over the low-rank matrix factors, enabling faster convergence. We demonstrate significant improvements over state-of-the-art baselines on a real-world dataset from the world's largest maternal mHealth program. Our approach achieves a $16\%$ reduction in the number of calls compared to existing methods and a $47$\% reduction compared to the deployed random policy. This efficiency gain translates to a potential increase in program capacity by $0.5-1.4$ million beneficiaries, granting them access to vital ante-natal and post-natal care information. Furthermore, we observe a $7\%$ and $29\%$ improvement in beneficiary retention (an extremely hard metric to impact) compared to state-of-the-art and deployed baselines, respectively. Synthetic simulations further demonstrate the superiority of our approach, particularly in low-data regimes and in effectively utilizing prior information. We also provide a theoretical analysis of our algorithm in a special setting using Eluder dimension.
Authors: Ilias Diakonikolas, Sushrut Karmalkar, Shuo Pang, Aaron Potechin
Abstract: Non-Gaussian Component Analysis (NGCA) is the statistical task of finding a non-Gaussian direction in a high-dimensional dataset. Specifically, given i.i.d.\ samples from a distribution $P^A_{v}$ on $\mathbb{R}^n$ that behaves like a known distribution $A$ in a hidden direction $v$ and like a standard Gaussian in the orthogonal complement, the goal is to approximate the hidden direction. The standard formulation posits that the first $k-1$ moments of $A$ match those of the standard Gaussian and the $k$-th moment differs. Under mild assumptions, this problem has sample complexity $O(n)$. On the other hand, all known efficient algorithms require $\Omega(n^{k/2})$ samples. Prior work developed sharp Statistical Query and low-degree testing lower bounds suggesting an information-computation tradeoff for this problem. Here we study the complexity of NGCA in the Sum-of-Squares (SoS) framework. Our main contribution is the first super-constant degree SoS lower bound for NGCA. Specifically, we show that if the non-Gaussian distribution $A$ matches the first $(k-1)$ moments of $\mathcal{N}(0, 1)$ and satisfies other mild conditions, then with fewer than $n^{(1 - \varepsilon)k/2}$ many samples from the normal distribution, with high probability, degree $(\log n)^{{1\over 2}-o_n(1)}$ SoS fails to refute the existence of such a direction $v$. Our result significantly strengthens prior work by establishing a super-polynomial information-computation tradeoff against a broader family of algorithms. As corollaries, we obtain SoS lower bounds for several problems in robust statistics and the learning of mixture models. Our SoS lower bound proof introduces a novel technique, that we believe may be of broader interest, and a number of refinements over existing methods.
Authors: Remi Genet, Hugo Inzirillo
Abstract: Recent research has challenged the necessity of complex deep learning architectures for time series forecasting, demonstrating that simple linear models can often outperform sophisticated approaches. Building upon this insight, we introduce a novel architecture the Temporal Linear Net (TLN), that extends the capabilities of linear models while maintaining interpretability and computational efficiency. TLN is designed to effectively capture both temporal and feature-wise dependencies in multivariate time series data. Our approach is a variant of TSMixer that maintains strict linearity throughout its architecture. TSMixer removes activation functions, introduces specialized kernel initializations, and incorporates dilated convolutions to handle various time scales, while preserving the linear nature of the model. Unlike transformer-based models that may lose temporal information due to their permutation-invariant nature, TLN explicitly preserves and leverages the temporal structure of the input data. A key innovation of TLN is its ability to compute an equivalent linear model, offering a level of interpretability not found in more complex architectures such as TSMixer. This feature allows for seamless conversion between the full TLN model and its linear equivalent, facilitating both training flexibility and inference optimization.
Authors: Wassim Bouaziz, El-Mahdi El-Mhamdi, Nicolas Usunier
Abstract: Gradient attacks and data poisoning tamper with the training of machine learning algorithms to maliciously alter them and have been proven to be equivalent in convex settings. The extent of harm these attacks can produce in non-convex settings is still to be determined. Gradient attacks can affect far less systems than data poisoning but have been argued to be more harmful since they can be arbitrary, whereas data poisoning reduces the attacker's power to only being able to inject data points to training sets, via e.g. legitimate participation in a collaborative dataset. This raises the question of whether the harm made by gradient attacks can be matched by data poisoning in non-convex settings. In this work, we provide a positive answer in a worst-case scenario and show how data poisoning can mimic a gradient attack to perform an availability attack on (non-convex) neural networks. Through gradient inversion, commonly used to reconstruct data points from actual gradients, we show how reconstructing data points out of malicious gradients can be sufficient to perform a range of attacks. This allows us to show, for the first time, an availability attack on neural networks through data poisoning, that degrades the model's performances to random-level through a minority (as low as 1%) of poisoned points.
Authors: Hanshi Sun, Li-Wen Chang, Wenlei Bao, Size Zheng, Ningxin Zheng, Xin Liu, Harry Dong, Yuejie Chi, Beidi Chen
Abstract: With the widespread deployment of long-context large language models (LLMs), there has been a growing demand for efficient support of high-throughput inference. However, as the key-value (KV) cache expands with the sequence length, the increasing memory footprint and the need to access it for each token generation both result in low throughput when serving long-context LLMs. While various dynamic sparse attention methods have been proposed to speed up inference while maintaining generation quality, they either fail to sufficiently reduce GPU memory consumption or introduce significant decoding latency by offloading the KV cache to the CPU. We present ShadowKV, a high-throughput long-context LLM inference system that stores the low-rank key cache and offloads the value cache to reduce the memory footprint for larger batch sizes and longer sequences. To minimize decoding latency, ShadowKV employs an accurate KV selection strategy that reconstructs minimal sparse KV pairs on-the-fly. By evaluating ShadowKV on a broad range of benchmarks, including RULER, LongBench, and Needle In A Haystack, and models like Llama-3.1-8B, Llama-3-8B-1M, GLM-4-9B-1M, Yi-9B-200K, Phi-3-Mini-128K, and Qwen2-7B-128K, we demonstrate that it can support up to 6$\times$ larger batch sizes and boost throughput by up to 3.04$\times$ on an A100 GPU without sacrificing accuracy, even surpassing the performance achievable with infinite batch size under the assumption of infinite GPU memory. The code is available at https://github.com/bytedance/ShadowKV.
Authors: Brendan Hogan, Anmol Kabra, Felipe Siqueira Pacheco, Laura Greenstreet, Joshua Fan, Aaron Ferber, Marta Ummus, Alecsander Brito, Olivia Graham, Lillian Aoki, Drew Harvell, Alex Flecker, Carla Gomes
Abstract: Trust and interpretability are crucial for the use of Artificial Intelligence (AI) in scientific research, but current models often operate as black boxes offering limited transparency and justifications for their outputs. We introduce AiSciVision, a framework that specializes Large Multimodal Models (LMMs) into interactive research partners and classification models for image classification tasks in niche scientific domains. Our framework uses two key components: (1) Visual Retrieval-Augmented Generation (VisRAG) and (2) domain-specific tools utilized in an agentic workflow. To classify a target image, AiSciVision first retrieves the most similar positive and negative labeled images as context for the LMM. Then the LMM agent actively selects and applies tools to manipulate and inspect the target image over multiple rounds, refining its analysis before making a final prediction. These VisRAG and tooling components are designed to mirror the processes of domain experts, as humans often compare new data to similar examples and use specialized tools to manipulate and inspect images before arriving at a conclusion. Each inference produces both a prediction and a natural language transcript detailing the reasoning and tool usage that led to the prediction. We evaluate AiSciVision on three real-world scientific image classification datasets: detecting the presence of aquaculture ponds, diseased eelgrass, and solar panels. Across these datasets, our method outperforms fully supervised models in low and full-labeled data settings. AiSciVision is actively deployed in real-world use, specifically for aquaculture research, through a dedicated web application that displays and allows the expert users to converse with the transcripts. This work represents a crucial step toward AI systems that are both interpretable and effective, advancing their use in scientific research and scientific discovery.
Authors: Ren\'e Manass\'e Galekwa, Jean Marie Tshimula, Etienne Gael Tajeuna, Kyamakya Kyandoghere
Abstract: The sports betting industry has experienced rapid growth, driven largely by technological advancements and the proliferation of online platforms. Machine learning (ML) has played a pivotal role in the transformation of this sector by enabling more accurate predictions, dynamic odds-setting, and enhanced risk management for both bookmakers and bettors. This systematic review explores various ML techniques, including support vector machines, random forests, and neural networks, as applied in different sports such as soccer, basketball, tennis, and cricket. These models utilize historical data, in-game statistics, and real-time information to optimize betting strategies and identify value bets, ultimately improving profitability. For bookmakers, ML facilitates dynamic odds adjustment and effective risk management, while bettors leverage data-driven insights to exploit market inefficiencies. This review also underscores the role of ML in fraud detection, where anomaly detection models are used to identify suspicious betting patterns. Despite these advancements, challenges such as data quality, real-time decision-making, and the inherent unpredictability of sports outcomes remain. Ethical concerns related to transparency and fairness are also of significant importance. Future research should focus on developing adaptive models that integrate multimodal data and manage risk in a manner akin to financial portfolios. This review provides a comprehensive examination of the current applications of ML in sports betting, and highlights both the potential and the limitations of these technologies.
Authors: Hongyang Li, Caesar Wu, Mohammed Chadli, Said Mammar, Pascal Bouvry
Abstract: Distributed learning (DL) leverages multiple nodes to accelerate training, enabling the efficient optimization of large-scale models. Stochastic Gradient Descent (SGD), a key optimization algorithm, plays a central role in this process. However, communication bottlenecks often limit scalability and efficiency, leading to the increasing adoption of compressed SGD techniques to alleviate these challenges. Despite addressing communication overheads, compressed SGD introduces trustworthiness concerns, as gradient exchanges among nodes are vulnerable to attacks like gradient inversion (GradInv) and membership inference attacks (MIA). The trustworthiness of compressed SGD remains underexplored, leaving important questions about its reliability unanswered. In this paper, we provide a trustworthiness evaluation of compressed versus uncompressed SGD. Specifically, we conduct empirical studies using GradInv attacks, revealing that compressed SGD demonstrates significantly higher resistance to privacy leakage compared to uncompressed SGD. Moreover, our findings suggest that MIA may not be a reliable metric for assessing privacy risks in machine learning.
Authors: Joe Benton, Misha Wagner, Eric Christiansen, Cem Anil, Ethan Perez, Jai Srivastav, Esin Durmus, Deep Ganguli, Shauna Kravec, Buck Shlegeris, Jared Kaplan, Holden Karnofsky, Evan Hubinger, Roger Grosse, Samuel R. Bowman, David Duvenaud
Abstract: Sufficiently capable models could subvert human oversight and decision-making in important contexts. For example, in the context of AI development, models could covertly sabotage efforts to evaluate their own dangerous capabilities, to monitor their behavior, or to make decisions about their deployment. We refer to this family of abilities as sabotage capabilities. We develop a set of related threat models and evaluations. These evaluations are designed to provide evidence that a given model, operating under a given set of mitigations, could not successfully sabotage a frontier model developer or other large organization's activities in any of these ways. We demonstrate these evaluations on Anthropic's Claude 3 Opus and Claude 3.5 Sonnet models. Our results suggest that for these models, minimal mitigations are currently sufficient to address sabotage risks, but that more realistic evaluations and stronger mitigations seem likely to be necessary soon as capabilities improve. We also survey related evaluations we tried and abandoned. Finally, we discuss the advantages of mitigation-aware capability evaluations, and of simulating large-scale deployments using small-scale statistics.
Authors: Wenxian Shi, Menghua Wu, Regina Barzilay
Abstract: Forecasting the change in the distribution of viral variants is crucial for therapeutic design and disease surveillance. This task poses significant modeling challenges due to the sharp differences in virus distributions across sub-populations (e.g., countries) and their dynamic interactions. Existing machine learning approaches that model the variant distribution as a whole are incapable of making location-specific predictions and ignore transmissions that shape the viral landscape. In this paper, we propose a sub-population specific protein evolution model, which predicts the time-resolved distributions of viral proteins in different locations. The algorithm explicitly models the transmission rates between sub-populations and learns their interdependence from data. The change in protein distributions across all sub-populations is defined through a linear ordinary differential equation (ODE) parametrized by transmission rates. Solving this ODE yields the likelihood of a given protein occurring in particular sub-populations. Multi-year evaluation on both SARS-CoV-2 and influenza A/H3N2 demonstrates that our model outperforms baselines in accurately predicting distributions of viral proteins across continents and countries. We also find that the transmission rates learned from data are consistent with the transmission pathways discovered by retrospective phylogenetic analysis.
Authors: Xinrui He, Yikun Ban, Jiaru Zou, Tianxin Wei, Curtiss B. Cook, Jingrui He
Abstract: Missing data imputation is a critical challenge in tabular datasets, especially in healthcare, where data completeness is vital for accurate analysis. Large language models (LLMs), trained on vast corpora, have shown strong potential in data generation, making them a promising tool for tabular data imputation. However, challenges persist in designing effective prompts for a finetuning-free process and in mitigating the risk of LLM hallucinations. To address these issues, we propose a novel framework, LLM-Forest, which introduces a "forest" of few-shot learning LLM "trees" with confidence-based weighted voting. This framework is established on a new concept of bipartite information graphs to identify high-quality relevant neighboring entries with both feature and value granularity. Extensive experiments on four real-world healthcare datasets demonstrate the effectiveness and efficiency of LLM-Forest.
Authors: Sriniketh Vangaru, Daniel Rosen, Dylan Green, Raphael Rodriguez, Maxwell Wiecek, Amos Johnson, Alyse M. Jones, William C. Headley
Abstract: Technological trends show that Radio Frequency Reinforcement Learning (RFRL) will play a prominent role in the wireless communication systems of the future. Applications of RFRL range from military communications jamming to enhancing WiFi networks. Before deploying algorithms for these purposes, they must be trained in a simulation environment to ensure adequate performance. For this reason, we previously created the RFRL Gym: a standardized, accessible tool for the development and testing of reinforcement learning (RL) algorithms in the wireless communications space. This environment leveraged the OpenAI Gym framework and featured customizable simulation scenarios within the RF spectrum. However, the RFRL Gym was limited to training a single RL agent per simulation; this is not ideal, as most real-world RF scenarios will contain multiple intelligent agents in cooperative, competitive, or mixed settings, which is a natural consequence of spectrum congestion. Therefore, through integration with Ray RLlib, multi-agent reinforcement learning (MARL) functionality for training and assessment has been added to the RFRL Gym, making it even more of a robust tool for RF spectrum simulation. This paper provides an overview of the updated RFRL Gym environment. In this work, the general framework of the tool is described relative to comparable existing resources, highlighting the significant additions and refactoring we have applied to the Gym. Afterward, results from testing various RF scenarios in the MARL environment and future additions are discussed.
Authors: Hengrui Zhang, Liancheng Fang, Qitian Wu, Philip S. Yu
Abstract: Autoregressive models are predominant in natural language generation, while their application in tabular data remains underexplored. We posit that this can be attributed to two factors: 1) tabular data contains heterogeneous data type, while the autoregressive model is primarily designed to model discrete-valued data; 2) tabular data is column permutation-invariant, requiring a generation model to generate columns in arbitrary order. This paper proposes a Diffusion-nested Autoregressive model (TabDAR) to address these issues. To enable autoregressive methods for continuous columns, TabDAR employs a diffusion model to parameterize the conditional distribution of continuous features. To ensure arbitrary generation order, TabDAR resorts to masked transformers with bi-directional attention, which simulate various permutations of column order, hence enabling it to learn the conditional distribution of a target column given an arbitrary combination of other columns. These designs enable TabDAR to not only freely handle heterogeneous tabular data but also support convenient and flexible unconditional/conditional sampling. We conduct extensive experiments on ten datasets with distinct properties, and the proposed TabDAR outperforms previous state-of-the-art methods by 18% to 45% on eight metrics across three distinct aspects.
Authors: Hsun-Yu Kuo, Yin-Hsiang Liao, Yu-Chieh Chao, Wei-Yun Ma, Pu-Jen Cheng
Abstract: Synthetic data augmentation via large language models (LLMs) allows researchers to leverage additional training data, thus enhancing the performance of downstream tasks, especially when real-world data is scarce. However, the generated data can deviate from the real-world data, and this misalignment can bring deficient outcomes while applying the trained model to applications. Therefore, we proposed efficient weighted-loss approaches to align synthetic data with real-world distribution by emphasizing high-quality and diversified data generated by LLMs with using merely a little real-world data. We empirically assessed the effectiveness of our method on multiple text classification tasks, and the results showed leveraging our approaches on a BERT-level model robustly outperformed standard cross-entropy and other data weighting approaches, providing potential solutions to effectively leveraging synthetic data from any suitable data generator for model training.
Authors: Guneet S. Dhillon, Xingjian Shi, Yee Whye Teh, Alex Smola
Abstract: Supervised fine-tuning (SFT) and alignment of large language models (LLMs) are key steps in providing a good user experience. However, the concept of an appropriate alignment is inherently application-dependent, and current methods often rely on heuristic choices to drive the optimization. In this work, we formulate SFT and alignment as a constrained optimization problem, where the LLM is trained on a task while being required to meet application-specific requirements, without resorting to heuristics. To solve this, we propose Lagrange Large Language Models (L3Ms), which employ logarithmic barriers to enforce the constraints. This approach allows for the customization of L3Ms across diverse applications while avoiding heuristic-driven processes. We demonstrate experimentally the versatility and efficacy of L3Ms in achieving tailored alignments for various applications.
Authors: Muhammad Farhan Tanvir, Md Maruf Hossain, Md Asifuzzaman Jishan
Abstract: In the highly competitive environment of the banking industry, it is essential to precisely forecast the behavior of customers in order to maximize the effectiveness of marketing initiatives and improve financial consequences. The purpose of this research is to examine the efficacy of logit and probit models in predicting term deposit subscriptions using a Portuguese bank's direct marketing data. There are several demographic, economic, and behavioral characteristics in the dataset that affect the probability of subscribing. To increase model performance and provide an unbiased evaluation, the target variable was balanced, considering the inherent imbalance in the dataset. The two model's prediction abilities were evaluated using Bayesian techniques and Leave-One-Out Cross-Validation (LOO-CV). The logit model performed better than the probit model in handling this classification problem. The results highlight the relevance of model selection when dealing with complicated decision-making processes in the financial services industry and imbalanced datasets. Findings from this study shed light on how banks can optimize their decision-making processes, improve their client segmentation, and boost their marketing campaigns by utilizing machine learning models.
Authors: Pouya M. Ghari, Yanning Shen
Abstract: Federated learning is renowned for its efficacy in distributed model training, ensuring that users, called clients, retain data privacy by not disclosing their data to the central server that orchestrates collaborations. Most previous work on federated learning assumes that clients possess static batches of training data. However, clients may also need to make real-time predictions on streaming data in non-stationary environments. In such dynamic environments, employing pre-trained models may be inefficient, as they struggle to adapt to the constantly evolving data streams. To address this challenge, clients can fine-tune models online, leveraging their observed data to enhance performance. Despite the potential benefits of client participation in federated online model fine-tuning, existing analyses have not conclusively demonstrated its superiority over local model fine-tuning. To bridge this gap, the present paper develops a novel personalized federated learning algorithm, wherein each client constructs a personalized model by combining a locally fine-tuned model with multiple federated models learned by the server over time. Theoretical analysis and experiments on real datasets corroborate the effectiveness of this approach for real-time predictions and federated model fine-tuning.
Authors: Shaorong Zhang, Yuanbin Cheng, Xianghao Kong, Greg Ver Steeg
Abstract: Diffusion bridge models effectively facilitate image-to-image (I2I) translation by connecting two distributions. However, existing methods overlook the impact of noise in sampling SDEs, transition kernel, and the base distribution on sampling efficiency, image quality and diversity. To address this gap, we propose the Stochasticity-controlled Diffusion Bridge (SDB), a novel theoretical framework that extends the design space of diffusion bridges, and provides strategies to mitigate singularities during both training and sampling. By controlling stochasticity in the sampling SDEs, our sampler achieves speeds up to 5 times faster than the baseline, while also producing lower FID scores. After training, SDB sets new benchmarks in image quality and sampling efficiency via managing stochasticity within the transition kernel. Furthermore, introducing stochasticity into the base distribution significantly improves image diversity, as quantified by a newly introduced metric.
Authors: Alexander Christie, Matan Leibovitch, Miguel Moscoso, Alexei Novikov, George Papanicolaou, Chrysoula Tsogka
Abstract: We propose a methodology that exploits large and diverse data sets to accurately estimate the ambient medium's Green's functions in strongly scattering media. Given these estimates, obtained with and without the use of neural networks, excellent imaging results are achieved, with a resolution that is better than that of a homogeneous medium. This phenomenon, also known as super-resolution, occurs because the ambient scattering medium effectively enhances the physical imaging aperture.
Authors: Juyoung Yun
Abstract: In deep learning, Residual Networks (ResNets) have proven effective in addressing the vanishing gradient problem, allowing for the successful training of very deep networks. However, skip connections in ResNets can lead to gradient overlap, where gradients from both the learned transformation and the skip connection combine, potentially resulting in overestimated gradients. This overestimation can cause inefficiencies in optimization, as some updates may overshoot optimal regions, affecting weight updates. To address this, we examine Z-score Normalization (ZNorm) as a technique to manage gradient overlap. ZNorm adjusts the gradient scale, standardizing gradients across layers and reducing the negative impact of overlapping gradients. Our experiments demonstrate that ZNorm improves training process, especially in non-convex optimization scenarios common in deep learning, where finding optimal solutions is challenging. These findings suggest that ZNorm can affect the gradient flow, enhancing performance in large-scale data processing where accuracy is critical.
Authors: Dimitris Bertsimas, Lisa Everest, Jiayi Gu, Matthew Peroni, Vasiliki Stoumpou
Abstract: Decision Trees have remained a popular machine learning method for tabular datasets, mainly due to their interpretability. However, they lack the expressiveness needed to handle highly nonlinear or unstructured datasets. Motivated by recent advances in tree-based machine learning (ML) techniques and first-order optimization methods, we introduce Generalized Soft Trees (GSTs), which extend soft decision trees (STs) and are capable of processing images directly. We demonstrate their advantages with respect to tractability, performance, and interpretability. We develop a tractable approach to growing GSTs, given by the DeepTree algorithm, which, in addition to new regularization terms, produces high-quality models with far fewer nodes and greater interpretability than traditional soft trees. We test the performance of our GSTs on benchmark tabular and image datasets, including MIMIC-IV, MNIST, Fashion MNIST, CIFAR-10 and Celeb-A. We show that our approach outperforms other popular tree methods (CART, Random Forests, XGBoost) in almost all of the datasets, with Convolutional Trees having a significant edge in the hardest CIFAR-10 and Fashion MNIST datasets. Finally, we explore the interpretability of our GSTs and find that even the most complex GSTs are considerably more interpretable than deep neural networks. Overall, our approach of Generalized Soft Trees provides a tractable method that is high-performing on (un)structured datasets and preserves interpretability more than traditional deep learning methods.
Authors: Tyler Sam, Yudong Chen, Christina Lee Yu
Abstract: Many reinforcement learning (RL) algorithms are too costly to use in practice due to the large sizes $S, A$ of the problem's state and action space. To resolve this issue, we study transfer RL with latent low rank structure. We consider the problem of transferring a latent low rank representation when the source and target MDPs have transition kernels with Tucker rank $(S , d, A )$, $(S , S , d), (d, S, A )$, or $(d , d , d )$. In each setting, we introduce the transfer-ability coefficient $\alpha$ that measures the difficulty of representational transfer. Our algorithm learns latent representations in each source MDP and then exploits the linear structure to remove the dependence on $S, A $, or $S A$ in the target MDP regret bound. We complement our positive results with information theoretic lower bounds that show our algorithms (excluding the ($d, d, d$) setting) are minimax-optimal with respect to $\alpha$.
Authors: Claudius Krause, Michele Faucci Giannelli, Gregor Kasieczka, Benjamin Nachman, Dalila Salamani, David Shih, Anna Zaborowska, Oz Amram, Kerstin Borras, Matthew R. Buckley, Erik Buhmann, Thorsten Buss, Renato Paulo Da Costa Cardoso, Anthony L. Caterini, Nadezda Chernyavskaya, Federico A. G. Corchia, Jesse C. Cresswell, Sascha Diefenbacher, Etienne Dreyer, Vijay Ekambaram, Engin Eren, Florian Ernst, Luigi Favaro, Matteo Franchini, Frank Gaede, Eilam Gross, Shih-Chieh Hsu, Kristina Jaruskova, Benno K\"ach, Jayant Kalagnanam, Raghav Kansal, Taewoo Kim, Dmitrii Kobylianskii, Anatolii Korol, William Korcari, Dirk Kr\"ucker, Katja Kr\"uger, Marco Letizia, Shu Li, Qibin Liu, Xiulong Liu, Gabriel Loaiza-Ganem, Thandikire Madula, Peter McKeown, Isabell-A. Melzer-Pellmann, Vinicius Mikuni, Nam Nguyen, Ayodele Ore, Sofia Palacios Schweitzer, Ian Pang, Kevin Pedro, Tilman Plehn, Witold Pokorski, Huilin Qu, Piyush Raikwar, John A. Raine, Humberto Reyes-Gonzalez, Lorenzo Rinaldi, Brendan Leigh Ross, Moritz A. W. Scham, Simon Schnake, Chase Shimmin, Eli Shlizerman, Nathalie Soybelman, Mudhakar Srivatsa, Kalliopi Tsolaki, Sofia Vallecorsa, Kyongmin Yeo, Rui Zhang
Abstract: We present the results of the "Fast Calorimeter Simulation Challenge 2022" - the CaloChallenge. We study state-of-the-art generative models on four calorimeter shower datasets of increasing dimensionality, ranging from a few hundred voxels to a few tens of thousand voxels. The 31 individual submissions span a wide range of current popular generative architectures, including Variational AutoEncoders (VAEs), Generative Adversarial Networks (GANs), Normalizing Flows, Diffusion models, and models based on Conditional Flow Matching. We compare all submissions in terms of quality of generated calorimeter showers, as well as shower generation time and model size. To assess the quality we use a broad range of different metrics including differences in 1-dimensional histograms of observables, KPD/FPD scores, AUCs of binary classifiers, and the log-posterior of a multiclass classifier. The results of the CaloChallenge provide the most complete and comprehensive survey of cutting-edge approaches to calorimeter fast simulation to date. In addition, our work provides a uniquely detailed perspective on the important problem of how to evaluate generative models. As such, the results presented here should be applicable for other domains that use generative AI and require fast and faithful generation of samples in a large phase space.
Authors: Yiwen Qiu, Yujia Zheng, Kun Zhang
Abstract: When solving long-horizon tasks, it is intriguing to decompose the high-level task into subtasks. Decomposing experiences into reusable subtasks can improve data efficiency, accelerate policy generalization, and in general provide promising solutions to multi-task reinforcement learning and imitation learning problems. However, the concept of subtasks is not sufficiently understood and modeled yet, and existing works often overlook the true structure of the data generation process: subtasks are the results of a $\textit{selection}$ mechanism on actions, rather than possible underlying confounders or intermediates. Specifically, we provide a theory to identify, and experiments to verify the existence of selection variables in such data. These selections serve as subgoals that indicate subtasks and guide policy. In light of this idea, we develop a sequential non-negative matrix factorization (seq- NMF) method to learn these subgoals and extract meaningful behavior patterns as subtasks. Our empirical results on a challenging Kitchen environment demonstrate that the learned subtasks effectively enhance the generalization to new tasks in multi-task imitation learning scenarios. The codes are provided at https://anonymous.4open.science/r/Identifying\_Selections\_for\_Unsupervised\_Subtask\_Discovery/README.md.
Authors: Yuntian He, Pranav Maneriker, Anutam Srinivasan, Aditya T. Vadlamani, Srinivasan Parthasarathy
Abstract: Conformal Prediction is a robust framework that ensures reliable coverage across machine learning tasks. Although recent studies have applied conformal prediction to graph neural networks, they have largely emphasized post-hoc prediction set generation. Improving conformal prediction during the training stage remains unaddressed. In this work, we tackle this challenge from a denoising perspective by introducing SparGCP, which incorporates graph sparsification and a conformal prediction-specific objective into GNN training. SparGCP employs a parameterized graph sparsification module to filter out task-irrelevant edges, thereby improving conformal prediction efficiency. Extensive experiments on real-world graph datasets demonstrate that SparGCP outperforms existing methods, reducing prediction set sizes by an average of 32\% and scaling seamlessly to large networks on commodity GPUs.
Authors: Jiahe Bai, Baojian Zhou, Deqing Yang, Yanghua Xiao
Abstract: Efficient computation of graph diffusion equations (GDEs), such as Personalized PageRank, Katz centrality, and the Heat kernel, is crucial for clustering, training neural networks, and many other graph-related problems. Standard iterative methods require accessing the whole graph per iteration, making them time-consuming for large-scale graphs. While existing local solvers approximate diffusion vectors through heuristic local updates, they often operate sequentially and are typically designed for specific diffusion types, limiting their applicability. Given that diffusion vectors are highly localizable, as measured by the participation ratio, this paper introduces a novel framework for approximately solving GDEs using a local diffusion process. This framework reveals the suboptimality of existing local solvers. Furthermore, our approach effectively localizes standard iterative solvers by designing simple and provably sublinear time algorithms. These new local solvers are highly parallelizable, making them well-suited for implementation on GPUs. We demonstrate the effectiveness of our framework in quickly obtaining approximate diffusion vectors, achieving up to a hundred-fold speed improvement, and its applicability to large-scale dynamic graphs. Our framework could also facilitate more efficient local message-passing mechanisms for GNNs.
Authors: Kazuki Uematsu, Kosuke Haruki, Taiji Suzuki, Mitsuhiro Kimura, Takahiro Takimoto, Hideyuki Nakagawa
Abstract: Out-of-distribution (OOD) detection is a critical issue for the stable and reliable operation of systems using a deep neural network (DNN). Although many OOD detection methods have been proposed, it remains unclear how the differences between in-distribution (ID) and OOD samples are generated by each processing step inside DNNs. We experimentally clarify this issue by investigating the layer dependence of feature representations from multiple perspectives. We find that intrinsic low dimensionalization of DNNs is essential for understanding how OOD samples become more distinct from ID samples as features propagate to deeper layers. Based on these observations, we provide a simple picture that consistently explains various properties of OOD samples. Specifically, low-dimensional weights eliminate most information from OOD samples, resulting in misclassifications due to excessive attention to dataset bias. In addition, we demonstrate the utility of dimensionality by proposing a dimensionality-aware OOD detection method based on alignment of features and weights, which consistently achieves high performance for various datasets with lower computational cost.
Authors: M. Reza Ebrahimi, Jun Chen, Ashish Khisti
Abstract: This paper investigates a novel lossy compression framework operating under logarithmic loss, designed to handle situations where the reconstruction distribution diverges from the source distribution. This framework is especially relevant for applications that require joint compression and retrieval, and in scenarios involving distributional shifts due to processing. We show that the proposed formulation extends the classical minimum entropy coupling framework by integrating a bottleneck, allowing for a controlled degree of stochasticity in the coupling. We explore the decomposition of the Minimum Entropy Coupling with Bottleneck (MEC-B) into two distinct optimization problems: Entropy-Bounded Information Maximization (EBIM) for the encoder, and Minimum Entropy Coupling (MEC) for the decoder. Through extensive analysis, we provide a greedy algorithm for EBIM with guaranteed performance, and characterize the optimal solution near functional mappings, yielding significant theoretical insights into the structural complexity of this problem. Furthermore, we illustrate the practical application of MEC-B through experiments in Markov Coding Games (MCGs) under rate limits. These games simulate a communication scenario within a Markov Decision Process, where an agent must transmit a compressed message from a sender to a receiver through its actions. Our experiments highlight the trade-offs between MDP rewards and receiver accuracy across various compression rates, showcasing the efficacy of our method compared to conventional compression baseline.
Authors: Rajeev Kohli, Kriste Krstovski, Hengyu Kuang, Hengxu Lin
Abstract: Experience goods such as sporting and artistic events, songs, videos, news stories, podcasts, and television series, are often packaged and consumed in bundles. Many such bundles are ordered in the sense that the individual items are consumed sequentially, one at a time. We examine if an individual's decision to consume the next item in an ordered bundle can be predicted based on his/her consumption pattern for the preceding items. We evaluate several predictive models, including two custom Transformers using decoder-only and encoder-decoder architectures, fine-tuned GPT-3, a custom LSTM model, a reinforcement learning model, two Markov models, and a zero-order model. Using data from Spotify, we find that the custom Transformer with a decoder-only architecture provides the most accurate predictions, both for individual choices and aggregate demand. This model captures a general form of state dependence. Analysis of Transformer attention weights suggests that the consumption of the next item in a bundle is based on approximately equal weighting of all preceding choices. Our results indicate that the Transformer can assist in queuing the next item that an individual is likely to consume from an ordered bundle, predicting the demand for individual items, and personalizing promotions to increase demand.
Authors: Hanlin Zhang, Depen Morwani, Nikhil Vyas, Jingfeng Wu, Difan Zou, Udaya Ghai, Dean Foster, Sham Kakade
Abstract: Training large-scale models under given resources requires careful design of parallelism strategies. In particular, the efficiency notion of critical batch size, concerning the compromise between time and compute, marks the threshold beyond which greater data parallelism leads to diminishing returns. To operationalize it, we propose a measure of CBS and pre-train a series of auto-regressive language models, ranging from 85 million to 1.2 billion parameters, on the C4 dataset. Through extensive hyper-parameter sweeps and careful control on factors such as batch size, momentum, and learning rate along with its scheduling, we systematically investigate the impact of scale on CBS. Then we fit scaling laws with respect to model and data sizes to decouple their effects. Overall, our results demonstrate that CBS scales primarily with data size rather than model size, a finding we justify theoretically through the analysis of infinite-width limits of neural networks and infinite-dimensional least squares regression. Of independent interest, we highlight the importance of common hyper-parameter choices and strategies for studying large-scale pre-training beyond fixed training durations.
Authors: Zihan Pengmei, Zhengyuan Shen, Zichen Wang, Marcus Collins, Huzefa Rangwala
Abstract: Constructing transferable descriptors for conformation representation of molecular and biological systems finds numerous applications in drug discovery, learning-based molecular dynamics, and protein mechanism analysis. Geometric graph neural networks (Geom-GNNs) with all-atom information have transformed atomistic simulations by serving as a general learnable geometric descriptors for downstream tasks including prediction of interatomic potential and molecular properties. However, common practices involve supervising Geom-GNNs on specific downstream tasks, which suffer from the lack of high-quality data and inaccurate labels leading to poor generalization and performance degradation on out-of-distribution (OOD) scenarios. In this work, we explored the possibility of using pre-trained Geom-GNNs as transferable and highly effective geometric descriptors for improved generalization. To explore their representation power, we studied the scaling behaviors of Geom-GNNs under self-supervised pre-training, supervised and unsupervised learning setups. We find that the expressive power of different architectures can differ on the pre-training task. Interestingly, Geom-GNNs do not follow the power-law scaling on the pre-training task, and universally lack predictable scaling behavior on the supervised tasks with quantum chemical labels important for screening and design of novel molecules. More importantly, we demonstrate how all-atom graph embedding can be organically combined with other neural architectures to enhance the expressive power. Meanwhile, the low-dimensional projection of the latent space shows excellent agreement with conventional geometrical descriptors.
Authors: Khashayar Gatmiry, Nikunj Saunshi, Sashank J. Reddi, Stefanie Jegelka, Sanjiv Kumar
Abstract: The intriguing in-context learning (ICL) abilities of deep Transformer models have lately garnered significant attention. By studying in-context linear regression on unimodal Gaussian data, recent empirical and theoretical works have argued that ICL emerges from Transformers' abilities to simulate learning algorithms like gradient descent. However, these works fail to capture the remarkable ability of Transformers to learn multiple tasks in context. To this end, we study in-context learning for linear regression with diverse tasks, characterized by data covariance matrices with condition numbers ranging from $[1, \kappa]$, and highlight the importance of depth in this setting. More specifically, (a) we show theoretical lower bounds of $\log(\kappa)$ (or $\sqrt{\kappa}$) linear attention layers in the unrestricted (or restricted) attention setting and, (b) we show that multilayer Transformers can indeed solve such tasks with a number of layers that matches the lower bounds. However, we show that this expressivity of multilayer Transformer comes at the price of robustness. In particular, multilayer Transformers are not robust to even distributional shifts as small as $O(e^{-L})$ in Wasserstein distance, where $L$ is the depth of the network. We then demonstrate that Looped Transformers -- a special class of multilayer Transformers with weight-sharing -- not only exhibit similar expressive power but are also provably robust under mild assumptions. Besides out-of-distribution generalization, we also show that Looped Transformers are the only models that exhibit a monotonic behavior of loss with respect to depth.
Authors: Shaan Ul Haque, Siva Theja Maguluri
Abstract: Motivated by engineering applications such as resource allocation in networks and inventory systems, we consider average-reward Reinforcement Learning with unbounded state space and reward function. Recent works studied this problem in the actor-critic framework and established finite sample bounds assuming access to a critic with certain error guarantees. We complement their work by studying Temporal Difference (TD) learning with linear function approximation and establishing finite-time bounds with the optimal $\mathcal{O}\left(1/\epsilon^2\right)$ sample complexity. These results are obtained using the following general-purpose theorem for non-linear Stochastic Approximation (SA). Suppose that one constructs a Lyapunov function for a non-linear SA with certain drift condition. Then, our theorem establishes finite-time bounds when this SA is driven by unbounded Markovian noise under suitable conditions. It serves as a black box tool to generalize sample guarantees on SA from i.i.d. or martingale difference case to potentially unbounded Markovian noise. The generality and the mild assumption of the setup enables broad applicability of our theorem. We illustrate its power by studying two more systems: (i) We improve upon the finite-time bounds of $Q$-learning by tightening the error bounds and also allowing for a larger class of behavior policies. (ii) We establish the first ever finite-time bounds for distributed stochastic optimization of high-dimensional smooth strongly convex function using cyclic block coordinate descent.
Authors: Xuetong Li, Xiao-Dong Zhang
Abstract: Multisource data has spurred the development of advanced clustering algorithms, such as multi-view clustering, which critically relies on constructing similarity matrices. Traditional algorithms typically generate these matrices from sample attributes alone. However, real-world networks often include pairwise directed topological structures critical for clustering. This paper introduces a novel multi-view clustering algorithm, AAS. It utilizes a two-step proximity approach via anchors in each view, integrating attribute and directed structural information. This approach enhances the clarity of category characteristics in the similarity matrices. The anchor structural similarity matrix leverages strongly connected components of directed graphs. The entire process-from similarity matrices construction to clustering - is consolidated into a unified optimization framework. Comparative experiments on the modified Attribute SBM dataset against eight algorithms affirm the effectiveness and superiority of AAS.
Authors: Julien Pallage, Bertrand Scherrer, Salma Naccache, Christophe B\'elanger, Antoine Lesage-Landry
Abstract: In this work, we present a new unsupervised anomaly (outlier) detection (AD) method using the sliced-Wasserstein metric. This filtering technique is conceptually interesting for integration in MLOps pipelines deploying trustworthy machine learning models in critical sectors like energy. Additionally, we open the first dataset showcasing localized critical peak rebate demand response in a northern climate. We demonstrate the capabilities of our method on synthetic datasets as well as standard AD datasets and use it in the making of a first benchmark for our open-source localized critical peak rebate dataset.
Authors: Dang Nguyen, Sunil Gupta, Kien Do, Thin Nguyen, Svetha Venkatesh
Abstract: While most generative models show achievements in image data generation, few are developed for tabular data generation. Recently, due to success of large language models (LLM) in diverse tasks, they have also been used for tabular data generation. However, these methods do not capture the correct correlation between the features and the target variable, hindering their applications in downstream predictive tasks. To address this problem, we propose a LLM-based method with three important improvements to correctly capture the ground-truth feature-class correlation in the real data. First, we propose a novel permutation strategy for the input data in the fine-tuning phase. Second, we propose a feature-conditional sampling approach to generate synthetic samples. Finally, we generate the labels by constructing prompts based on the generated samples to query our fine-tuned LLM. Our extensive experiments show that our method significantly outperforms 10 SOTA baselines on 20 datasets in downstream tasks. It also produces highly realistic synthetic samples in terms of quality and diversity. More importantly, classifiers trained with our synthetic data can even compete with classifiers trained with the original data on half of the benchmark datasets, which is a significant achievement in tabular data generation.
Authors: Yang Xiang, Li Fan, Tulika Saha, Yushan Pan, Haiyang Zhang, Chengtao Ji
Abstract: Graph clustering is an essential aspect of network analysis that involves grouping nodes into separate clusters. Recent developments in deep learning have resulted in advanced deep graph clustering techniques, which have proven effective in many applications. Nonetheless, these methods often encounter difficulties when dealing with the complexities of real-world graphs, particularly in the presence of noisy edges. Additionally, many denoising graph clustering strategies tend to suffer from lower performance compared to their non-denoised counterparts, training instability, and challenges in scaling to large datasets. To tackle these issues, we introduce a new framework called the Dual Adaptive Assignment Approach for Robust Graph-Based Clustering (RDSA). RDSA consists of three key components: (i) a node embedding module that effectively integrates the graph's topological features and node attributes; (ii) a structure-based soft assignment module that improves graph modularity by utilizing an affinity matrix for node assignments; and (iii) a node-based soft assignment module that identifies community landmarks and refines node assignments to enhance the model's robustness. We assess RDSA on various real-world datasets, demonstrating its superior performance relative to existing state-of-the-art methods. Our findings indicate that RDSA provides robust clustering across different graph types, excelling in clustering effectiveness and robustness, including adaptability to noise, stability, and scalability.
Authors: Bo Jiang, Hao Wu, Beibei Wang, Jin Tang, Bin Luo
Abstract: Recently, graph prompt learning has garnered increasing attention in adapting pre-trained GNN models for downstream graph learning tasks. However, existing works generally conduct prompting over all graph elements (e.g., nodes, edges, node attributes, etc.), which is suboptimal and obviously redundant. To address this issue, we propose exploiting sparse representation theory for graph prompting and present Graph Sparse Prompting (GSP). GSP aims to adaptively and sparsely select the optimal elements (e.g., certain node attributes) to achieve compact prompting for downstream tasks. Specifically, we propose two kinds of GSP models, termed Graph Sparse Feature Prompting (GSFP) and Graph Sparse multi-Feature Prompting (GSmFP). Both GSFP and GSmFP provide a general scheme for tuning any specific pre-trained GNNs that can achieve attribute selection and compact prompt learning simultaneously. A simple yet effective algorithm has been designed for solving GSFP and GSmFP models. Experiments on 16 widely-used benchmark datasets validate the effectiveness and advantages of the proposed GSFPs.
Authors: Meitong Liu, Xiaoyuan Zhang, Chulin Xie, Kate Donahue, Han Zhao
Abstract: The goal of multi-objective optimization (MOO) is to learn under multiple, potentially conflicting, objectives. One widely used technique to tackle MOO is through linear scalarization, where one fixed preference vector is used to combine the objectives into a single scalar value for optimization. However, recent work (Hu et al., 2024) has shown linear scalarization often fails to capture the non-convex regions of the Pareto Front, failing to recover the complete set of Pareto optimal solutions. In light of the above limitations, this paper focuses on Tchebycheff scalarization that optimizes for the worst-case objective. In particular, we propose an online mirror descent algorithm for Tchebycheff scalarization, which we call OMD-TCH. We show that OMD-TCH enjoys a convergence rate of $O(\sqrt{\log m/T})$ where $m$ is the number of objectives and $T$ is the number of iteration rounds. We also propose a novel adaptive online-to-batch conversion scheme that significantly improves the practical performance of OMD-TCH while maintaining the same convergence guarantees. We demonstrate the effectiveness of OMD-TCH and the adaptive conversion scheme on both synthetic problems and federated learning tasks under fairness constraints, showing state-of-the-art performance.
Authors: Li Shen, Anke Tang, Enneng Yang, Guibing Guo, Yong Luo, Lefei Zhang, Xiaochun Cao, Bo Du, Dacheng Tao
Abstract: Multi-task learning (MTL) leverages a shared model to accomplish multiple tasks and facilitate knowledge transfer. Recent research on task arithmetic-based MTL demonstrates that merging the parameters of independently fine-tuned models can effectively achieve MTL. However, existing merging methods primarily seek a static optimal solution within the original model parameter space, which often results in performance degradation due to the inherent diversity among tasks and potential interferences. To address this challenge, in this paper, we propose a Weight-Ensembling Mixture of Experts (WEMoE) method for multi-task model merging. Specifically, we first identify critical (or sensitive) modules by analyzing parameter variations in core modules of Transformer-based models before and after finetuning. Then, our WEMoE statically merges non-critical modules while transforming critical modules into a mixture-of-experts (MoE) structure. During inference, expert modules in the MoE are dynamically merged based on input samples, enabling a more flexible and adaptive merging approach. Building on WEMoE, we further introduce an efficient-and-effective WEMoE (E-WEMoE) method, whose core mechanism involves eliminating non-essential elements in the critical modules of WEMoE and implementing shared routing across multiple MoE modules, thereby significantly reducing both the trainable parameters, the overall parameter count, and computational overhead of the merged model by WEMoE. Experimental results across various architectures and tasks demonstrate that both WEMoE and E-WEMoE outperform state-of-the-art (SOTA) model merging methods in terms of MTL performance, generalization, and robustness.
Authors: Shaobo Wang, Hongxuan Tang, Mingyang Wang, Hongrui Zhang, Xuyang Liu, Weiya Li, Xuming Hu, Linfeng Zhang
Abstract: The debate between self-interpretable models and post-hoc explanations for black-box models is central to Explainable AI (XAI). Self-interpretable models, such as concept-based networks, offer insights by connecting decisions to human-understandable concepts but often struggle with performance and scalability. Conversely, post-hoc methods like Shapley values, while theoretically robust, are computationally expensive and resource-intensive. To bridge the gap between these two lines of research, we propose a novel method that combines their strengths, providing theoretically guaranteed self-interpretability for black-box models without compromising prediction accuracy. Specifically, we introduce a parameter-efficient pipeline, *AutoGnothi*, which integrates a small side network into the black-box model, allowing it to generate Shapley value explanations without changing the original network parameters. This side-tuning approach significantly reduces memory, training, and inference costs, outperforming traditional parameter-efficient methods, where full fine-tuning serves as the optimal baseline. *AutoGnothi* enables the black-box model to predict and explain its predictions with minimal overhead. Extensive experiments show that *AutoGnothi* offers accurate explanations for both vision and language tasks, delivering superior computational efficiency with comparable interpretability.
Authors: Gyeonghoon Ko, Hyunsu Kim, Juho Lee
Abstract: Exploiting symmetry inherent in data can significantly improve the sample efficiency of a learning procedure and the generalization of learned models. When data clearly reveals underlying symmetry, leveraging this symmetry can naturally inform the design of model architectures or learning strategies. Yet, in numerous real-world scenarios, identifying the specific symmetry within a given data distribution often proves ambiguous. To tackle this, some existing works learn symmetry in a data-driven manner, parameterizing and learning expected symmetry through data. However, these methods often rely on explicit knowledge, such as pre-defined Lie groups, which are typically restricted to linear or affine transformations. In this paper, we propose a novel symmetry learning algorithm based on transformations defined with one-parameter groups, continuously parameterized transformations flowing along the directions of vector fields called infinitesimal generators. Our method is built upon minimal inductive biases, encompassing not only commonly utilized symmetries rooted in Lie groups but also extending to symmetries derived from nonlinear generators. To learn these symmetries, we introduce a notion of a validity score that examine whether the transformed data is still valid for the given task. The validity score is designed to be fully differentiable and easily computable, enabling effective searches for transformations that achieve symmetries innate to the data. We apply our method mainly in two domains: image data and partial differential equations, and demonstrate its advantages. Our codes are available at \url{https://github.com/kogyeonghoon/learning-symmetry-from-scratch.git}.
URLs: https://github.com/kogyeonghoon/learning-symmetry-from-scratch.git
Authors: Patrik Reizinger, Alice Bizeul, Attila Juhos, Julia E. Vogt, Randall Balestriero, Wieland Brendel, David Klindt
Abstract: Supervised learning has become a cornerstone of modern machine learning, yet a comprehensive theory explaining its effectiveness remains elusive. Empirical phenomena, such as neural analogy-making and the linear representation hypothesis, suggest that supervised models can learn interpretable factors of variation in a linear fashion. Recent advances in self-supervised learning, particularly nonlinear Independent Component Analysis, have shown that these methods can recover latent structures by inverting the data generating process. We extend these identifiability results to parametric instance discrimination, then show how insights transfer to the ubiquitous setting of supervised learning with cross-entropy minimization. We prove that even in standard classification tasks, models learn representations of ground-truth factors of variation up to a linear transformation. We corroborate our theoretical contribution with a series of empirical studies. First, using simulated data matching our theoretical assumptions, we demonstrate successful disentanglement of latent factors. Second, we show that on DisLib, a widely-used disentanglement benchmark, simple classification tasks recover latent structures up to linear transformations. Finally, we reveal that models trained on ImageNet encode representations that permit linear decoding of proxy factors of variation. Together, our theoretical findings and experiments offer a compelling explanation for recent observations of linear representations, such as superposition in neural networks. This work takes a significant step toward a cohesive theory that accounts for the unreasonable effectiveness of supervised deep learning.
Authors: Rajana Akter, Shahnure Rabib, Rahul Deb Mohalder, Laboni Paul, Ferdous Bin Ali
Abstract: Intrusion detection system (IDS) is a piece of hardware or software that looks for malicious activity or policy violations in a network. It looks for malicious activity or security flaws on a network or system. IDS protects hosts or networks by looking for indications of known attacks or deviations from normal behavior (Network-based intrusion detection system, or NIDS for short). Due to the rapidly increasing amount of network data, traditional intrusion detection systems (IDSs) are far from being able to quickly and efficiently identify complex and varied network attacks, especially those linked to low-frequency attacks. The SCGNet (Stacked Convolution with Gated Recurrent Unit Network) is a novel deep learning architecture that we propose in this study. It exhibits promising results on the NSL-KDD dataset in both task, network attack detection, and attack type classification with 99.76% and 98.92% accuracy, respectively. We have also introduced a general data preprocessing pipeline that is easily applicable to other similar datasets. We have also experimented with conventional machine-learning techniques to evaluate the performance of the data processing pipeline.
Authors: Gabriele Onorato
Abstract: This study investigates the application of Bayesian Optimization (BO) for the hyperparameter tuning of neural networks, specifically targeting the enhancement of Convolutional Neural Networks (CNN) for image classification tasks. Bayesian Optimization is a derivative-free global optimization method suitable for expensive black-box functions with continuous inputs and limited evaluation budgets. The BO algorithm leverages Gaussian Process regression and acquisition functions like Upper Confidence Bound (UCB) and Expected Improvement (EI) to identify optimal configurations effectively. Using the Ax and BOTorch frameworks, this work demonstrates the efficiency of BO in reducing the number of hyperparameter tuning trials while achieving competitive model performance. Experimental outcomes reveal that BO effectively balances exploration and exploitation, converging rapidly towards optimal settings for CNN architectures. This approach underlines the potential of BO in automating neural network tuning, contributing to improved accuracy and computational efficiency in machine learning pipelines.
Authors: Kaustubh Kislay, Shlok Singh, Soham Joshi, Rohan Dutta, Jay Shim George Flint, Kevin Zhu
Abstract: Symbolic Regression remains an NP-Hard problem, with extensive research focusing on AI models for this task. Transformer models have shown promise in Symbolic Regression, but performance suffers with smaller datasets. We propose applying k-fold cross-validation to a transformer-based symbolic regression model trained on a significantly reduced dataset (15,000 data points, down from 500,000). This technique partitions the training data into multiple subsets (folds), iteratively training on some while validating on others. Our aim is to provide an estimate of model generalization and mitigate overfitting issues associated with smaller datasets. Results show that this process improves the model's output consistency and generalization by a relative improvement in validation loss of 53.31%. Potentially enabling more efficient and accessible symbolic regression in resource-constrained environments.
Authors: Senne Deproost, Denis Steckelmacher, Ann Now\'e
Abstract: With Deep Reinforcement Learning (DRL) being increasingly considered for the control of real-world systems, the lack of transparency of the neural network at the core of RL becomes a concern. Programmatic Reinforcement Learning (PRL) is able to to create representations of this black-box in the form of source code, not only increasing the explainability of the controller but also allowing for user adaptations. However, these methods focus on distilling a black-box policy into a program and do so after learning using the Mean Squared Error between produced and wanted behaviour, discarding other elements of the RL algorithm. The distilled policy may therefore perform significantly worse than the black-box learned policy. In this paper, we propose to directly learn a program as the policy of an RL agent. We build on TD3 and use its critics as the basis of the objective function of a genetic algorithm that syntheses the program. Our approach builds the program during training, as opposed to after the fact. This steers the program to actual high rewards, instead of a simple Mean Squared Error. Also, our approach leverages the TD3 critics to achieve high sample-efficiency, as opposed to pure genetic methods that rely on Monte-Carlo evaluations. Our experiments demonstrate the validity, explainability and sample-efficiency of our approach in a simple gridworld environment.
Authors: Emanuele Ledda, Giovanni Scodeller, Daniele Angioni, Giorgio Piras, Antonio Emanuele Cin\`a, Giorgio Fumera, Battista Biggio, Fabio Roli
Abstract: In learning problems, the noise inherent to the task at hand hinders the possibility to infer without a certain degree of uncertainty. Quantifying this uncertainty, regardless of its wide use, assumes high relevance for security-sensitive applications. Within these scenarios, it becomes fundamental to guarantee good (i.e., trustworthy) uncertainty measures, which downstream modules can securely employ to drive the final decision-making process. However, an attacker may be interested in forcing the system to produce either (i) highly uncertain outputs jeopardizing the system's availability or (ii) low uncertainty estimates, making the system accept uncertain samples that would instead require a careful inspection (e.g., human intervention). Therefore, it becomes fundamental to understand how to obtain robust uncertainty estimates against these kinds of attacks. In this work, we reveal both empirically and theoretically that defending against adversarial examples, i.e., carefully perturbed samples that cause misclassification, additionally guarantees a more secure, trustworthy uncertainty estimate under common attack scenarios without the need for an ad-hoc defense strategy. To support our claims, we evaluate multiple adversarial-robust models from the publicly available benchmark RobustBench on the CIFAR-10 and ImageNet datasets.
Authors: Kun Song, Ruben Solozabal, Li hao, Lu Ren, Moloud Abdar, Qing Li, Fakhri Karray, Martin Takac
Abstract: Hyperbolic representation learning is well known for its ability to capture hierarchical information. However, the distance between samples from different levels of hierarchical classes can be required large. We reveal that the hyperbolic discriminant objective forces the backbone to capture this hierarchical information, which may inevitably increase the Lipschitz constant of the backbone. This can hinder the full utilization of the backbone's generalization ability. To address this issue, we introduce second-order pooling into hyperbolic representation learning, as it naturally increases the distance between samples without compromising the generalization ability of the input features. In this way, the Lipschitz constant of the backbone does not necessarily need to be large. However, current off-the-shelf low-dimensional bilinear pooling methods cannot be directly employed in hyperbolic representation learning because they inevitably reduce the distance expansion capability. To solve this problem, we propose a kernel approximation regularization, which enables the low-dimensional bilinear features to approximate the kernel function well in low-dimensional space. Finally, we conduct extensive experiments on graph-structured datasets to demonstrate the effectiveness of the proposed method.
Authors: Nikolaos Tsilivis, Gal Vardi, Julia Kempe
Abstract: We study the implicit bias of the general family of steepest descent algorithms, which includes gradient descent, sign descent and coordinate descent, in deep homogeneous neural networks. We prove that an algorithm-dependent geometric margin starts increasing once the networks reach perfect training accuracy and characterize the late-stage bias of the algorithms. In particular, we define a generalized notion of stationarity for optimization problems and show that the algorithms progressively reduce a (generalized) Bregman divergence, which quantifies proximity to such stationary points of a margin-maximization problem. We then experimentally zoom into the trajectories of neural networks optimized with various steepest descent algorithms, highlighting connections to the implicit bias of Adam.
Authors: Zhiqi Bu, Xiaomeng Jin, Bhanukiran Vinzamuri, Anil Ramakrishna, Kai-Wei Chang, Volkan Cevher, Mingyi Hong
Abstract: Machine unlearning has been used to remove unwanted knowledge acquired by large language models (LLMs). In this paper, we examine machine unlearning from an optimization perspective, framing it as a regularized multi-task optimization problem, where one task optimizes a forgetting objective and another optimizes the model performance. In particular, we introduce a normalized gradient difference (NGDiff) algorithm, enabling us to have better control over the trade-off between the objectives, while integrating a new, automatic learning rate scheduler. We provide a theoretical analysis and empirically demonstrate the superior performance of NGDiff among state-of-the-art unlearning methods on the TOFU and MUSE datasets while exhibiting stable training.
Authors: Xinyue Feng, Jinquan Hang, Yuequn Zhang, Haotian Wang, Desheng Zhang, Guang Wang
Abstract: Heterogeneous graph is an important structure for modeling complex relational data in real-world scenarios and usually involves various node prediction tasks within a single graph. Training these tasks separately may neglect beneficial information sharing, hence a preferred way is to learn several tasks in a same model by Multi-Task Learning (MTL). However, MTL introduces the issue of negative transfer, where the training of different tasks interferes with each other as they may focus on different information from the data, resulting in suboptimal performance. To solve the issue, existing MTL methods use separate backbones for each task, then selectively exchange beneficial features through interactions among the output embeddings from each layer of different backbones, which we refer to as outer-layer exchange. However, the negative transfer in heterogeneous graphs arises not simply from the varying importance of an individual node feature across tasks, but also from the varying importance of inter-relation between two nodes across tasks. These inter-relations are entangled in the output embedding, making it difficult for existing methods to discriminate beneficial information from the embedding. To address this challenge, we propose the Inner-Layer Information Exchange (InLINE) model that facilitate fine-grained information exchanges within each graph layer rather than through output embeddings. Specifically, InLINE consists of (1) Structure Disentangled Experts for layer-wise structure disentanglement, (2) Structure Disentangled Gates for assigning disentangled information to different tasks. Evaluations on two public datasets and a large industry dataset show that our model effectively alleviates the significant performance drop on specific tasks caused by negative transfer, improving Macro F1 by 6.3% on DBLP dataset and AUC by 3.6% on the industry dataset compared to SoA methods.
Authors: Lior Dikstein, Ariel Lapid, Arnon Netzer, Hai Victor Habi
Abstract: Zero-shot quantization (ZSQ) using synthetic data is a key approach for post-training quantization (PTQ) under privacy and security constraints. However, existing data generation methods often struggle to effectively generate data suitable for hardware-friendly quantization, where all model layers are quantized. We analyze existing data generation methods based on batch normalization (BN) matching and identify several gaps between synthetic and real data: 1) Current generation algorithms do not optimize the entire synthetic dataset simultaneously; 2) Data augmentations applied during training are often overlooked; and 3) A distribution shift occurs in the final model layers due to the absence of BN in those layers. These gaps negatively impact ZSQ performance, particularly in hardware-friendly quantization scenarios. In this work, we propose Data Generation for Hardware-friendly quantization (DGH), a novel method that addresses these gaps. DGH jointly optimizes all generated images, regardless of the image set size or GPU memory constraints. To address data augmentation mismatches, DGH includes a preprocessing stage that mimics the augmentation process and enhances image quality by incorporating natural image priors. Finally, we propose a new distribution-stretching loss that aligns the support of the feature map distribution between real and synthetic data. This loss is applied to the model's output and can be adapted to various tasks. DGH demonstrates significant improvements in quantization performance across multiple tasks, achieving up to a 30% increase in accuracy for hardware-friendly ZSQ in both classification and object detection, often performing on par with real data.
Authors: Ildus Sadrtdinov, Maxim Kodryan, Eduard Pokonechny, Ekaterina Lobacheva, Dmitry Vetrov
Abstract: It is generally accepted that starting neural networks training with large learning rates (LRs) improves generalization. Following a line of research devoted to understanding this effect, we conduct an empirical study in a controlled setting focusing on two questions: 1) how large an initial LR is required for obtaining optimal quality, and 2) what are the key differences between models trained with different LRs? We discover that only a narrow range of initial LRs slightly above the convergence threshold lead to optimal results after fine-tuning with a small LR or weight averaging. By studying the local geometry of reached minima, we observe that using LRs from this optimal range allows for the optimization to locate a basin that only contains high-quality minima. Additionally, we show that these initial LRs result in a sparse set of learned features, with a clear focus on those most relevant for the task. In contrast, starting training with too small LRs leads to unstable minima and attempts to learn all features simultaneously, resulting in poor generalization. Conversely, using initial LRs that are too large fails to detect a basin with good solutions and extract meaningful patterns from the data.
Authors: Qiuhao Wang, Shaohang Xu, Chin Pang Ho, Marek Petrick
Abstract: We develop a generic policy gradient method with the global optimality guarantee for robust Markov Decision Processes (MDPs). While policy gradient methods are widely used for solving dynamic decision problems due to their scalable and efficient nature, adapting these methods to account for model ambiguity has been challenging, often making it impractical to learn robust policies. This paper introduces a novel policy gradient method, Double-Loop Robust Policy Mirror Descent (DRPMD), for solving robust MDPs. DRPMD employs a general mirror descent update rule for the policy optimization with adaptive tolerance per iteration, guaranteeing convergence to a globally optimal policy. We provide a comprehensive analysis of DRPMD, including new convergence results under both direct and softmax parameterizations, and provide novel insights into the inner problem solution through Transition Mirror Ascent (TMA). Additionally, we propose innovative parametric transition kernels for both discrete and continuous state-action spaces, broadening the applicability of our approach. Empirical results validate the robustness and global convergence of DRPMD across various challenging robust MDP settings.
Authors: Barbara Hoffmann, Jana Vatter, Ruben Mayer
Abstract: The European Union's Artificial Intelligence Act (AI Act) introduces comprehensive guidelines for the development and oversight of Artificial Intelligence (AI) and Machine Learning (ML) systems, with significant implications for Graph Neural Networks (GNNs). This paper addresses the unique challenges posed by the AI Act for GNNs, which operate on complex graph-structured data. The legislation's requirements for data management, data governance, robustness, human oversight, and privacy necessitate tailored strategies for GNNs. Our study explores the impact of these requirements on GNN training and proposes methods to ensure compliance. We provide an in-depth analysis of bias, robustness, explainability, and privacy in the context of GNNs, highlighting the need for fair sampling strategies and effective interpretability techniques. Our contributions fill the research gap by offering specific guidance for GNNs under the new legislative framework and identifying open questions and future research directions.
Authors: Pin-Yen Huang, Szu-Wei Fu, Yu Tsao
Abstract: State-of-the-art (SOTA) semi-supervised learning techniques, such as FixMatch and it's variants, have demonstrated impressive performance in classification tasks. However, these methods are not directly applicable to regression tasks. In this paper, we present RankUp, a simple yet effective approach that adapts existing semi-supervised classification techniques to enhance the performance of regression tasks. RankUp achieves this by converting the original regression task into a ranking problem and training it concurrently with the original regression objective. This auxiliary ranking classifier outputs a classification result, thus enabling integration with existing semi-supervised classification methods. Moreover, we introduce regression distribution alignment (RDA), a complementary technique that further enhances RankUp's performance by refining pseudo-labels through distribution alignment. Despite its simplicity, RankUp, with or without RDA, achieves SOTA results in across a range of regression benchmarks, including computer vision, audio, and natural language processing tasks. Our code and log data are open-sourced at https://github.com/pm25/semi-supervised-regression.
Authors: Raymond Chua, Arna Ghosh, Christos Kaplanis, Blake A. Richards, Doina Precup
Abstract: In Deep Reinforcement Learning (RL), it is a challenge to learn representations that do not exhibit catastrophic forgetting or interference in non-stationary environments. Successor Features (SFs) offer a potential solution to this challenge. However, canonical techniques for learning SFs from pixel-level observations often lead to representation collapse, wherein representations degenerate and fail to capture meaningful variations in the data. More recent methods for learning SFs can avoid representation collapse, but they often involve complex losses and multiple learning phases, reducing their efficiency. We introduce a novel, simple method for learning SFs directly from pixels. Our approach uses a combination of a Temporal-difference (TD) loss and a reward prediction loss, which together capture the basic mathematical definition of SFs. We show that our approach matches or outperforms existing SF learning techniques in both 2D (Minigrid), 3D (Miniworld) mazes and Mujoco, for both single and continual learning scenarios. As well, our technique is efficient, and can reach higher levels of performance in less time than other approaches. Our work provides a new, streamlined technique for learning SFs directly from pixel observations, with no pretraining required.
Authors: Jonghong Jeon
Abstract: Artificial Intelligence (AI) has rapidly evolved over the past decade and has advanced in areas such as language comprehension, image and video recognition, programming, and scientific reasoning. Recent AI technologies based on large language models and foundation models are approaching or surpassing artificial general intelligence. These systems demonstrate superior performance in complex problem solving, natural language processing, and multi-domain tasks, and can potentially transform fields such as science, industry, healthcare, and education. However, these advancements have raised concerns regarding the safety and trustworthiness of advanced AI, including risks related to uncontrollability, ethical conflicts, long-term socioeconomic impacts, and safety assurance. Efforts are being expended to develop internationally agreed-upon standards to ensure the safety and reliability of AI. This study analyzes international trends in safety and trustworthiness standardization for advanced AI, identifies key areas for standardization, proposes future directions and strategies, and draws policy implications. The goal is to support the safe and trustworthy development of advanced AI and enhance international competitiveness through effective standardization.
Authors: Matin Mortaheb, Priyanka Kaswan, Sennur Ulukus
Abstract: Federated learning (FL) is a collaborative approach where multiple clients, coordinated by a parameter server (PS), train a unified machine-learning model. The approach, however, suffers from two key challenges: data heterogeneity and communication overhead. Data heterogeneity refers to inconsistencies in model training arising from heterogeneous data at different clients. Communication overhead arises from the large volumes of parameter updates exchanged between the PS and clients. Existing solutions typically address these challenges separately. This paper introduces a new communication-efficient algorithm that uses the age of information metric to simultaneously tackle both limitations of FL. We introduce age vectors at the PS, which keep track of how often the different model parameters are updated from the clients. The PS uses this to selectively request updates for specific gradient indices from each client. Further, the PS employs age vectors to identify clients with statistically similar data and group them into clusters. The PS combines the age vectors of the clustered clients to efficiently coordinate gradient index updates among clients within a cluster. We evaluate our approach using the MNIST and CIFAR10 datasets in highly non-i.i.d. settings. The experimental results show that our proposed method can expedite training, surpassing other communication-efficient strategies in efficiency.
Authors: Bowen Liu, Haoyang Li, Shuning Wang, Shuo Nie, Shanghang Zhang
Abstract: Out-of-distribution (OOD) generalization in Graph Neural Networks (GNNs) has gained significant attention due to its critical importance in graph-based predictions in real-world scenarios. Existing methods primarily focus on extracting a single causal subgraph from the input graph to achieve generalizable predictions. However, relying on a single subgraph can lead to susceptibility to spurious correlations and is insufficient for learning invariant patterns behind graph data. Moreover, in many real-world applications, such as molecular property prediction, multiple critical subgraphs may influence the target label property. To address these challenges, we propose a novel framework, SubGraph Aggregation (SuGAr), designed to learn a diverse set of subgraphs that are crucial for OOD generalization on graphs. Specifically, SuGAr employs a tailored subgraph sampler and diversity regularizer to extract a diverse set of invariant subgraphs. These invariant subgraphs are then aggregated by averaging their representations, which enriches the subgraph signals and enhances coverage of the underlying causal structures, thereby improving OOD generalization. Extensive experiments on both synthetic and real-world datasets demonstrate that \ours outperforms state-of-the-art methods, achieving up to a 24% improvement in OOD generalization on graphs. To the best of our knowledge, this is the first work to study graph OOD generalization by learning multiple invariant subgraphs.
Authors: Saeed Mahloujifar, Luca Melis, Kamalika Chaudhuri
Abstract: Empirical auditing has emerged as a means of catching some of the flaws in the implementation of privacy-preserving algorithms. Existing auditing mechanisms, however, are either computationally inefficient requiring multiple runs of the machine learning algorithms or suboptimal in calculating an empirical privacy. In this work, we present a tight and efficient auditing procedure and analysis that can effectively assess the privacy of mechanisms. Our approach is efficient; similar to the recent work of Steinke, Nasr, and Jagielski (2023), our auditing procedure leverages the randomness of examples in the input dataset and requires only a single run of the target mechanism. And it is more accurate; we provide a novel analysis that enables us to achieve tight empirical privacy estimates by using the hypothesized $f$-DP curve of the mechanism, which provides a more accurate measure of privacy than the traditional $\epsilon,\delta$ differential privacy parameters. We use our auditing procure and analysis to obtain empirical privacy, demonstrating that our auditing procedure delivers tighter privacy estimates.
Authors: Pulkit Gopalani, Ekdeep Singh Lubana, Wei Hu
Abstract: Recent analysis on the training dynamics of Transformers has unveiled an interesting characteristic: the training loss plateaus for a significant number of training steps, and then suddenly (and sharply) drops to near--optimal values. To understand this phenomenon in depth, we formulate the low-rank matrix completion problem as a masked language modeling (MLM) task, and show that it is possible to train a BERT model to solve this task to low error. Furthermore, the loss curve shows a plateau early in training followed by a sudden drop to near-optimal values, despite no changes in the training procedure or hyper-parameters. To gain interpretability insights into this sudden drop, we examine the model's predictions, attention heads, and hidden states before and after this transition. Concretely, we observe that (a) the model transitions from simply copying the masked input to accurately predicting the masked entries; (b) the attention heads transition to interpretable patterns relevant to the task; and (c) the embeddings and hidden states encode information relevant to the problem. We also analyze the training dynamics of individual model components to understand the sudden drop in loss.
Authors: Hongyi Xu
Abstract: Multivariate time series anomaly detection technology plays an important role in many fields including aerospace, water treatment, cloud service providers, etc. Excellent anomaly detection models can greatly improve work efficiency and avoid major economic losses. However, with the development of technology, the increasing size and complexity of data, and the lack of labels for relevant abnormal data, it is becoming increasingly challenging to perform effective and accurate anomaly detection in high-dimensional and complex data sets. In this paper, we propose a hypergraph based spatiotemporal graph convolutional neural network model STGCN_Hyper, which explicitly captures high-order, multi-hop correlations between multiple variables through a hypergraph based dynamic graph structure learning module. On this basis, we further use the hypergraph based spatiotemporal graph convolutional network to utilize the learned hypergraph structure to effectively propagate and aggregate one-hop and multi-hop related node information in the convolutional network, thereby obtaining rich spatial information. Furthermore, through the multi-scale TCN dilated convolution module, the STGCN_hyper model can also capture the dependencies of features at different scales in the temporal dimension. An unsupervised anomaly detector based on PCA and GMM is also integrated into the STGCN_hyper model. Through the anomaly score of the detector, the model can detect the anomalies in an unsupervised way. Experimental results on multiple time series datasets show that our model can flexibly learn the multi-scale time series features in the data and the dependencies between features, and outperforms most existing baseline models in terms of precision, recall, F1-score on anomaly detection tasks. Our code is available on: https://git.ecdf.ed.ac.uk/msc-23-24/s2044819
Authors: Patricia Pauli, Ruigang Wang, Ian Manchester, Frank Allg\"ower
Abstract: We propose a novel layer-wise parameterization for convolutional neural networks (CNNs) that includes built-in robustness guarantees by enforcing a prescribed Lipschitz bound. Each layer in our parameterization is designed to satisfy a linear matrix inequality (LMI), which in turn implies dissipativity with respect to a specific supply rate. Collectively, these layer-wise LMIs ensure Lipschitz boundedness for the input-output mapping of the neural network, yielding a more expressive parameterization than through spectral bounds or orthogonal layers. Our new method LipKernel directly parameterizes dissipative convolution kernels using a 2-D Roesser-type state space model. This means that the convolutional layers are given in standard form after training and can be evaluated without computational overhead. In numerical experiments, we show that the run-time using our method is orders of magnitude faster than state-of-the-art Lipschitz-bounded networks that parameterize convolutions in the Fourier domain, making our approach particularly attractive for improving robustness of learning-based real-time perception or control in robotics, autonomous vehicles, or automation systems. We focus on CNNs, and in contrast to previous works, our approach accommodates a wide variety of layers typically used in CNNs, including 1-D and 2-D convolutional layers, maximum and average pooling layers, as well as strided and dilated convolutions and zero padding. However, our approach naturally extends beyond CNNs as we can incorporate any layer that is incrementally dissipative.
Authors: Jacob L. Block, Sundararajan Srinivasan, Liam Collins, Aryan Mokhtari, Sanjay Shakkottai
Abstract: The power of foundation models (FMs) lies in their capacity to learn highly expressive representations that can be adapted to a broad spectrum of tasks. However, these pretrained models require multiple stages of fine-tuning to become effective for downstream applications. Conventionally, the model is first retrained on the aggregate of a diverse set of tasks of interest and then adapted to specific low-resource downstream tasks by utilizing a parameter-efficient fine-tuning (PEFT) scheme. While this two-phase procedure seems reasonable, the independence of the retraining and fine-tuning phases causes a major issue, as there is no guarantee the retrained model will achieve good performance post-fine-tuning. To explicitly address this issue, we introduce a meta-learning framework infused with PEFT in this intermediate retraining stage to learn a model that can be easily adapted to unseen tasks. For our theoretical results, we focus on linear models using low-rank adaptations. In this setting, we demonstrate the suboptimality of standard retraining for finding an adaptable set of parameters. Further, we prove that our method recovers the optimally adaptable parameters. We then apply these theoretical insights to retraining the RoBERTa model to predict the continuation of conversations between different personas within the ConvAI2 dataset. Empirically, we observe significant performance benefits using our proposed meta-learning scheme during retraining relative to the conventional approach.
Authors: Nate Gillman, Daksh Aggarwal, Michael Freeman, Saurabh Singh, Chen Sun
Abstract: As the quality of large language models has improved, there has been increased interest in using them to model non-linguistic tokens. For example, the Decision Transformer recasts agentic decision making as a sequence modeling problem, using a decoder-only LLM to model the distribution over the discrete action space for an Atari agent. However, when adapting LLMs to non-linguistic domains, it remains unclear if softmax over discrete bins captures the continuous structure of the tokens and the potentially complex distributions needed for high quality token generation. We introduce a neural network layer, constructed using Fourier series, which we can easily substitute for any linear layer if we want the outputs to have a more continuous structure. We perform extensive analysis on synthetic datasets, as well as on large-scale decision making and time series forecasting tasks. We also provide theoretical evidence that this layer can better learn signal from data while ignoring high-frequency noise. All of our results support the effectiveness of our proposed Fourier head in scenarios where the underlying data distribution has a natural continuous structure. For example, the Fourier head improves a Decision Transformer agent's returns by 46% on the Atari Seaquest game, and increases a state-of-the-art times series foundation model's forecasting performance by 3.5% across 20 benchmarks unseen during training.
Authors: Angelica Chen, Samuel D. Stanton, Robert G. Alberstein, Andrew M. Watkins, Richard Bonneau, Vladimir Gligorijevi, Kyunghyun Cho, Nathan C. Frey
Abstract: Large language models (LLMs) have recently shown significant potential in various biological tasks such as protein engineering and molecule design. These tasks typically involve black-box discrete sequence optimization, where the challenge lies in generating sequences that are not only biologically feasible but also adhere to hard fine-grained constraints. However, LLMs often struggle with such constraints, especially in biological contexts where verifying candidate solutions is costly and time-consuming. In this study, we explore the possibility of employing LLMs as highly-constrained bilevel optimizers through a methodology we refer to as Language Model Optimization with Margin Expectation (LLOME). This approach combines both offline and online optimization, utilizing limited oracle evaluations to iteratively enhance the sequences generated by the LLM. We additionally propose a novel training objective -- Margin-Aligned Expectation (MargE) -- that trains the LLM to smoothly interpolate between the reward and reference distributions. Lastly, we introduce a synthetic test suite that bears strong geometric similarity to real biophysical problems and enables rapid evaluation of LLM optimizers without time-consuming lab validation. Our findings reveal that, in comparison to genetic algorithm baselines, LLMs achieve significantly lower regret solutions while requiring fewer test function evaluations. However, we also observe that LLMs exhibit moderate miscalibration, are susceptible to generator collapse, and have difficulty finding the optimal solution when no explicit ground truth rewards are available.
Authors: Yifan Sun, Yuhang Li, Yue Zhang, Yuchen Jin, Huan Zhang
Abstract: Open-source Large Language Models (LLMs) have recently demonstrated remarkable capabilities in natural language understanding and generation, leading to widespread adoption across various domains. However, their increasing model sizes render local deployment impractical for individual users, pushing many to rely on computing service providers for inference through a blackbox API. This reliance introduces a new risk: a computing provider may stealthily substitute the requested LLM with a smaller, less capable model without consent from users, thereby delivering inferior outputs while benefiting from cost savings. In this paper, we formalize the problem of verifiable inference for LLMs. Existing verifiable computing solutions based on cryptographic or game-theoretic techniques are either computationally uneconomical or rest on strong assumptions. We introduce SVIP, a secret-based verifiable LLM inference protocol that leverages intermediate outputs from LLM as unique model identifiers. By training a proxy task on these outputs and requiring the computing provider to return both the generated text and the processed intermediate outputs, users can reliably verify whether the computing provider is acting honestly. In addition, the integration of a secret mechanism further enhances the security of our protocol. We thoroughly analyze our protocol under multiple strong and adaptive adversarial scenarios. Our extensive experiments demonstrate that SVIP is accurate, generalizable, computationally efficient, and resistant to various attacks. Notably, SVIP achieves false negative rates below 5% and false positive rates below 3%, while requiring less than 0.01 seconds per query for verification.
Authors: Karthik Prakhya, Tolga Birdal, Alp Yurtsever
Abstract: Solving non-convex, NP-hard optimization problems is crucial for training machine learning models, including neural networks. However, non-convexity often leads to black-box machine learning models with unclear inner workings. While convex formulations have been used for verifying neural network robustness, their application to training neural networks remains less explored. In response to this challenge, we reformulate the problem of training infinite-width two-layer ReLU networks as a convex completely positive program in a finite-dimensional (lifted) space. Despite the convexity, solving this problem remains NP-hard due to the complete positivity constraint. To overcome this challenge, we introduce a semidefinite relaxation that can be solved in polynomial time. We then experimentally evaluate the tightness of this relaxation, demonstrating its competitive performance in test accuracy across a range of classification tasks.
Authors: Can Chen, Jun-Kun Wang
Abstract: Developing algorithms to differentiate between machine-generated texts and human-written texts has garnered substantial attention in recent years. Existing methods in this direction typically concern an offline setting where a dataset containing a mix of real and machine-generated texts is given upfront, and the task is to determine whether each sample in the dataset is from a large language model (LLM) or a human. However, in many practical scenarios, sources such as news websites, social media accounts, or on other forums publish content in a streaming fashion. Therefore, in this online scenario, how to quickly and accurately determine whether the source is an LLM with strong statistical guarantees is crucial for these media or platforms to function effectively and prevent the spread of misinformation and other potential misuse of LLMs. To tackle the problem of online detection, we develop an algorithm based on the techniques of sequential hypothesis testing by betting that not only builds upon and complements existing offline detection techniques but also enjoys statistical guarantees, which include a controlled false positive rate and the expected time to correctly identify a source as an LLM. Experiments were conducted to demonstrate the effectiveness of our method.
Authors: Taiwo A. Adebiyi, Bach Do, Ruda Zhang
Abstract: Bayesian optimization devolves the global optimization of a costly objective function to the global optimization of a sequence of acquisition functions. This inner-loop optimization can be catastrophically difficult if it involves posterior samples, especially in higher dimensions. We introduce an efficient global optimization strategy for posterior samples based on global rootfinding. It provides gradient-based optimizers with judiciously selected starting points, designed to combine exploitation and exploration. The algorithm scales practically linearly to high dimensions. For posterior sample-based acquisition functions such as Gaussian process Thompson sampling (GP-TS) and variants of entropy search, we demonstrate remarkable improvement in both inner- and outer-loop optimization, surprisingly outperforming alternatives like EI and GP-UCB in most cases. We also propose a sample-average formulation of GP-TS, which has a parameter to explicitly control exploitation and can be computed at the cost of one posterior sample. Our implementation is available at https://github.com/UQUH/TSRoots .
Authors: Joongwon Chae, Zhenyu Wang, Peiwu Qin
Abstract: Recent advancements in protein structure prediction, particularly AlphaFold2, have revolutionized structural biology by achieving near-experimental accuracy. However, the computational intensity of these models limits their application in high-throughput protein screening. Concurrently, large language models like ESM (Evolutionary Scale Modeling) have demonstrated the potential to extract rich structural information directly from protein sequences. Despite these advances, a significant gap remains in rapidly assessing protein structure quality for large-scale analyses. We introduce pLDDT-Predictor, a high-speed protein screening tool that bridges this gap by leveraging pre-trained ESM2 protein embeddings and a Transformer architecture to accurately predict AlphaFold2`s pLDDT (predicted Local Distance Difference Test) scores. Our model addresses the critical need for fast, accurate protein structure quality assessment without the computational burden of full structure prediction. By combining the evolutionary information captured in ESM2 embeddings with the sequence-wide context modeling of Transformers, pLDDT-Predictor achieves a balance between structural insight and computational efficiency. Our experimental results, conducted on a diverse dataset of 1.5 million protein sequences, demonstrate that pLDDT-Predictor can classify more than 90 percent of proteins with a pLDDT score above 70, closely matching AlphaFold2`s confidence level.
Authors: Angel Varela
Abstract: Predicting the stock market is a big challenge for the machine learning world. It is known how difficult it is to have accurate and consistent predictions with ML models. Some architectures are able to capture the movement of stocks but almost never are able to be launched to the production world. We present Achilles, with a classical architecture of LSTM(Long Short Term Memory) neural network this model is able to predict the Gold vs USD commodity. With the predictions minute-per-minute of this model we implemented a trading bot to run during 23 days of testing excluding weekends. At the end of the testing period we generated $1623.52 in profit with the methodology used. The results of our method demonstrate Machine Learning can successfully be implemented to predict the Gold vs USD commodity.
Authors: Azam Asilian Bidgoli, Shahryar Rahnamayan
Abstract: Feature selection is a crucial step in machine learning, especially for high-dimensional datasets, where irrelevant and redundant features can degrade model performance and increase computational costs. This paper proposes a novel large-scale multi-objective evolutionary algorithm based on the search space shrinking, termed LMSSS, to tackle the challenges of feature selection particularly as a sparse optimization problem. The method includes a shrinking scheme to reduce dimensionality of the search space by eliminating irrelevant features before the main evolutionary process. This is achieved through a ranking-based filtering method that evaluates features based on their correlation with class labels and frequency in an initial, cost-effective evolutionary process. Additionally, a smart crossover scheme based on voting between parent solutions is introduced, giving higher weight to the parent with better classification accuracy. An intelligent mutation process is also designed to target features prematurely excluded from the population, ensuring they are evaluated in combination with other features. These integrated techniques allow the evolutionary process to explore the search space more efficiently and effectively, addressing the sparse and high-dimensional nature of large-scale feature selection problems. The effectiveness of the proposed algorithm is demonstrated through comprehensive experiments on 15 large-scale datasets, showcasing its potential to identify more accurate feature subsets compared to state-of-the-art large-scale feature selection algorithms. These results highlight LMSSS's capability to improve model performance and computational efficiency, setting a new benchmark in the field.
Authors: Wen Ge, Guanyi Mou, Emmanuel O. Agu, Kyumin Lee
Abstract: Human Activity Recognition (HAR) is essential in ubiquitous computing, with far-reaching real-world applications. While recent SOTA HAR research has demonstrated impressive performance, some key aspects remain under-explored. Firstly, HAR can be both highly contextualized and personalized. However, prior work has predominantly focused on being Context-Aware (CA) while largely ignoring the necessity of being User-Aware (UA). We argue that addressing the impact of innate user action-performing differences is equally crucial as considering external contextual environment settings in HAR tasks. Secondly, being user-aware makes the model acknowledge user discrepancies but does not necessarily guarantee mitigation of these discrepancies, i.e., unified predictions under the same activities. There is a need for a methodology that explicitly enforces closer (different user, same activity) representations. To bridge this gap, we introduce CLAUDIA, a novel framework designed to address these issues. Specifically, we expand the contextual scope of the CA-HAR task by integrating User Identification (UI) within the CA-HAR framework, jointly predicting both CA-HAR and UI in a new task called User and Context-Aware HAR (UCA-HAR). This approach enriches personalized and contextual understanding by jointly learning user-invariant and user-specific patterns. Inspired by SOTA designs in the visual domain, we introduce a supervised contrastive loss objective on instance-instance pairs to enhance model efficacy and improve learned feature quality. Evaluation across three real-world CA-HAR datasets reveals substantial performance enhancements, with average improvements ranging from 5.8% to 14.1% in Matthew's Correlation Coefficient and 3.0% to 7.2% in Macro F1 score.
Authors: Liam Moroy, Guillaume Bourmaud, Fr\'ed\'eric Champagnat, Jean-Fran\c{c}ois Giovannelli
Abstract: Plug&Play (PnP) diffusion models are state-of-the-art methods in computed tomography (CT) reconstruction. Such methods usually consider applications where the sinogram contains a sufficient amount of information for the posterior distribution to be peaked, and consequently are evaluated using image-to-image metrics such as PSNR/SSIM. Instead, we are interested in reconstructing compressible flow images from sinograms having a small number of projections, which results in a posterior distribution no longer peaked or even multimodal. Thus, in this paper, we aim at evaluating the approximate posterior of PnP diffusion models and introduce two posterior evaluation criteria. We quantitatively evaluate three PnP diffusion methods on three different datasets for several numbers of projections. We surprisingly find that, for each method, the approximate posterior deviates from the true posterior when the number of projections decreases.
Authors: Marcel Roth, Micha V. Nowak
Abstract: Video capsule endoscopy has transformed gastrointestinal endoscopy (GIE) diagnostics by offering a non-invasive method for capturing detailed images of the gastrointestinal tract, enabling early disease detection. However, its potential is limited by the sheer volume of images generated during the imaging procedure, which can take anywhere from 6-8 hours and often produce up to 1 million images, necessitating automated analysis. Additionally, the variability of these images, combined with the need for expert annotations and the scarcity of large, high-quality labeled datasets, constrains the effectiveness of current medical image analysis models. To address this, we introduce a novel large gastrointestinal endoscopy dataset, called EndoExtend24, created by merging and re-stratifying the train/test splits of ten existing public and private datasets, ensuring no overlap of patient data across splits. EndoExtend24 includes over 226,000 labeled images, as well as dynamic class mappings, which allow unified training across datasets with differing labeling granularity, supporting up to 123 distinct pathological findings. Further, we propose to leverage domain adaptive pre-training of foundation models in computer vision trained with self-supervision on generic image data, to adapt them to the task of GIE medical diagnosis. Specifically, the EVA-02 model, which is based on the vision transformer architecture and was trained on ImageNet-22k with masked image modeling (using EVA-CLIP as a MIM teacher), is pre-trained on the novel EndoExtend24 dataset to achieve domain adaptation, and finally trained on the Capsule Endoscopy 2024 Challenge dataset. Experimental results show promising results on the challenge validation set, with an AUC Macro score of 0.993 and a balanced accuracy of 89.3%.
Authors: Serkan Sulun, Paula Viana, Matthew E. P. Davies
Abstract: We introduce VEMOCLAP: Video EMOtion Classifier using Pretrained features, the first readily available and open-source web application that analyzes the emotional content of any user-provided video. We improve our previous work, which exploits open-source pretrained models that work on video frames and audio, and then efficiently fuse the resulting pretrained features using multi-head cross-attention. Our approach increases the state-of-the-art classification accuracy on the Ekman-6 video emotion dataset by 4.3% and offers an online application for users to run our model on their own videos or YouTube videos. We invite the readers to try our application at serkansulun.com/app.
Authors: Chika Maduabuchi, Ericmoore Jossou, Matteo Bucci
Abstract: High-speed video (HSV) segmentation is essential for analyzing dynamic physical processes in scientific and industrial applications, such as boiling heat transfer. Existing models like U-Net struggle with generalization and accurately segmenting complex bubble formations. We present VideoSAM, a specialized adaptation of the Segment Anything Model (SAM), fine-tuned on a diverse HSV dataset for phase detection. Through diverse experiments, VideoSAM demonstrates superior performance across four fluid environments -- Water, FC-72, Nitrogen, and Argon -- significantly outperforming U-Net in complex segmentation tasks. In addition to introducing VideoSAM, we contribute an open-source HSV segmentation dataset designed for phase detection, enabling future research in this domain. Our findings underscore VideoSAM's potential to set new standards in robust and accurate HSV segmentation. The code and dataset used in this study are available online at https://github.com/chikap421/videosam .
Authors: Haoyue Bai
Abstract: Deep learning has been demonstrated with tremendous success in recent years. Despite so, its performance in practice often degenerates drastically when encountering out-of-distribution (OoD) data, i.e. training and test data are sampled from different distributions. In this thesis, we study ways toward robust OoD generalization for deep learning, i.e., its performance is not susceptible to distribution shift in the test data. We first propose a novel and effective approach to disentangle the spurious correlation between features that are not essential for recognition. It employs decomposed feature representation by orthogonalizing the two gradients of losses for category and context branches. Furthermore, we perform gradient-based augmentation on context-related features (e.g., styles, backgrounds, or scenes of target objects) to improve the robustness of learned representations. Results show that our approach generalizes well for different distribution shifts. We then study the problem of strengthening neural architecture search in OoD scenarios. We propose to optimize the architecture parameters that minimize the validation loss on synthetic OoD data, under the condition that corresponding network parameters minimize the training loss. Moreover, to obtain a proper validation set, we learn a conditional generator by maximizing their losses computed by different neural architectures. Results show that our approach effectively discovers robust architectures that perform well for OoD generalization.
Authors: E. Zhixuan Zeng, Yuhao Chen, Alexander Wong
Abstract: Recent advances in image generation have made diffusion models powerful tools for creating high-quality images. However, their iterative denoising process makes understanding and interpreting their semantic latent spaces more challenging than other generative models, such as GANs. Recent methods have attempted to address this issue by identifying semantically meaningful directions within the latent space. However, they often need manual interpretation or are limited in the number of vectors that can be trained, restricting their scope and utility. This paper proposes a novel framework for unsupervised exploration of diffusion latent spaces. We directly leverage natural language prompts and image captions to map latent directions. This method allows for the automatic understanding of hidden features and supports a broader range of analysis without the need to train specific vectors. Our method provides a more scalable and interpretable understanding of the semantic knowledge encoded within diffusion models, facilitating comprehensive analysis of latent biases and the nuanced representations these models learn. Experimental results show that our framework can uncover hidden patterns and associations in various domains, offering new insights into the interpretability of diffusion model latent spaces.
Authors: Vishesh Prasad, Brian Kim, Nickvash Kani
Abstract: Recent advances in natural language processing (NLP), particularly with the emergence of large language models (LLMs), have significantly enhanced the field of textual analysis. However, while these developments have yielded substantial progress in analyzing textual data, applying analysis to mathematical equations and their relationships within texts has produced mixed results. In this paper, we take the initial steps toward understanding the dependency relationships between mathematical expressions in STEM articles. Our dataset, sourced from a random sampling of the arXiv corpus, contains an analysis of 107 published STEM manuscripts whose inter-equation dependency relationships have been hand-labeled, resulting in a new object we refer to as a derivation graph that summarizes the mathematical content of the manuscript. We exhaustively evaluate analytical and NLP-based models to assess their capability to identify and extract the derivation relationships for each article and compare the results with the ground truth. Our comprehensive testing finds that both analytical and NLP models (including LLMs) achieve $\sim$40-50% F1 scores for extracting derivation graphs from articles, revealing that the recent advances in NLP have not made significant inroads in comprehending mathematical texts compared to simpler analytic models. While current approaches offer a solid foundation for extracting mathematical information, further research is necessary to improve accuracy and depth in this area.
Authors: Mikko Partio, Leila Hieta, Anniina Kokkonen
Abstract: Cloud cover plays a critical role in weather prediction and impacts several sectors, including agriculture, solar power generation, and aviation. Despite advancements in numerical weather prediction (NWP) models, forecasting total cloud cover remains challenging due to the small-scale nature of cloud formation processes. In this study, we introduce CloudCast, a convolutional neural network (CNN) based on the U-Net architecture, designed to predict total cloud cover (TCC) up to five hours ahead. Trained on five years of satellite data, CloudCast significantly outperforms traditional NWP models and optical flow methods. Compared to a reference NWP model, CloudCast achieves a 24% lower mean absolute error and reduces multi-category prediction errors by 46%. The model demonstrates strong performance, particularly in capturing the large-scale structure of cloud cover in the first few forecast hours, though later predictions are subject to blurring and underestimation of cloud formation. An ablation study identified the optimal input features and loss functions, with MAE-based models performing the best. CloudCast has been integrated into the Finnish Meteorological Institute's operational nowcasting system, where it improves cloud cover forecasts used by public and private sector clients. While CloudCast is limited by a relatively short skillful lead time of about three hours, future work aims to extend this through more complex network architectures and higher-resolution data. CloudCast code is available at https://github.com/fmidev/cloudcast.
Authors: Yue Yu, Prayag Tiwari
Abstract: Large Language Models (LLMs), such as ChatGPT, Phi3 and Llama-3, are leading a significant leap in AI, as they can generalize knowledge from their training to new tasks without fine-tuning. However, their application in the financial domain remains relatively limited. The financial field is inherently complex, requiring a deep understanding across various perspectives, from macro, micro economic trend to quantitative analysis. Motivated by this complexity, a mixture of expert LLMs tailored to specific financial domains could offer a more comprehensive understanding for intricate financial tasks. In this paper, we present the FinTeamExperts, a role-specialized LLM framework structured as a Mixture of Experts (MOEs) for financial analysis. The framework simulates a collaborative team setting by training each model to specialize in distinct roles: Macro Analysts, Micro analysts, and Quantitative Analysts. This role-specific specialization enhances the model's ability to integrate their domain-specific expertise. We achieve this by training three 8-billion parameter models on different corpus, each dedicated to excelling in specific finance-related roles. We then instruct-tune FinTeamExperts on downstream tasks to align with practical financial tasks. The experimental results show that FinTeamExperts outperform all models of the same size and larger on three out of four datasets. On the fourth dataset, which presents a more complex task, FinTeamExperts still surpass all models of the same size. This highlights the success of our role-based specialization approach and the continued training approach for FinTeamExperts.
Authors: Dong Yao, Caizhi Tang, Qing Cui, Longfei Li
Abstract: Data from observational studies (OSs) is widely available and readily obtainable yet frequently contains confounding biases. On the other hand, data derived from randomized controlled trials (RCTs) helps to reduce these biases; however, it is expensive to gather, resulting in a tiny size of randomized data. For this reason, effectively fusing observational data and randomized data to better estimate heterogeneous treatment effects (HTEs) has gained increasing attention. However, existing methods for integrating observational data with randomized data must require \textit{complete} observational data, meaning that both treated subjects and untreated subjects must be included in OSs. This prerequisite confines the applicability of such methods to very specific situations, given that including all subjects, whether treated or untreated, in observational studies is not consistently achievable. In our paper, we propose a resilient approach to \textbf{C}ombine \textbf{I}ncomplete \textbf{O}bservational data and randomized data for HTE estimation, which we abbreviate as \textbf{CIO}. The CIO is capable of estimating HTEs efficiently regardless of the completeness of the observational data, be it full or partial. Concretely, a confounding bias function is first derived using the pseudo-experimental group from OSs, in conjunction with the pseudo-control group from RCTs, via an effect estimation procedure. This function is subsequently utilized as a corrective residual to rectify the observed outcomes of observational data during the HTE estimation by combining the available observational data and the all randomized data. To validate our approach, we have conducted experiments on a synthetic dataset and two semi-synthetic datasets.
Authors: Zehui Li, Yuhao Ni, Guoxuan Xia, William Beardall, Akashaditya Das, Guy-Bart Stan, Yiren Zhao
Abstract: Abstract Recent advances in immunology and synthetic biology have accelerated the development of deep generative methods for DNA sequence design. Two dominant approaches in this field are AutoRegressive (AR) models and Diffusion Models (DMs). However, genomic sequences are functionally heterogeneous, consisting of multiple connected regions (e.g., Promoter Regions, Exons, and Introns) where elements within each region come from the same probability distribution, but the overall sequence is non-homogeneous. This heterogeneous nature presents challenges for a single model to accurately generate genomic sequences. In this paper, we analyze the properties of AR models and DMs in heterogeneous genomic sequence generation, pointing out crucial limitations in both methods: (i) AR models capture the underlying distribution of data by factorizing and learning the transition probability but fail to capture the global property of DNA sequences. (ii) DMs learn to recover the global distribution but tend to produce errors at the base pair level. To overcome the limitations of both approaches, we propose a post-training sampling method, termed Absorb & Escape (A&E) to perform compositional generation from AR models and DMs. This approach starts with samples generated by DMs and refines the sample quality using an AR model through the alternation of the Absorb and Escape steps. To assess the quality of generated sequences, we conduct extensive experiments on 15 species for conditional and unconditional DNA generation. The experiment results from motif distribution, diversity checks, and genome integration tests unequivocally show that A&E outperforms state-of-the-art AR models and DMs in genomic sequence generation.
Authors: Minkai Xu, Tomas Geffner, Karsten Kreis, Weili Nie, Yilun Xu, Jure Leskovec, Stefano Ermon, Arash Vahdat
Abstract: Despite remarkable progress in autoregressive language models, alternative generative paradigms beyond left-to-right generation are still being actively explored. Discrete diffusion models, with the capacity for parallel generation, have recently emerged as a promising alternative. Unfortunately, these models still underperform the autoregressive counterparts, with the performance gap increasing when reducing the number of sampling steps. Our analysis reveals that this degradation is a consequence of an imperfect approximation used by diffusion models. In this work, we propose Energy-based Diffusion Language Model (EDLM), an energy-based model operating at the full sequence level for each diffusion step, introduced to improve the underlying approximation used by diffusion models. More specifically, we introduce an EBM in a residual form, and show that its parameters can be obtained by leveraging a pretrained autoregressive model or by finetuning a bidirectional transformer via noise contrastive estimation. We also propose an efficient generation algorithm via parallel important sampling. Comprehensive experiments on language modeling benchmarks show that our model can consistently outperform state-of-the-art diffusion models by a significant margin, and approaches autoregressive models' perplexity. We further show that, without any generation performance drop, our framework offers a 1.3$\times$ sampling speedup over existing diffusion models.
Authors: Ji Ma
Abstract: As Large Language Model (LLM)-based agents increasingly undertake real-world tasks and engage with human society, how well do we understand their behaviors? This study (1) investigates how LLM agents' prosocial behaviors -- a fundamental social norm -- can be induced by different personas and benchmarked against human behaviors; and (2) introduces a behavioral approach to evaluate the performance of LLM agents in complex decision-making scenarios. We explored how different personas and experimental framings affect these AI agents' altruistic behavior in dictator games and compared their behaviors within the same LLM family, across various families, and with human behaviors. Our findings reveal substantial variations and inconsistencies among LLMs and notable differences compared to human behaviors. Merely assigning a human-like identity to LLMs does not produce human-like behaviors. Despite being trained on extensive human-generated data, these AI agents cannot accurately predict human decisions. LLM agents are not able to capture the internal processes of human decision-making, and their alignment with human behavior is highly variable and dependent on specific model architectures and prompt formulations; even worse, such dependence does not follow a clear pattern.
Authors: Mohammad Fahes, Tuan-Hung Vu, Andrei Bursuc, Patrick P\'erez, Raoul de Charette
Abstract: Domain adaptation has been extensively investigated in computer vision but still requires access to target data at the training time, which might be difficult to obtain in some uncommon conditions. In this paper, we present a new framework for domain adaptation relying on a single Vision-Language (VL) latent embedding instead of full target data. First, leveraging a contrastive language-image pre-training model (CLIP), we propose prompt/photo-driven instance normalization (PIN). PIN is a feature augmentation method that mines multiple visual styles using a single target VL latent embedding, by optimizing affine transformations of low-level source features. The VL embedding can come from a language prompt describing the target domain, a partially optimized language prompt, or a single unlabeled target image. Second, we show that these mined styles (i.e., augmentations) can be used for zero-shot (i.e., target-free) and one-shot unsupervised domain adaptation. Experiments on semantic segmentation demonstrate the effectiveness of the proposed method, which outperforms relevant baselines in the zero-shot and one-shot settings.
Authors: Aseem Paranjape (IUCAA), Ravi K. Sheth (UPenn/ICTP)
Abstract: We consider approximating the linearly evolved 2-point correlation function (2pcf) of dark matter $\xi_{\rm lin}(r;\boldsymbol{\theta})$ in a cosmological model with parameters $\boldsymbol{\theta}$ as the linear combination $\xi_{\rm lin}(r;\boldsymbol{\theta})\approx\sum_i\,b_i(r)\,w_i(\boldsymbol{\theta})$, where the functions $\mathcal{B}=\{b_i(r)\}$ form a $\textit{model-agnostic basis}$ for the linear 2pcf. This decomposition is important for model-agnostic analyses of the baryon acoustic oscillation (BAO) feature in the nonlinear 2pcf of galaxies that fix $\mathcal{B}$ and leave the coefficients $\{w_i\}$ free. To date, such analyses have made simple but sub-optimal choices for $\mathcal{B}$, such as monomials. We develop a machine learning framework for systematically discovering a $\textit{minimal}$ basis $\mathcal{B}$ that describes $\xi_{\rm lin}(r)$ near the BAO feature in a wide class of cosmological models. We use a custom architecture, denoted $\texttt{BiSequential}$, for a neural network (NN) that explicitly realizes the separation between $r$ and $\boldsymbol{\theta}$ above. The optimal NN trained on data in which only $\{\Omega_{\rm m},h\}$ are varied in a $\textit{flat}$ $\Lambda$CDM model produces a basis $\mathcal{B}$ comprising $9$ functions capable of describing $\xi_{\rm lin}(r)$ to $\sim0.6\%$ accuracy in $\textit{curved}$ $w$CDM models varying 7 parameters within $\sim5\%$ of their fiducial, flat $\Lambda$CDM values. Scales such as the peak, linear point and zero-crossing of $\xi_{\rm lin}(r)$ are also recovered with very high accuracy. We compare our approach to other compression schemes in the literature, and speculate that $\mathcal{B}$ may also encompass $\xi_{\rm lin}(r)$ in modified gravity models near our fiducial $\Lambda$CDM model. Using our basis functions in model-agnostic BAO analyses can potentially lead to significant statistical gains.
Authors: Henrik Madsen, Gudmund Grov, Federico Mancini, Magnus Baksaas, {\AA}vald {\AA}slaugson Sommervoll
Abstract: Unmanned vehicles able to conduct advanced operations without human intervention are being developed at a fast pace for many purposes. Not surprisingly, they are also expected to significantly change how military operations can be conducted. To leverage the potential of this new technology in a physically and logically contested environment, security risks are to be assessed and managed accordingly. Research on this topic points to autonomous cyber defence as one of the capabilities that may be needed to accelerate the adoption of these vehicles for military purposes. Here, we pursue this line of investigation by exploring reinforcement learning to train an agent that can autonomously respond to cyber attacks on unmanned vehicles in the context of a military operation. We first developed a simple simulation environment to quickly prototype and test some proof-of-concept agents for an initial evaluation. This agent was then applied to a more realistic simulation environment and finally deployed on an actual unmanned ground vehicle for even more realism. A key contribution of our work is demonstrating that reinforcement learning is a viable approach to train an agent that can be used for autonomous cyber defence on a real unmanned ground vehicle, even when trained in a simple simulation environment.
Authors: He Jiang, Yutong Wang, Rishi Veerapaneni, Tanishq Duhan, Guillaume Sartoretti, Jiaoyang Li
Abstract: Lifelong Multi-Agent Path Finding (LMAPF) is a variant of MAPF where agents are continually assigned new goals, necessitating frequent re-planning to accommodate these dynamic changes. Recently, this field has embraced learning-based methods, which reactively generate single-step actions based on individual local observations. However, it is still challenging for them to match the performance of the best search-based algorithms, especially in large-scale settings. This work proposes an imitation-learning-based LMAPF solver that introduces a novel communication module and systematic single-step collision resolution and global guidance techniques. Our proposed solver, Scalable Imitation Learning for LMAPF (SILLM), inherits the fast reasoning speed of learning-based methods and the high solution quality of search-based methods with the help of modern GPUs. Across six large-scale maps with up to 10,000 agents and varying obstacle structures, SILLM surpasses the best learning- and search-based baselines, achieving average throughput improvements of 137.7% and 16.0%, respectively. Furthermore, SILLM also beats the winning solution of the 2023 League of Robot Runners, an international LMAPF competition sponsored by Amazon Robotics. Finally, we validated SILLM with 10 real robots and 100 virtual robots in a mockup warehouse environment.
Authors: Chris Cama\~no, Daniel Huang
Abstract: We introduce Soft Kernel Interpolation (SoftKI) designed for scalable Gaussian Process (GP) regression on high-dimensional datasets. Inspired by Structured Kernel Interpolation (SKI), which approximates a GP kernel via interpolation from a structured lattice, SoftKI approximates a kernel via softmax interpolation from a smaller number of learned interpolation (i.e, inducing) points. By abandoning the lattice structure used in SKI-based methods, SoftKI separates the cost of forming an approximate GP kernel from the dimensionality of the data, making it well-suited for high-dimensional datasets. We demonstrate the effectiveness of SoftKI across various examples, and demonstrate that its accuracy exceeds that of other scalable GP methods when the data dimensionality is modest (around $10$).
Authors: Zhichao Wang, Bin Bi, Zixu Zhu, Xiangbo Mao, Jun Wang, Shiyu Wang
Abstract: By pretraining on trillions of tokens, an LLM gains the capability of text generation. However, to enhance its utility and reduce potential harm, SFT and alignment are applied sequentially to the pretrained model. Due to the differing nature and objective functions of SFT and alignment, catastrophic forgetting has become a significant issue. To address this, we introduce Unified Fine-Tuning (UFT), which integrates SFT and alignment into a single training stage using the same objective and loss functions through an implicit reward function. Our experimental results demonstrate that UFT outperforms SFT on instruction-tuning data alone. Moreover, when combining instruction-tuning data with alignment data, UFT effectively prevents catastrophic forgetting across these two stages and shows a clear advantage over sequentially applying SFT and alignment. This is evident in the significant improvements observed in the \textbf{ifeval} task for instruction-following and the \textbf{truthful-qa} task for factuality. The proposed general fine-tuning framework UFT establishes an effective and efficient pretraining-UFT paradigm for LLM training.
Authors: Adonisz Dimitriu, Tam\'as Michaletzky, Viktor Remeli
Abstract: Adversarial attacks threaten the reliability of machine learning models in critical applications like autonomous vehicles and defense systems. As object detectors become more robust with models like YOLOv8, developing effective adversarial methodologies is increasingly challenging. We present Truck Adversarial Camouflage Optimization (TACO), a novel framework that generates adversarial camouflage patterns on 3D vehicle models to deceive state-of-the-art object detectors. Adopting Unreal Engine 5, TACO integrates differentiable rendering with a Photorealistic Rendering Network to optimize adversarial textures targeted at YOLOv8. To ensure the generated textures are both effective in deceiving detectors and visually plausible, we introduce the Convolutional Smooth Loss function, a generalized smooth loss function. Experimental evaluations demonstrate that TACO significantly degrades YOLOv8's detection performance, achieving an AP@0.5 of 0.0099 on unseen test data. Furthermore, these adversarial patterns exhibit strong transferability to other object detection models such as Faster R-CNN and earlier YOLO versions.
Authors: Timothy D. Gebhard, Jonas Wildberger, Maximilian Dax, Annalena Kofler, Daniel Angerhausen, Sascha P. Quanz, Bernhard Sch\"olkopf
Abstract: Inferring atmospheric properties of exoplanets from observed spectra is key to understanding their formation, evolution, and habitability. Since traditional Bayesian approaches to atmospheric retrieval (e.g., nested sampling) are computationally expensive, a growing number of machine learning (ML) methods such as neural posterior estimation (NPE) have been proposed. We seek to make ML-based atmospheric retrieval (1) more reliable and accurate with verified results, and (2) more flexible with respect to the underlying neural networks and the choice of the assumed noise models. First, we adopt flow matching posterior estimation (FMPE) as a new ML approach to atmospheric retrieval. FMPE maintains many advantages of NPE, but provides greater architectural flexibility and scalability. Second, we use importance sampling (IS) to verify and correct ML results, and to compute an estimate of the Bayesian evidence. Third, we condition our ML models on the assumed noise level of a spectrum (i.e., error bars), thus making them adaptable to different noise models. Both our noise level-conditional FMPE and NPE models perform on par with nested sampling across a range of noise levels when tested on simulated data. FMPE trains about 3 times faster than NPE and yields higher IS efficiencies. IS successfully corrects inaccurate ML results, identifies model failures via low efficiencies, and provides accurate estimates of the Bayesian evidence. FMPE is a powerful alternative to NPE for fast, amortized, and parallelizable atmospheric retrieval. IS can verify results, thus helping to build confidence in ML-based approaches, while also facilitating model comparison via the evidence ratio. Noise level conditioning allows design studies for future instruments to be scaled up, for example, in terms of the range of signal-to-noise ratios.
Authors: Iftach Arbel, Yehonathan Refael, Ofir Lindenbaum
Abstract: Large Language Models (LLMs) have shown promise in highly-specialized domains, however challenges are still present in aspects of accuracy and costs. These limitations restrict the usage of existing models in domain-specific tasks. While fine-tuning pre-trained models have shown promising results, this process can be computationally expensive and require massive datasets of the specialized application in hand. In this work, we bridge that gap. We have developed Phi-2-Legal and Mistral-Legal-7B, which are language models specifically designed for legal applications. These models are based on Phi-2 and Mistral-7B-v0.1, and have gone through continued pre-training with over 500 million tokens of legal texts. Our innovative approach significantly improves capabilities in legal tasks by using Large Language Models (LLMs) to convert raw training data into reading comprehension text. Our legal LLMs have demonstrated superior performance in legal benchmarks, even outperforming models trained on much larger datasets with more resources. This work emphasizes the effectiveness of continued pre-training on domain-specific texts, while using affordable LLMs for data conversion, which gives these models domain expertise while retaining general language understanding capabilities. While this work uses the legal domain as a test case, our method can be scaled and applied to any pre-training dataset, resulting in significant improvements across different tasks. These findings underscore the potential of domain-adaptive pre-training and reading comprehension for the development of highly effective domain-specific language models.
Authors: Vu-Anh Le, Mehmet Dik
Abstract: Neural operators have emerged as transformative tools for learning mappings between infinite-dimensional function spaces, offering useful applications in solving complex partial differential equations (PDEs). This paper presents a rigorous mathematical framework for analyzing the behaviors of neural operators, with a focus on their stability, convergence, clustering dynamics, universality, and generalization error. By proposing a list of novel theorems, we provide stability bounds in Sobolev spaces and demonstrate clustering in function space via gradient flow interpretation, guiding neural operator design and optimization. Based on these theoretical gurantees, we aim to offer clear and unified guidance in a single setting for the future design of neural operator-based methods.
Authors: Yuening Wang, Man Chen, Yaochen Hu, Wei Guo, Yingxue Zhang, Huifeng Guo, Yong Liu, Mark Coates
Abstract: Many platforms, such as e-commerce websites, offer both search and recommendation services simultaneously to better meet users' diverse needs. Recommendation services suggest items based on user preferences, while search services allow users to search for items before providing recommendations. Since users and items are often shared between the search and recommendation domains, there is a valuable opportunity to enhance the recommendation domain by leveraging user preferences extracted from the search domain. Existing approaches either overlook the shift in user intention between these domains or fail to capture the significant impact of learning from users' search queries on understanding their interests. In this paper, we propose a framework that learns from user search query embeddings within the context of user preferences in the recommendation domain. Specifically, user search query sequences from the search domain are used to predict the items users will click at the next time point in the recommendation domain. Additionally, the relationship between queries and items is explored through contrastive learning. To address issues of data sparsity, the diffusion model is incorporated to infer positive items the user will select after searching with certain queries in a denoising manner, which is particularly effective in preventing false positives. Effectively extracting this information, the queries are integrated into click-through rate prediction in the recommendation domain. Experimental analysis demonstrates that our model outperforms state-of-the-art models in the recommendation domain.
Authors: Lijie Hu, Songning Lai, Wenshuo Chen, Hongru Xiao, Hongbin Lin, Lu Yu, Jingfeng Zhang, Di Wang
Abstract: The lack of interpretability in the field of medical image analysis has significant ethical and legal implications. Existing interpretable methods in this domain encounter several challenges, including dependency on specific models, difficulties in understanding and visualization, as well as issues related to efficiency. To address these limitations, we propose a novel framework called Med-MICN (Medical Multi-dimensional Interpretable Concept Network). Med-MICN provides interpretability alignment for various angles, including neural symbolic reasoning, concept semantics, and saliency maps, which are superior to current interpretable methods. Its advantages include high prediction accuracy, interpretability across multiple dimensions, and automation through an end-to-end concept labeling process that reduces the need for extensive human training effort when working with new datasets. To demonstrate the effectiveness and interpretability of Med-MICN, we apply it to four benchmark datasets and compare it with baselines. The results clearly demonstrate the superior performance and interpretability of our Med-MICN.
Authors: Jamal Al-Nabulsi, Mohammad Al-Sayed Ahmad, Baraa Hasaneiah, Fayhaa AlZoubi
Abstract: Diagnosing knee osteoarthritis (OA) early is crucial for managing symptoms and preventing further joint damage, ultimately improving patient outcomes and quality of life. In this paper, a bioimpedance-based diagnostic tool that combines precise hardware and deep learning for effective non-invasive diagnosis is proposed. system features a relay-based circuit and strategically placed electrodes to capture comprehensive bioimpedance data. The data is processed by a neural network model, which has been optimized using convolutional layers, dropout regularization, and the Adam optimizer. This approach achieves a 98% test accuracy, making it a promising tool for detecting knee osteoarthritis musculoskeletal disorders.
Authors: Sophia M Rein, Jing Li, Miguel Hernan, Andrew Beam
Abstract: The g-formula can be used to estimate causal effects of sustained treatment strategies using observational data under the identifying assumptions of consistency, positivity, and exchangeability. The non-iterative conditional expectation (NICE) estimator of the g-formula also requires correct estimation of the conditional distribution of the time-varying treatment, confounders, and outcome. Parametric models, which have been traditionally used for this purpose, are subject to model misspecification, which may result in biased causal estimates. Here, we propose a unified deep learning framework for the NICE g-formula estimator that uses multitask recurrent neural networks for estimation of the joint conditional distributions. Using simulated data, we evaluated our model's bias and compared it with that of the parametric g-formula estimator. We found lower bias in the estimates of the causal effect of sustained treatment strategies on a survival outcome when using the deep learning estimator compared with the parametric NICE estimator in settings with simple and complex temporal dependencies between covariates. These findings suggest that our Deep Learning g-formula estimator may be less sensitive to model misspecification than the classical parametric NICE estimator when estimating the causal effect of sustained treatment strategies from complex observational data.
Authors: Noel Elias, Homa Esfahanizadeh, Kaan Kale, Sriram Vishwanath, Muriel Medard
Abstract: Large language models have drastically changed the prospects of AI by introducing technologies for more complex natural language processing. However, current methodologies to train such LLMs require extensive resources including but not limited to large amounts of data, expensive machinery, and lengthy training. To solve this problem, this paper proposes a new tokenization method inspired by universal Lempel-Ziv-Welch data compression that compresses repetitive phrases into multi-word tokens. With MultiTok as a new tokenizing tool, we show that language models are able to be trained notably more efficiently while offering a similar accuracy on more succinct and compressed training data. In fact, our results demonstrate that MultiTok achieves a comparable performance to the BERT standard as a tokenizer while also providing close to 2.5x faster training with more than 30% less training data.
Authors: Noel Elias
Abstract: Sonar based audio classification techniques are a growing area of research in the field of underwater acoustics. Usually, underwater noise picked up by passive sonar transducers contains all types of signals that travel through the ocean and is transformed into spectrographic images. As a result, the corresponding spectrograms intended to display the temporal-frequency data of a certain object often include the tonal regions of abundant extraneous noise that can effectively interfere with a 'contact'. So, a majority of spectrographic samples extracted from underwater audio signals are rendered unusable due to their clutter and lack the required indistinguishability between different objects. With limited clean true data for supervised training, creating classification models for these audio signals is severely bottlenecked. This paper derives several new techniques to combat this problem by developing a novel Score-CAM based denoiser to extract an object's signature from noisy spectrographic data without being given any ground truth data. In particular, this paper proposes a novel generative adversarial network architecture for learning and producing spectrographic training data in similar distributions to low-feature spectrogram inputs. In addition, this paper also a generalizable class activation mapping based denoiser for different distributions of acoustic data, even real-world data distributions. Utilizing these novel architectures and proposed denoising techniques, these experiments demonstrate state-of-the-art noise reduction accuracy and improved classification accuracy than current audio classification standards. As such, this approach has applications not only to audio data but for countless data distributions used all around the world for machine learning.
Authors: Noel Elias
Abstract: Modern day audio signal classification techniques lack the ability to classify low feature audio signals in the form of spectrographic temporal frequency data representations. Additionally, currently utilized techniques rely on full diverse data sets that are often not representative of real-world distributions. This paper derives several first-of-its-kind machine learning methodologies to analyze these low feature audio spectrograms given data distributions that may have normalized, skewed, or even limited training sets. In particular, this paper proposes several novel customized convolutional architectures to extract identifying features using binary, one-class, and siamese approaches to identify the spectrographic signature of a given audio signal. Utilizing these novel convolutional architectures as well as the proposed classification methods, these experiments demonstrate state-of-the-art classification accuracy and improved efficiency than traditional audio classification methods.
Authors: Chirag Modi
Abstract: Hamiltonian Monte-Carlo (HMC) and its auto-tuned variant, the No U-Turn Sampler (NUTS) can struggle to accurately sample distributions with complex geometries, e.g., varying curvature, due to their constant step size for leapfrog integration and fixed mass matrix. In this work, we develop a strategy to locally adapt the step size parameter of HMC at every iteration by evaluating a low-rank approximation of the local Hessian and estimating its largest eigenvalue. We combine it with a strategy to similarly adapt the trajectory length by monitoring the no U-turn condition, resulting in an adaptive sampler, ATLAS: adapting trajectory length and step-size. We further use a delayed rejection framework for making multiple proposals that improves the computational efficiency of ATLAS, and develop an approach for automatically tuning its hyperparameters during warmup. We compare ATLAS with state-of-the-art samplers like NUTS on a suite of synthetic and real world examples, and show that i) unlike NUTS, ATLAS is able to accurately sample difficult distributions with complex geometries, ii) it is computationally competitive to NUTS for simpler distributions, and iii) it is more robust to the tuning of hyperparamters.
Authors: David Yunis, Siyu Huo, Chulaka Gunasekara, Danish Contractor
Abstract: Language models are now deployed in a wide variety of user-facing applications, often for specific purposes like answering questions about documentation or acting as coding assistants. As these models are intended for particular purposes, they should not be able to answer irrelevant queries like requests for poetry or questions about physics, or even worse, queries that can only be answered by humans like sensitive company policies. Instead we would like them to only answer queries corresponding to desired behavior and refuse all other requests, which we refer to as scoping. We find that, despite the use of system prompts, two representative language models can be poorly scoped and respond to queries they should not be addressing. We then conduct a comprehensive empirical evaluation of methods which could be used for scoping the behavior of language models. Among many other results, we show that a recently-proposed method for general alignment, Circuit Breakers (CB), can be adapted to scope language models to very specific tasks like sentiment analysis or summarization or even tasks with finer-grained scoping (e.g. summarizing only news articles). When compared to standard methods like fine-tuning or preference learning, CB is more robust both for out of distribution tasks, and to adversarial prompting techniques. We also show that layering SFT and CB together often results in the best of both worlds: improved performance only on relevant queries, while rejecting irrelevant ones.
Authors: Yamin Arefeen, Brett Levac, Jonathan I. Tamir
Abstract: Neonatal Magnetic Resonance Imaging (MRI) enables non-invasive assessment of potential brain abnormalities during the critical phase of early life development. Recently, interest has developed in lower field (i.e., below 1.5 Tesla) MRI systems that trade-off magnetic field strength for portability and access in the neonatal intensive care unit (NICU). Unfortunately, lower-field neonatal MRI still suffers from long scan times and motion artifacts that can limit its clinical utility for neonates. This work improves motion robustness and accelerates lower field neonatal MRI through diffusion-based generative modeling and signal processing based motion modeling. We first gather a training dataset of clinical neonatal MRI images. Then we train a diffusion-based generative model to learn the statistical distribution of fully-sampled images by applying several signal processing methods to handle the lower signal-to-noise ratio and lower quality of our MRI images. Finally, we present experiments demonstrating the utility of our generative model to improve reconstruction performance across two tasks: accelerated MRI and motion correction.
Authors: Jian Qian, Alexander Rakhlin, Nikita Zhivotovskiy
Abstract: We revisit the sequential variants of linear regression with the squared loss, classification problems with hinge loss, and logistic regression, all characterized by unbounded losses in the setup where no assumptions are made on the magnitude of design vectors and the norm of the optimal vector of parameters. The key distinction from existing results lies in our assumption that the set of design vectors is known in advance (though their order is not), a setup sometimes referred to as transductive online learning. While this assumption seems similar to fixed design regression or denoising, we demonstrate that the sequential nature of our algorithms allows us to convert our bounds into statistical ones with random design without making any additional assumptions about the distribution of the design vectors--an impossibility for standard denoising results. Our key tools are based on the exponential weights algorithm with carefully chosen transductive (design-dependent) priors, which exploit the full horizon of the design vectors. Our classification regret bounds have a feature that is only attributed to bounded losses in the literature: they depend solely on the dimension of the parameter space and on the number of rounds, independent of the design vectors or the norm of the optimal solution. For linear regression with squared loss, we further extend our analysis to the sparse case, providing sparsity regret bounds that additionally depend on the magnitude of the response variables. We argue that these improved bounds are specific to the transductive setting and unattainable in the worst-case sequential setup. Our algorithms, in several cases, have polynomial time approximations and reduce to sampling with respect to log-concave measures instead of aggregating over hard-to-construct $\varepsilon$-covers of classes.
Authors: Andrew Zhao
Abstract: We study the problem of learning an unknown quantum many-body Hamiltonian $H$ from black-box queries to its time evolution $e^{-\mathrm{i} H t}$. Prior proposals for solving this task either impose some assumptions on $H$, such as its interaction structure or locality, or otherwise use an exponential amount of computational postprocessing. In this paper, we present efficient algorithms to learn any $n$-qubit Hamiltonian, assuming only a bound on the number of Hamiltonian terms, $m \leq \mathrm{poly}(n)$. Our algorithms do not need to know the terms in advance, nor are they restricted to local interactions. We consider two models of control over the time evolution: the first has access to time reversal ($t < 0$), enabling an algorithm that outputs an $\epsilon$-accurate classical description of $H$ after querying its dynamics for a total of $\widetilde{O}(m/\epsilon)$ evolution time. The second access model is more conventional, allowing only forward-time evolutions; our algorithm requires $\widetilde{O}(\|H\|^3/\epsilon^4)$ evolution time in this setting. Central to our results is the recently introduced concept of a pseudo-Choi state of $H$. We extend the utility of this learning resource by showing how to use it to learn the Fourier spectrum of $H$, how to achieve nearly Heisenberg-limited scaling with it, and how to prepare it even under our more restricted access models.
Authors: Rafael Rivera Soto, Barry Chen, Nicholas Andrews
Abstract: Large language models can produce highly fluent paraphrases while retaining much of the original meaning. While this capability has a variety of helpful applications, it may also be abused by bad actors, for example to plagiarize content or to conceal their identity. This motivates us to consider the problem of paraphrase inversion: given a paraphrased document, attempt to recover the original text. To explore the feasibility of this task, we fine-tune paraphrase inversion models, both with and without additional author-specific context to help guide the inversion process. We explore two approaches to author-specific inversion: one using in-context examples of the target author's writing, and another using learned style representations that capture distinctive features of the author's style. We show that, when starting from paraphrased machine-generated text, we can recover significant portions of the document using a learned inversion model. When starting from human-written text, the variety of source writing styles poses a greater challenge for invertability. However, even when the original tokens can't be recovered, we find the inverted text is stylistically similar to the original, which significantly improves the performance of plagiarism detectors and authorship identification systems that rely on stylistic markers.
Authors: Hira Saleem, Flora Salim, Cormac Purcell
Abstract: Climate models serve as critical tools for evaluating the effects of climate change and projecting future climate scenarios. However, the reliance on numerical simulations of physical equations renders them computationally intensive and inefficient. While deep learning methodologies have made significant progress in weather forecasting, they are still unstable for climate emulation tasks. Here, we propose PACE, a lightweight 684K parameter Physics Informed Uncertainty Aware Climate Emulator. PACE emulates temperature and precipitation stably for 86 years while only being trained on greenhouse gas emissions data. We incorporate a fundamental physical law of advection-diffusion in PACE accounting for boundary conditions and empirically estimating the diffusion co-efficient and flow velocities from emissions data. PACE has been trained on 15 climate models provided by ClimateSet outperforming baselines across most of the climate models and advancing a new state of the art in a climate diagnostic task.
Authors: Jiaqi Han, Mingjian Jiang, Yuxuan Song, Jure Leskovec, Stefano Ermon, Minkai Xu
Abstract: Preference optimization has made significant progress recently, with numerous methods developed to align language models with human preferences. This paper introduces $f$-divergence Preference Optimization ($f$-PO), a novel framework that generalizes and extends existing approaches. $f$-PO minimizes $f$-divergences between the optimized policy and the optimal policy, encompassing a broad family of alignment methods using various divergences. Our approach unifies previous algorithms like DPO and EXO, while offering new variants through different choices of $f$-divergences. We provide theoretical analysis of $f$-PO's properties and conduct extensive experiments on state-of-the-art language models using benchmark datasets. Results demonstrate $f$-PO's effectiveness across various tasks, achieving superior performance compared to existing methods on popular benchmarks such as AlpacaEval 2, Arena-Hard, and MT-Bench. Additionally, we present ablation studies exploring the impact of different $f$-divergences, offering insights into the trade-offs between regularization and performance in offline preference optimization. Our work contributes both practical algorithms and theoretical understanding to the field of language model alignment. Code is available at https://github.com/MinkaiXu/fPO.
Authors: Apostolos Kokolis, Michael Kuchnik, John Hoffman, Adithya Kumar, Parth Malani, Faye Ma, Zachary DeVito, Shubho Sengupta, Kalyan Saladi, Carole-Jean Wu
Abstract: Reliability is a fundamental challenge in operating large-scale machine learning (ML) infrastructures, particularly as the scale of ML models and training clusters continues to grow. Despite decades of research on infrastructure failures, the impact of job failures across different scales remains unclear. This paper presents a view of managing two large, multi-tenant ML clusters, providing quantitative analysis, operational experience, and our own perspective in understanding and addressing reliability concerns at scale. Our analysis reveals that while large jobs are most vulnerable to failures, smaller jobs make up the majority of jobs in the clusters and should be incorporated into optimization objectives. We identify key workload properties, compare them across clusters, and demonstrate essential reliability requirements for pushing the boundaries of ML training at scale. We hereby introduce a taxonomy of failures and key reliability metrics, analyze 11 months of data from two state-of-the-art ML environments with over 150 million A100 GPU hours and 4 million jobs. Building on our data, we fit a failure model to project Mean Time to Failure for various GPU scales. We further propose a method to estimate a related metric, Effective Training Time Ratio, as a function of job parameters, and we use this model to gauge the efficacy of potential software mitigations at scale. Our work provides valuable insights and future research directions for improving the reliability of AI supercomputer clusters, emphasizing the need for flexible, workload-agnostic, and reliability-aware infrastructure, system software, and algorithms.
Authors: Zhihao Liu, Chenhui Hu
Abstract: As large language models (LLMs) rapidly evolve, they bring significant conveniences to our work and daily lives, but also introduce considerable safety risks. These models can generate texts with social biases or unethical content, and under specific adversarial instructions, may even incite illegal activities. Therefore, rigorous safety assessments of LLMs are crucial. In this work, we introduce a safety assessment benchmark, CFSafety, which integrates 5 classic safety scenarios and 5 types of instruction attacks, totaling 10 categories of safety questions, to form a test set with 25k prompts. This test set was used to evaluate the natural language generation (NLG) capabilities of LLMs, employing a combination of simple moral judgment and a 1-5 safety rating scale for scoring. Using this benchmark, we tested eight popular LLMs, including the GPT series. The results indicate that while GPT-4 demonstrated superior safety performance, the safety effectiveness of LLMs, including this model, still requires improvement. The data and code associated with this study are available on GitHub.
Authors: Julia Nakhleh, Joseph Shenouda, Robert D. Nowak
Abstract: This paper studies the properties of solutions to multi-task shallow ReLU neural network learning problems, wherein the network is trained to fit a dataset with minimal sum of squared weights. Remarkably, the solutions learned for each individual task resemble those obtained by solving a kernel method, revealing a novel connection between neural networks and kernel methods. It is known that single-task neural network training problems are equivalent to minimum norm interpolation problem in a non-Hilbertian Banach space, and that the solutions of such problems are generally non-unique. In contrast, we prove that the solutions to univariate-input, multi-task neural network interpolation problems are almost always unique, and coincide with the solution to a minimum-norm interpolation problem in a Sobolev (Reproducing Kernel) Hilbert Space. We also demonstrate a similar phenomenon in the multivariate-input case; specifically, we show that neural network learning problems with large numbers of diverse tasks are approximately equivalent to an $\ell^2$ (Hilbert space) minimization problem over a fixed kernel determined by the optimal neurons.
Authors: Pierre Alquier, William Kengne
Abstract: In a groundbreaking work, Schmidt-Hieber (2020) proved the minimax optimality of deep neural networks with ReLu activation for least-square regression estimation over a large class of functions defined by composition. In this paper, we extend these results in many directions. First, we remove the i.i.d. assumption on the observations, to allow some time dependence. The observations are assumed to be a Markov chain with a non-null pseudo-spectral gap. Then, we study a more general class of machine learning problems, which includes least-square and logistic regression as special cases. Leveraging on PAC-Bayes oracle inequalities and a version of Bernstein inequality due to Paulin (2015), we derive upper bounds on the estimation risk for a generalized Bayesian estimator. In the case of least-square regression, this bound matches (up to a logarithmic factor) the lower bound of Schmidt-Hieber (2020). We establish a similar lower bound for classification with the logistic loss, and prove that the proposed DNN estimator is optimal in the minimax sense.
Authors: Azim Ospanov, Farzan Farnia
Abstract: Reference-free evaluation metrics for generative models have recently been studied in the machine learning community. As a reference-free metric, the VENDI score quantifies the diversity of generative models using matrix-based entropy from information theory. The VENDI score is usually computed through the eigendecomposition of an $n \times n$ kernel matrix for $n$ generated samples. However, due to the high computational cost of eigendecomposition for large $n$, the score is often computed on sample sizes limited to a few tens of thousands. In this paper, we explore the statistical convergence of the VENDI score and demonstrate that for kernel functions with an infinite feature map dimension, the evaluated score for a limited sample size may not converge to the matrix-based entropy statistic. We introduce an alternative statistic called the $t$-truncated VENDI statistic. We show that the existing Nystr\"om method and the FKEA approximation method for the VENDI score will both converge to the defined truncated VENDI statistic given a moderate sample size. We perform several numerical experiments to illustrate the concentration of the empirical VENDI score around the truncated VENDI statistic and discuss how this statistic correlates with the visual diversity of image data.
Authors: Matheus Farias, H. T. Kung
Abstract: We introduce a novel approach to reduce the number of times required for reprogramming memristors on bit-sliced compute-in-memory crossbars for deep neural networks (DNNs). Our idea addresses the limited non-volatile memory endurance, which restrict the number of times they can be reprogrammed. To reduce reprogramming demands, we employ two techniques: (1) we organize weights into sorted sections to schedule reprogramming of similar crossbars, maximizing memristor state reuse, and (2) we reprogram only a fraction of randomly selected memristors in low-order columns, leveraging their bit-level distribution and recognizing their relatively small impact on model accuracy. We evaluate our approach for state-of-the-art models on the ImageNet-1K dataset. We demonstrate a substantial reduction in crossbar reprogramming by 3.7x for ResNet-50 and 21x for ViT-Base, while maintaining model accuracy within a 1% margin.
Authors: Anubhav Shrimal, Stanley Kanagaraj, Kriti Biswas, Swarnalatha Raghuraman, Anish Nediyanchath, Yi Zhang, Promod Yenigalla
Abstract: Large language model advancements have enabled the development of multi-agent frameworks to tackle complex, real-world problems such as to automate tasks that require interactions with diverse tools, reasoning, and human collaboration. We present MARCO, a Multi-Agent Real-time Chat Orchestration framework for automating tasks using LLMs. MARCO addresses key challenges in utilizing LLMs for complex, multi-step task execution. It incorporates robust guardrails to steer LLM behavior, validate outputs, and recover from errors that stem from inconsistent output formatting, function and parameter hallucination, and lack of domain knowledge. Through extensive experiments we demonstrate MARCO's superior performance with 94.48% and 92.74% accuracy on task execution for Digital Restaurant Service Platform conversations and Retail conversations datasets respectively along with 44.91% improved latency and 33.71% cost reduction. We also report effects of guardrails in performance gain along with comparisons of various LLM models, both open-source and proprietary. The modular and generic design of MARCO allows it to be adapted for automating tasks across domains and to execute complex usecases through multi-turn interactions.
Authors: Lina Zhu, Lin Zhou
Abstract: In multiple classification, one aims to determine whether a testing sequence is generated from the same distribution as one of the M training sequences or not. Unlike most of existing studies that focus on discrete-valued sequences with perfect distribution match, we study multiple classification for continuous sequences with distribution uncertainty, where the generating distributions of the testing and training sequences deviate even under the true hypothesis. In particular, we propose distribution free tests and prove that the error probabilities of our tests decay exponentially fast for three different test designs: fixed-length, sequential, and two-phase tests. We first consider the simple case without the null hypothesis, where the testing sequence is known to be generated from a distribution close to the generating distribution of one of the training sequences. Subsequently, we generalize our results to a more general case with the null hypothesis by allowing the testing sequence to be generated from a distribution that is vastly different from the generating distributions of all training sequences.
Authors: Damir Filipovic, Paul Schneider
Abstract: We propose a novel nonparametric kernel-based estimator of cross-sectional conditional mean and covariance matrices for large unbalanced panels. We show its consistency and provide finite-sample guarantees. In an empirical application, we estimate conditional mean and covariance matrices for a large unbalanced panel of monthly stock excess returns given macroeconomic and firm-specific covariates from 1962 to 2021.The estimator performs well with respect to statistical measures. It is informative for empirical asset pricing, generating conditional mean-variance efficient portfolios with substantial out-of-sample Sharpe ratios far beyond equal-weighted benchmarks.
Authors: Massimo Bilancia, Samuele Magro
Abstract: This paper presents a variant of the Multinomial mixture model tailored for the unsupervised classification of short text data. Traditionally, the Multinomial probability vector in this hierarchical model is assigned a Dirichlet prior distribution. Here, however, we explore an alternative prior - the Beta-Liouville distribution - which offers a more flexible correlation structure than the Dirichlet. We examine the theoretical properties of the Beta-Liouville distribution, focusing on its conjugacy with the Multinomial likelihood. This property enables the derivation of update equations for a CAVI (Coordinate Ascent Variational Inference) variational algorithm, facilitating the approximate posterior estimation of model parameters. Additionally, we propose a stochastic variant of the CAVI algorithm that enhances scalability. The paper concludes with data examples that demonstrate effective strategies for setting the Beta-Liouville hyperparameters.
Authors: Xiao Xia, Dan Zhang, Zibo Liao, Zhenyu Hou, Tianrui Sun, Jing Li, Ling Fu, Yuxiao Dong
Abstract: The modeling of industrial scenes is essential for simulations in industrial manufacturing. While large language models (LLMs) have shown significant progress in generating general 3D scenes from textual descriptions, generating industrial scenes with LLMs poses a unique challenge due to their demand for precise measurements and positioning, requiring complex planning over spatial arrangement. To address this challenge, we introduce SceneGenAgent, an LLM-based agent for generating industrial scenes through C# code. SceneGenAgent ensures precise layout planning through a structured and calculable format, layout verification, and iterative refinement to meet the quantitative requirements of industrial scenarios. Experiment results demonstrate that LLMs powered by SceneGenAgent exceed their original performance, reaching up to 81.0% success rate in real-world industrial scene generation tasks and effectively meeting most scene generation requirements. To further enhance accessibility, we construct SceneInstruct, a dataset designed for fine-tuning open-source LLMs to integrate into SceneGenAgent. Experiments show that fine-tuning open-source LLMs on SceneInstruct yields significant performance improvements, with Llama3.1-70B approaching the capabilities of GPT-4o. Our code and data are available at https://github.com/THUDM/SceneGenAgent .
Authors: Yuanyuan Wang, Biwei Huang, Wei Huang, Xi Geng, Mingming Gong
Abstract: The identifiability analysis of linear Ordinary Differential Equation (ODE) systems is a necessary prerequisite for making reliable causal inferences about these systems. While identifiability has been well studied in scenarios where the system is fully observable, the conditions for identifiability remain unexplored when latent variables interact with the system. This paper aims to address this gap by presenting a systematic analysis of identifiability in linear ODE systems incorporating hidden confounders. Specifically, we investigate two cases of such systems. In the first case, latent confounders exhibit no causal relationships, yet their evolution adheres to specific functional forms, such as polynomial functions of time $t$. Subsequently, we extend this analysis to encompass scenarios where hidden confounders exhibit causal dependencies, with the causal structure of latent variables described by a Directed Acyclic Graph (DAG). The second case represents a more intricate variation of the first case, prompting a more comprehensive identifiability analysis. Accordingly, we conduct detailed identifiability analyses of the second system under various observation conditions, including both continuous and discrete observations from single or multiple trajectories. To validate our theoretical results, we perform a series of simulations, which support and substantiate our findings.
Authors: Blanka Balogh, David Saint-Martin, Olivier Geoffroy
Abstract: In this study, we present the integration of a neural network-based parameterization into the global atmospheric model ARP-GEM1, leveraging the Python interface of the OASIS coupler. This approach facilitates the exchange of fields between the Fortran-based ARP-GEM1 model and a Python component responsible for neural network inference. As a proof-of-concept experiment, we trained a neural network to emulate the deep convection parameterization of ARP-GEM1. Using the flexible Fortran/Python interface, we have successfully replaced ARP-GEM1's deep convection scheme with a neural network emulator. To assess the performance of the neural network deep convection scheme, we have run a 5-years ARP-GEM1 simulation using the neural network emulator. The evaluation of averaged fields showed good agreement with output from an ARP-GEM1 simulation using the physics-based deep convection scheme. The Python component was deployed on a separate partition from the general circulation model, using GPUs to increase inference speed of the neural network.
Authors: Boris Wolfson, Erman Acar
Abstract: Current trends in Machine Learning prefer explainability even when it comes at the cost of performance. Therefore, explainable AI methods are particularly important in the field of Fraud Detection. This work investigates the applicability of Differentiable Inductive Logic Programming (DILP) as an explainable AI approach to Fraud Detection. Although the scalability of DILP is a well-known issue, we show that with some data curation such as cleaning and adjusting the tabular and numerical data to the expected format of background facts statements, it becomes much more applicable. While in processing it does not provide any significant advantage on rather more traditional methods such as Decision Trees, or more recent ones like Deep Symbolic Classification, it still gives comparable results. We showcase its limitations and points to improve, as well as potential use cases where it can be much more useful compared to traditional methods, such as recursive rule learning.
Authors: Timur Mamedov, Anton Konushin, Vadim Konushin
Abstract: Modern person re-identification (Re-ID) methods have a weak generalization ability and experience a major accuracy drop when capturing environments change. This is because existing multi-camera Re-ID datasets are limited in size and diversity, since such data is difficult to obtain. At the same time, enormous volumes of unlabeled single-camera records are available. Such data can be easily collected, and therefore, it is more diverse. Currently, single-camera data is used only for self-supervised pre-training of Re-ID methods. However, the diversity of single-camera data is suppressed by fine-tuning on limited multi-camera data after pre-training. In this paper, we propose ReMix, a generalized Re-ID method jointly trained on a mixture of limited labeled multi-camera and large unlabeled single-camera data. Effective training of our method is achieved through a novel data sampling strategy and new loss functions that are adapted for joint use with both types of data. Experiments show that ReMix has a high generalization ability and outperforms state-of-the-art methods in generalizable person Re-ID. To the best of our knowledge, this is the first work that explores joint training on a mixture of multi-camera and single-camera data in person Re-ID.
Authors: Chuqiao Zhang, Crina Grosan, Dalia Chakrabarty
Abstract: Patients who are undergoing physical rehabilitation, benefit from feedback that follows from reliable assessment of their cumulative performance attained at a given time. In this paper, we provide a method for the learning of the recovery trajectory of an individual patient, as they undertake exercises as part of their physical therapy towards recovery of their loss of movement ability, following a critical illness. The difference between the Movement Recovery Scores (MRSs) attained by a patient, when undertaking a given exercise routine on successive instances, is given by a statistical distance/divergence between the (posterior) probabilities of random graphs that are Bayesianly learnt using time series data on locations of 20 of the patient's joints, recorded on an e-platform as the patient exercises. This allows for the computation of the MRS on every occasion the patient undertakes this exercise, using which, the recovery trajectory is drawn. We learn each graph as a Random Geometric Graph drawn in a probabilistic metric space, and identify the closed-form marginal posterior of any edge of the graph, given the correlation structure of the multivariate time series data on joint locations. On the basis of our recovery learning, we offer recommendations on the optimal exercise routines for patients with given level of mobility impairment.
Authors: Martin Gjorgjevski, Nicolas Keriven, Simon Barthelm\'e, Yohann De Castro
Abstract: Node regression consists in predicting the value of a graph label at a node, given observations at the other nodes. To gain some insight into the performance of various estimators for this task, we perform a theoretical study in a context where the graph is random. Specifically, we assume that the graph is generated by a Latent Position Model, where each node of the graph has a latent position, and the probability that two nodes are connected depend on the distance between the latent positions of the two nodes. In this context, we begin by studying the simplest possible estimator for graph regression, which consists in averaging the value of the label at all neighboring nodes. We show that in Latent Position Models this estimator tends to a Nadaraya Watson estimator in the latent space, and that its rate of convergence is in fact the same. One issue with this standard estimator is that it averages over a region consisting of all neighbors of a node, and that depending on the graph model this may be too much or too little. An alternative consists in first estimating the true distances between the latent positions, then injecting these estimated distances into a classical Nadaraya Watson estimator. This enables averaging in regions either smaller or larger than the typical graph neighborhood. We show that this method can achieve standard nonparametric rates in certain instances even when the graph neighborhood is too large or too small.
Authors: Huan-Chih Wang, Ja-Ling Wu
Abstract: Face verification is a well-known image analysis application and is widely used to recognize individuals in contemporary society. However, most real-world recognition systems ignore the importance of protecting the identity-sensitive facial images that are used for verification. To address this problem, we investigate how to implement a secure face verification system that protects the facial images from being imitated. In our work, we use the DeepID2 convolutional neural network to extract the features of a facial image and an EM algorithm to solve the facial verification problem. To maintain the privacy of facial images, we apply homomorphic encryption schemes to encrypt the facial data and compute the EM algorithm in the ciphertext domain. We develop three face verification systems for surveillance (or entrance) control of a local community based on three levels of privacy concerns. The associated timing performances are presented to demonstrate their feasibility for practical implementation.
Authors: Martin Holler, Erion Morina
Abstract: This paper addresses the problem of uniqueness in learning physical laws for systems of partial differential equations (PDEs). Contrary to most existing approaches, it considers a framework of structured model learning, where existing, approximately correct physical models are augmented with components that are learned from data. The main result of the paper is a uniqueness result that covers a large class of PDEs and a suitable class of neural networks used for approximating the unknown model components. The uniqueness result shows that, in the idealized setting of full, noiseless measurements, a unique identification of the unknown model components is possible as regularization-minimizing solution of the PDE system. Furthermore, the paper provides a convergence result showing that model components learned on the basis of incomplete, noisy measurements approximate the ground truth model component in the limit. These results are possible under specific properties of the approximating neural networks and due to a dedicated choice of regularization. With this, a practical contribution of this analytic paper is to provide a class of model learning frameworks different to standard settings where uniqueness can be expected in the limit of full measurements.
Authors: Spyridon Kantarelis, Konstantinos Thomas, Vassilis Lyberatos, Edmund Dervakos, Giorgos Stamou
Abstract: Chord progressions encapsulate important information about music, pertaining to its structure and conveyed emotions. They serve as the backbone of musical composition, and in many cases, they are the sole information required for a musician to play along and follow the music. Despite their importance, chord progressions as a data domain remain underexplored. There is a lack of large-scale datasets suitable for deep learning applications, and limited research exploring chord progressions as an input modality. In this work, we present Chordonomicon, a dataset of over 666,000 songs and their chord progressions, annotated with structural parts, genre, and release date - created by scraping various sources of user-generated progressions and associated metadata. We demonstrate the practical utility of the Chordonomicon dataset for classification and generation tasks, and discuss its potential to provide valuable insights to the research community. Chord progressions are unique in their ability to be represented in multiple formats (e.g. text, graph) and the wealth of information chords convey in given contexts, such as their harmonic function . These characteristics make the Chordonomicon an ideal testbed for exploring advanced machine learning techniques, including transformers, graph machine learning, and hybrid systems that combine knowledge representation and machine learning.
Authors: Vu C. Dinh, Lam Si Tung Ho, Cuong V. Nguyen
Abstract: We analyze the error rates of the Hamiltonian Monte Carlo algorithm with leapfrog integrator for Bayesian neural network inference. We show that due to the non-differentiability of activation functions in the ReLU family, leapfrog HMC for networks with these activation functions has a large local error rate of $\Omega(\epsilon)$ rather than the classical error rate of $O(\epsilon^3)$. This leads to a higher rejection rate of the proposals, making the method inefficient. We then verify our theoretical findings through empirical simulations as well as experiments on a real-world dataset that highlight the inefficiency of HMC inference on ReLU-based neural networks compared to analytical networks.
Authors: Qizhi Chen, Delin Qu, Yiwen Tang, Haoming Song, Yiting Zhang, Dong Wang, Bin Zhao, Xuelong Li
Abstract: Reconstructing controllable Gaussian splats from monocular video is a challenging task due to its inherently insufficient constraints. Widely adopted approaches supervise complex interactions with additional masks and control signal annotations, limiting their real-world applications. In this paper, we propose an annotation guidance-free method, dubbed FreeGaussian, that mathematically derives dynamic Gaussian motion from optical flow and camera motion using novel dynamic Gaussian constraints. By establishing a connection between 2D flows and 3D Gaussian dynamic control, our method enables self-supervised optimization and continuity of dynamic Gaussian motions from flow priors. Furthermore, we introduce a 3D spherical vector controlling scheme, which represents the state with a 3D Gaussian trajectory, thereby eliminating the need for complex 1D control signal calculations and simplifying controllable Gaussian modeling. Quantitative and qualitative evaluations on extensive experiments demonstrate the state-of-the-art visual performance and control capability of our method. Project page: https://freegaussian.github.io.
Authors: Malte Algren, Christopher Pollard, John Andrew Raine, Tobias Golling
Abstract: In this paper, we present a novel method for pile-up removal of pp interactions using variational inference with diffusion models, called Vipr. Instead of using classification methods to identify which particles are from the primary collision, a generative model is trained to predict the constituents of the hard-scatter particle jets with pile-up removed. This results in an estimate of the full posterior over hard-scatter jet constituents, which has not yet been explored in the context of pile-up removal. We evaluate the performance of Vipr in a sample of jets from simulated $t\bar{t}$ events overlain with pile-up contamination. Vipr outperforms SoftDrop in predicting the substructure of the hard-scatter jets over a wide range of pile-up scenarios.
Authors: Qiuchi Li, Christina Lioma
Abstract: The matching of competences, such as skills, occupations or knowledges, is a key desiderata for candidates to be fit for jobs. Automatic extraction of competences from CVs and Jobs can greatly promote recruiters' productivity in locating relevant candidates for job vacancies. This work presents the first model that jointly extracts and classifies competence from Danish job postings. Different from existing works on skill extraction and skill classification, our model is trained on a large volume of annotated Danish corpora and is capable of extracting a wide range of Danish competences, including skills, occupations and knowledges of different categories. More importantly, as a single BERT-like architecture for joint extraction and classification, our model is lightweight and efficient at inference. On a real-scenario job matching dataset, our model beats the state-of-the-art models in the overall performance of Danish competence extraction and classification, and saves over 50% time at inference.
Authors: Elisabeth Kirsten, Ivan Habernal, Vedant Nanda, Muhammad Bilal Zafar
Abstract: Last few years have seen unprecedented advances in capabilities of Large Language Models (LLMs). These advancements promise to deeply benefit a vast array of application domains. However, due to their immense size, performing inference with LLMs is both costly and slow. Consequently, a plethora of recent work has proposed strategies to enhance inference efficiency, e.g., quantization, pruning, and caching. These acceleration strategies reduce the inference cost and latency, often by several factors, while maintaining much of the predictive performance measured via common benchmarks. In this work, we explore another critical aspect of LLM performance: demographic bias in model generations due to inference acceleration optimizations. Using a wide range of metrics, we probe bias in model outputs from a number of angles. Analysis of outputs before and after inference acceleration shows significant change in bias. Worryingly, these bias effects are complex and unpredictable. A combination of an acceleration strategy and bias type may show little bias change in one model but may lead to a large effect in another. Our results highlight a need for in-depth and case-by-case evaluation of model bias after it has been modified to accelerate inference.
Authors: Zhi Chang, Chukwudi Obite, Shuang Zhou, Shiwei Lan
Abstract: Motivated by deep neural networks, the deep Gaussian process (DGP) generalizes the standard GP by stacking multiple layers of GPs. Despite the enhanced expressiveness, GP, as an $L_2$ regularization prior, tends to be over-smooth and sub-optimal for inhomogeneous subjects, such as images with edges. Recently, Q-exponential process (Q-EP) has been proposed as an $L_q$ relaxation to GP and demonstrated with more desirable regularization properties through a parameter $q>0$ with $q=2$ corresponding to GP. Sharing the similar tractability of posterior and predictive distributions with GP, Q-EP can also be stacked to improve its modeling flexibility. In this paper, we generalize Q-EP to deep Q-EP to enjoy both proper regularization and improved expressiveness. The generalization is realized by introducing shallow Q-EP as a latent variable model and then building a hierarchy of the shallow Q-EP layers. Sparse approximation by inducing points and scalable variational strategy are applied to facilitate the inference. We demonstrate the numerical advantages of the proposed deep Q-EP model by comparing with multiple state-of-the-art deep probabilistic models.
Authors: Koen Ponse, Aske Plaat, Niki van Stein, Thomas M. Moerland
Abstract: Accurate economic simulations often require many experimental runs, particularly when combined with reinforcement learning. Unfortunately, training reinforcement learning agents in multi-agent economic environments can be slow. This paper introduces EconoJax, a fast simulated economy, based on the AI economist. EconoJax, and its training pipeline, are completely written in JAX. This allows EconoJax to scale to large population sizes and perform large experiments, while keeping training times within minutes. Through experiments with populations of 100 agents, we show how real-world economic behavior emerges through training within 15 minutes, in contrast to previous work that required several days. To aid and inspire researchers to build more rich and dynamic economic simulations, we open-source EconoJax on Github at: https://github.com/ponseko/econojax.
Authors: Eric Battenberg, RJ Skerry-Ryan, Daisy Stanton, Soroosh Mariooryad, Matt Shannon, Julian Salazar, David Kao
Abstract: Autoregressive (AR) Transformer-based sequence models are known to have difficulty generalizing to sequences longer than those seen during training. When applied to text-to-speech (TTS), these models tend to drop or repeat words or produce erratic output, especially for longer utterances. In this paper, we introduce enhancements aimed at AR Transformer-based encoder-decoder TTS systems that address these robustness and length generalization issues. Our approach uses an alignment mechanism to provide cross-attention operations with relative location information. The associated alignment position is learned as a latent property of the model via backprop and requires no external alignment information during training. While the approach is tailored to the monotonic nature of TTS input-output alignment, it is still able to benefit from the flexible modeling power of interleaved multi-head self- and cross-attention operations. A system incorporating these improvements, which we call Very Attentive Tacotron, matches the naturalness and expressiveness of a baseline T5-based TTS system, while eliminating problems with repeated or dropped words and enabling generalization to any practical utterance length.
Authors: Adrian Iordache, Bogdan Alexe, Radu Tudor Ionescu
Abstract: We propose a novel teacher-student framework to distill knowledge from multiple teachers trained on distinct datasets. Each teacher is first trained from scratch on its own dataset. Then, the teachers are combined into a joint architecture, which fuses the features of all teachers at multiple representation levels. The joint teacher architecture is fine-tuned on samples from all datasets, thus gathering useful generic information from all data samples. Finally, we employ a multi-level feature distillation procedure to transfer the knowledge to a student model for each of the considered datasets. We conduct image classification experiments on seven benchmarks, and action recognition experiments on three benchmarks. To illustrate the power of our feature distillation procedure, the student architectures are chosen to be identical to those of the individual teachers. To demonstrate the flexibility of our approach, we combine teachers with distinct architectures. We show that our novel Multi-Level Feature Distillation (MLFD) can significantly surpass equivalent architectures that are either trained on individual datasets, or jointly trained on all datasets at once. Furthermore, we confirm that each step of the proposed training procedure is well motivated by a comprehensive ablation study. We publicly release our code at https://github.com/AdrianIordache/MLFD.
Authors: Dimitrios G. Patsatzis, Mario di Bernardo, Lucia Russo, Constantinos Siettos
Abstract: We present GRINNs: numerical analysis-informed neural networks for the solution of inverse problems of non-linear systems of conservation laws. GRINNs are based on high-resolution Godunov schemes for the solution of the Riemann problem in hyperbolic Partial Differential Equations (PDEs). In contrast to other existing machine learning methods that learn the numerical fluxes of conservative Finite Volume methods, GRINNs learn the physical flux function per se. Due to their structure, GRINNs provide interpretable, conservative schemes, that learn the solution operator on the basis of approximate Riemann solvers that satisfy the Rankine-Hugoniot condition. The performance of GRINNs is assessed via four benchmark problems, namely the Burgers', the Shallow Water, the Lighthill-Whitham-Richards and the Payne-Whitham traffic flow models. The solution profiles of these PDEs exhibit shock waves, rarefactions and/or contact discontinuities at finite times. We demonstrate that GRINNs provide a very high accuracy both in the smooth and discontinuous regions.
Authors: Rakesh R. Menon, Shashank Srivastava
Abstract: Despite their high predictive accuracies, current machine learning systems often exhibit systematic biases stemming from annotation artifacts or insufficient support for certain classes in the dataset. Recent work proposes automatic methods for identifying and explaining systematic biases using keywords. We introduce DISCERN, a framework for interpreting systematic biases in text classifiers using language explanations. DISCERN iteratively generates precise natural language descriptions of systematic errors by employing an interactive loop between two large language models. Finally, we use the descriptions to improve classifiers by augmenting classifier training sets with synthetically generated instances or annotated examples via active learning. On three text-classification datasets, we demonstrate that language explanations from our framework induce consistent performance improvements that go beyond what is achievable with exemplars of systematic bias. Finally, in human evaluations, we show that users can interpret systematic biases more effectively (by over 25% relative) and efficiently when described through language explanations as opposed to cluster exemplars.
Authors: Rishabh Jain, Vivek M. Bhasi, Adwait Jog, Anand Sivasubramaniam, Mahmut T. Kandemir, Chita R. Das
Abstract: Personalized recommendation is a ubiquitous application on the internet, with many industries and hyperscalers extensively leveraging Deep Learning Recommendation Models (DLRMs) for their personalization needs (like ad serving or movie suggestions). With growing model and dataset sizes pushing computation and memory requirements, GPUs are being increasingly preferred for executing DLRM inference. However, serving newer DLRMs, while meeting acceptable latencies, continues to remain challenging, making traditional deployments increasingly more GPU-hungry, resulting in higher inference serving costs. In this paper, we show that the embedding stage continues to be the primary bottleneck in the GPU inference pipeline, leading up to a 3.2x embedding-only performance slowdown. To thoroughly grasp the problem, we conduct a detailed microarchitecture characterization and highlight the presence of low occupancy in the standard embedding kernels. By leveraging direct compiler optimizations, we achieve optimal occupancy, pushing the performance by up to 53%. Yet, long memory latency stalls continue to exist. To tackle this challenge, we propose specialized plug-and-play-based software prefetching and L2 pinning techniques, which help in hiding and decreasing the latencies. Further, we propose combining them, as they complement each other. Experimental evaluations using A100 GPUs with large models and datasets show that our proposed techniques improve performance by up to 103% for the embedding stage, and up to 77% for the overall DLRM inference pipeline.
Authors: Yuanxi Wang, Zuowen Wang, Shih-Chii Liu
Abstract: This paper presents an efficient deep learning solution for decoding motor movements from neural recordings in non-human primates. An Autoencoder Gated Recurrent Unit (AEGRU) model was adopted as the model architecture for this task. The autoencoder is only used during the training stage to achieve better generalization. Together with the preprocessing techniques, our model achieved 0.71 $R^2$ score, surpassing the baseline models in Neurobench and is ranked first for $R^2$ in the IEEE BioCAS 2024 Grand Challenge on Neural Decoding. Model pruning is also applied leading to a reduction of 41.4% of the multiply-accumulate (MAC) operations with little change in the $R^2$ score compared to the unpruned model.
Authors: Md. Ahsan Ayub, Subhabrata Majumdar
Abstract: Large Language Models (LLMs) are seeing significant adoption in every type of organization due to their exceptional generative capabilities. However, LLMs are found to be vulnerable to various adversarial attacks, particularly prompt injection attacks, which trick them into producing harmful or inappropriate content. Adversaries execute such attacks by crafting malicious prompts to deceive the LLMs. In this paper, we propose a novel approach based on embedding-based Machine Learning (ML) classifiers to protect LLM-based applications against this severe threat. We leverage three commonly used embedding models to generate embeddings of malicious and benign prompts and utilize ML classifiers to predict whether an input prompt is malicious. Out of several traditional ML methods, we achieve the best performance with classifiers built using Random Forest and XGBoost. Our classifiers outperform state-of-the-art prompt injection classifiers available in open-source implementations, which use encoder-only neural networks.
Authors: Chirag Modi, Diana Cai, Lawrence K. Saul
Abstract: Black-box variational inference (BBVI) scales poorly to high dimensional problems when it is used to estimate a multivariate Gaussian approximation with a full covariance matrix. In this paper, we extend the batch-and-match (BaM) framework for score-based BBVI to problems where it is prohibitively expensive to store such covariance matrices, let alone to estimate them. Unlike classical algorithms for BBVI, which use gradient descent to minimize the reverse Kullback-Leibler divergence, BaM uses more specialized updates to match the scores of the target density and its Gaussian approximation. We extend the updates for BaM by integrating them with a more compact parameterization of full covariance matrices. In particular, borrowing ideas from factor analysis, we add an extra step to each iteration of BaM -- a patch -- that projects each newly updated covariance matrix into a more efficiently parameterized family of diagonal plus low rank matrices. We evaluate this approach on a variety of synthetic target distributions and real-world problems in high-dimensional inference.
Authors: Souraja Kundu, Saket Singh, Yuji Iwahori
Abstract: Generating music from images can enhance various applications, including background music for photo slideshows, social media experiences, and video creation. This paper presents an emotion-guided image-to-music generation framework that leverages the Valence-Arousal (VA) emotional space to produce music that aligns with the emotional tone of a given image. Unlike previous models that rely on contrastive learning for emotional consistency, the proposed approach directly integrates a VA loss function to enable accurate emotional alignment. The model employs a CNN-Transformer architecture, featuring pre-trained CNN image feature extractors and three Transformer encoders to capture complex, high-level emotional features from MIDI music. Three Transformer decoders refine these features to generate musically and emotionally consistent MIDI sequences. Experimental results on a newly curated emotionally paired image-MIDI dataset demonstrate the proposed model's superior performance across metrics such as Polyphony Rate, Pitch Entropy, Groove Consistency, and loss convergence.
Authors: Harish Karthikeyan, Antigoni Polychroniadou
Abstract: Our work aims to minimize interaction in secure computation due to the high cost and challenges associated with communication rounds, particularly in scenarios with many clients. In this work, we revisit the problem of secure aggregation in the single-server setting where a single evaluation server can securely aggregate client-held individual inputs. Our key contribution is the introduction of One-shot Private Aggregation ($\mathsf{OPA}$) where clients speak only once (or even choose not to speak) per aggregation evaluation. Since each client communicates only once per aggregation, this simplifies managing dropouts and dynamic participation, contrasting with multi-round protocols and aligning with plaintext secure aggregation, where clients interact only once. We construct $\mathsf{OPA}$ based on LWR, LWE, class groups, DCR and demonstrate applications to privacy-preserving Federated Learning (FL) where clients \emph{speak once}. This is a sharp departure from prior multi-round FL protocols whose study was initiated by Bonawitz et al. (CCS, 2017). Moreover, unlike the YOSO (You Only Speak Once) model for general secure computation, $\mathsf{OPA}$ eliminates complex committee selection protocols to achieve adaptive security. Beyond asymptotic improvements, $\mathsf{OPA}$ is practical, outperforming state-of-the-art solutions. We benchmark logistic regression classifiers for two datasets, while also building an MLP classifier to train on MNIST, CIFAR-10, and CIFAR-100 datasets. We build two flavors of $\caps$ (1) from (threshold) key homomorphic PRF and (2) from seed homomorphic PRG and secret sharing.
Authors: Yihe Deng, Paul Mineiro
Abstract: Mathematical reasoning is a crucial capability for Large Language Models (LLMs), yet generating detailed and accurate reasoning traces remains a significant challenge. This paper introduces a novel approach to produce high-quality reasoning traces for LLM fine-tuning using online learning \textbf{Flows}. Our method employs an incremental output production Flow, where component LLMs collaboratively construct solutions through iterative communication. We train the Flow using online Direct Preference Optimization (DPO) learning with rollouts, generating DPO pairs for each training example and updating models in real-time. We directly compare the quality of reasoning traces generated by our method with those produced through direct model inference, demonstrating the effectiveness of our approach in improving LLM performance in mathematical reasoning tasks.
Authors: Grace Luo, Trevor Darrell, Amir Bar
Abstract: We investigate the internal representations of vision-and-language models (VLMs) and how they encode task representations. We consider tasks specified through examples or instructions, using either text or image inputs. Surprisingly, we find that conceptually similar tasks are mapped to similar task vector representations, regardless of how they are specified. Our findings suggest that to output answers, tokens in VLMs undergo three distinct phases: input, task, and answer, a process which is consistent across different modalities and specifications. The task vectors we identify in VLMs are general enough to be derived in one modality (e.g., text) and transferred to another (e.g., image). Additionally, we find that ensembling exemplar and instruction based task vectors produce better task representations. Taken together, these insights shed light on the underlying mechanisms of VLMs, particularly their ability to represent tasks in a shared manner across different modalities and task specifications. Project page: https://task-vectors-are-cross-modal.github.io.
Authors: Murtaza Dalal, Min Liu, Walter Talbott, Chen Chen, Deepak Pathak, Jian Zhang, Ruslan Salakhutdinov
Abstract: Sim2real for robotic manipulation is difficult due to the challenges of simulating complex contacts and generating realistic task distributions. To tackle the latter problem, we introduce ManipGen, which leverages a new class of policies for sim2real transfer: local policies. Locality enables a variety of appealing properties including invariances to absolute robot and object pose, skill ordering, and global scene configuration. We combine these policies with foundation models for vision, language and motion planning and demonstrate SOTA zero-shot performance of our method to Robosuite benchmark tasks in simulation (97%). We transfer our local policies from simulation to reality and observe they can solve unseen long-horizon manipulation tasks with up to 8 stages with significant pose, object and scene configuration variation. ManipGen outperforms SOTA approaches such as SayCan, OpenVLA, LLMTrajGen and VoxPoser across 50 real-world manipulation tasks by 36%, 76%, 62% and 60% respectively. Video results at https://mihdalal.github.io/manipgen/
Authors: Mauricio Soto-Gomez, Peter Robinson, Carlos Cano, Ali Pashaeibarough, Emanuele Cavalleri, Justin Reese, Marco Mesiti, Giorgio Valentini, Elena Casiraghi
Abstract: Many real-world problems are naturally modeled as heterogeneous graphs, where nodes and edges represent multiple types of entities and relations. Existing learning models for heterogeneous graph representation usually depend on the computation of specific and user-defined heterogeneous paths, or in the application of large and often not scalable deep neural network architectures. We propose Het-node2vec, an extension of the node2vec algorithm, designed for embedding heterogeneous graphs. Het-node2vec addresses the challenge of capturing the topological and structural characteristics of graphs and the semantic information underlying the different types of nodes and edges of heterogeneous graphs, by introducing a simple stochastic node and edge type switching strategy in second order random walk processes. The proposed approach also introduces an ''attention mechanism'' to focus the random walks on specific node and edge types, thus allowing more accurate embeddings and more focused predictions on specific node and edge types of interest. Empirical results on benchmark datasets show that Hetnode2vec achieves comparable or superior performance with respect to state-of-the-art methods for heterogeneous graphs in node label and edge prediction tasks.
Authors: Lin Fan, Peter W. Glynn
Abstract: We study the behavior of Thompson sampling from the perspective of weak convergence. In the regime where the gaps between arm means scale as $1/\sqrt{n}$ with the time horizon $n$, we show that the dynamics of Thompson sampling evolve according to discrete versions of SDE's and stochastic ODE's. As $n \to \infty$, we show that the dynamics converge weakly to solutions of the corresponding SDE's and stochastic ODE's. Our weak convergence theory is developed from first principles using the Continuous Mapping Theorem, and can be easily adapted to analyze other sampling-based bandit algorithms. In this regime, we also show that the weak limits of the dynamics of many sampling-based algorithms -- including Thompson sampling designed for any exponential family of rewards, and algorithms involving bootstrap-based sampling -- coincide with those of Gaussian Thompson sampling. Moreover, in this regime, these algorithms are generally robust to model mis-specification.
Authors: Yunxiao Guo, Han Long, Xiaojun Duan, Kaiyuan Feng, Maochu Li, Xiaying Ma
Abstract: As a popular Deep Reinforcement Learning (DRL) algorithm, Proximal Policy Optimization (PPO) has demonstrated remarkable efficacy in numerous complex tasks. According to the penalty mechanism in a surrogate, PPO can be classified into PPO with KL divergence (PPO-KL) and PPO with Clip (PPO-Clip). In this paper, we analyze the impact of asymmetry in KL divergence on PPO-KL and highlight that when this asymmetry is pronounced, it will misguide the improvement of the surrogate. To address this issue, we represent the PPO-KL in inner product form and demonstrate that the KL divergence is a Correntropy Induced Metric (CIM) in Euclidean space. Subsequently, we extend the PPO-KL to the Reproducing Kernel Hilbert Space (RKHS), redefine the inner products with RKHS, and propose the PPO-CIM algorithm. Moreover, this paper states that the PPO-CIM algorithm has a lower computation cost in policy gradient and proves that PPO-CIM can guarantee the new policy is within the trust region while the kernel satisfies some conditions. Finally, we design experiments based on six Mujoco continuous-action tasks to validate the proposed algorithm. The experimental results validate that the asymmetry of KL divergence can affect the policy improvement of PPO-KL and show that the PPO-CIM can perform better than both PPO-KL and PPO-Clip in most tasks.
Authors: Andrew Patterson, Samuel Neumann, Martha White, Adam White
Abstract: Empirical design in reinforcement learning is no small task. Running good experiments requires attention to detail and at times significant computational resources. While compute resources available per dollar have continued to grow rapidly, so have the scale of typical experiments in reinforcement learning. It is now common to benchmark agents with millions of parameters against dozens of tasks, each using the equivalent of 30 days of experience. The scale of these experiments often conflict with the need for proper statistical evidence, especially when comparing algorithms. Recent studies have highlighted how popular algorithms are sensitive to hyper-parameter settings and implementation details, and that common empirical practice leads to weak statistical evidence (Machado et al., 2018; Henderson et al., 2018). Here we take this one step further. This manuscript represents both a call to action, and a comprehensive resource for how to do good experiments in reinforcement learning. In particular, we cover: the statistical assumptions underlying common performance measures, how to properly characterize performance variation and stability, hypothesis testing, special considerations for comparing multiple agents, baseline and illustrative example construction, and how to deal with hyper-parameters and experimenter bias. Throughout we highlight common mistakes found in the literature and the statistical consequences of those in example experiments. The objective of this document is to provide answers on how we can use our unprecedented compute to do good science in reinforcement learning, as well as stay alert to potential pitfalls in our empirical design.
Authors: Qi Wang, Junming Yang, Yunbo Wang, Xin Jin, Wenjun Zeng, Xiaokang Yang
Abstract: Training offline RL models using visual inputs poses two significant challenges, i.e., the overfitting problem in representation learning and the overestimation bias for expected future rewards. Recent work has attempted to alleviate the overestimation bias by encouraging conservative behaviors. This paper, in contrast, tries to build more flexible constraints for value estimation without impeding the exploration of potential advantages. The key idea is to leverage off-the-shelf RL simulators, which can be easily interacted with in an online manner, as the "test bed" for offline policies. To enable effective online-to-offline knowledge transfer, we introduce CoWorld, a model-based RL approach that mitigates cross-domain discrepancies in state and reward spaces. Experimental results demonstrate the effectiveness of CoWorld, outperforming existing RL approaches by large margins.
Authors: Ali Shirali, Moritz Hardt
Abstract: ImageNet was famously created from Flickr image search results. What if we recreated ImageNet instead by searching the massive LAION dataset based on image captions alone? In this work, we carry out this counterfactual investigation. We find that the resulting ImageNet recreation, which we call LAIONet, looks distinctly unlike the original. Specifically, the intra-class similarity of images in the original ImageNet is dramatically higher than it is for LAIONet. Consequently, models trained on ImageNet perform significantly worse on LAIONet. We propose a rigorous explanation for the discrepancy in terms of a subtle, yet important, difference in two plausible causal data-generating processes for the respective datasets, that we support with systematic experimentation. In a nutshell, searching based on an image caption alone creates an information bottleneck that mitigates the selection bias otherwise present in image-based filtering. Our explanation formalizes a long-held intuition in the community that ImageNet images are stereotypical, unnatural, and overly simple representations of the class category. At the same time, it provides a simple and actionable takeaway for future dataset creation efforts.
Authors: Akshay Mehra, Yunbei Zhang, Jihun Hamm
Abstract: The growing popularity of transfer learning, due to the availability of models pre-trained on vast amounts of data, makes it imperative to understand when the knowledge of these pre-trained models can be transferred to obtain high-performing models on downstream target tasks. However, the exact conditions under which transfer learning succeeds in a cross-domain cross-task setting are still poorly understood. To bridge this gap, we propose a novel analysis that analyzes the transferability of the representations of pre-trained models to downstream tasks in terms of their relatedness to a given reference task. Our analysis leads to an upper bound on transferability in terms of task-relatedness, quantified using the difference between the class priors, label sets, and features of the two tasks. Our experiments using state-of-the-art pre-trained models show the effectiveness of task-relatedness in explaining transferability on various vision and language tasks. The efficient computability of task-relatedness even without labels of the target task and its high correlation with the model's accuracy after end-to-end fine-tuning on the target task makes it a useful metric for transferability estimation. Our empirical results of using task-relatedness to select the best pre-trained model from a model zoo for a target task highlight its utility for practical problems.
Authors: Chaochao Chen, Xiaohua Feng, Yuyuan Li, Lingjuan Lyu, Jun Zhou, Xiaolin Zheng, Jianwei Yin
Abstract: As the parameter size of Large Language Models (LLMs) continues to expand, there is an urgent need to address the scarcity of high-quality data. In response, existing research has attempted to make a breakthrough by incorporating Federated Learning (FL) into LLMs. Conversely, considering the outstanding performance of LLMs in task generalization, researchers have also tried applying LLMs within FL to tackle challenges in relevant domains. The complementarity between LLMs and FL has already ignited widespread research interest. In this paper, we aim to deeply explore the integration of LLMs and FL. We propose a research framework, dividing the fusion of LLMs and FL into three parts: the combination of LLM sub-technologies with FL, the integration of FL sub-technologies with LLMs, and the overall merger of LLMs and FL. We first provide a comprehensive review of the current state of research in the domain of LLMs combined with FL, including their typical applications, integration advantages, challenges faced, and future directions for resolution. Subsequently, we discuss the practical applications of the combination of LLMs and FL in critical scenarios such as healthcare, finance, and education, and provide new perspectives and insights into future research directions for LLMs and FL.
Authors: Thanveer Shaik, Xiaohui Tao, Lin Li, Haoran Xie, Hong-Ning Dai, Feng Zhao, Jianming Yong
Abstract: Effective patient monitoring is vital for timely interventions and improved healthcare outcomes. Traditional monitoring systems often struggle to handle complex, dynamic environments with fluctuating vital signs, leading to delays in identifying critical conditions. To address this challenge, we propose a novel AI-driven patient monitoring framework using multi-agent deep reinforcement learning (DRL). Our approach deploys multiple learning agents, each dedicated to monitoring a specific physiological feature, such as heart rate, respiration, and temperature. These agents interact with a generic healthcare monitoring environment, learn the patients' behavior patterns, and make informed decisions to alert the corresponding Medical Emergency Teams (METs) based on the level of emergency estimated. In this study, we evaluate the performance of the proposed multi-agent DRL framework using real-world physiological and motion data from two datasets: PPG-DaLiA and WESAD. We compare the results with several baseline models, including Q-Learning, PPO, Actor-Critic, Double DQN, and DDPG, as well as monitoring frameworks like WISEML and CA-MAQL. Our experiments demonstrate that the proposed DRL approach outperforms all other baseline models, achieving more accurate monitoring of patient's vital signs. Furthermore, we conduct hyperparameter optimization to fine-tune the learning process of each agent. By optimizing hyperparameters, we enhance the learning rate and discount factor, thereby improving the agents' overall performance in monitoring patient health status.
Authors: Long Tan Le, Tuan Dung Nguyen, Tung-Anh Nguyen, Choong Seon Hong, Suranga Seneviratne, Wei Bao, Nguyen H. Tran
Abstract: Federated Learning (FL) has emerged as a groundbreaking distributed learning paradigm enabling clients to train a global model collaboratively without exchanging data. Despite enhancing privacy and efficiency in information retrieval and knowledge management contexts, training and deploying FL models confront significant challenges such as communication bottlenecks, data heterogeneity, and memory limitations. To comprehensively address these challenges, we introduce FeDEQ, a novel FL framework that incorporates deep equilibrium learning and consensus optimization to harness compact global data representations for efficient personalization. Specifically, we design a unique model structure featuring an equilibrium layer for global representation extraction, followed by explicit layers tailored for local personalization. We then propose a novel FL algorithm rooted in the alternating directions method of multipliers (ADMM), which enables the joint optimization of a shared equilibrium layer and individual personalized layers across distributed datasets. Our theoretical analysis confirms that FeDEQ converges to a stationary point, achieving both compact global representations and optimal personalized parameters for each client. Extensive experiments on various benchmarks demonstrate that FeDEQ matches the performance of state-of-the-art personalized FL methods, while significantly reducing communication size by up to 4 times and memory footprint by 1.5 times during training.
Authors: Xin Liu, Wei li, Dazhi Zhan, Yu Pan, Xin Ma, Yu Ding, Zhisong Pan
Abstract: Federated learning (FL) is a widely employed distributed paradigm for collaboratively training machine learning models from multiple clients without sharing local data. In practice, FL encounters challenges in dealing with partial client participation due to the limited bandwidth, intermittent connection and strict synchronized delay. Simultaneously, there exist few theoretical convergence guarantees in this practical setting, especially when associated with the non-convex optimization of neural networks. To bridge this gap, we focus on the training problem of federated averaging (FedAvg) method for two canonical models: a deep linear network and a two-layer ReLU network. Under the over-parameterized assumption, we provably show that FedAvg converges to a global minimum at a linear rate $\mathcal{O}\left((1-\frac{min_{i \in [t]}|S_i|}{N^2})^t\right)$ after $t$ iterations, where $N$ is the number of clients and $|S_i|$ is the number of the participated clients in the $i$-th iteration. Experimental evaluations confirm our theoretical results.
Authors: Khaled Mohammed Saifuddin, Mehmet Emin Aktas, Esra Akbas
Abstract: Hypergraphs, with their capacity to depict high-order relationships, have emerged as a significant extension of traditional graphs. Although Graph Neural Networks (GNNs) have remarkable performance in graph representation learning, their extension to hypergraphs encounters challenges due to their intricate structures. Furthermore, current hypergraph transformers, a special variant of GNN, utilize semantic feature-based self-attention, ignoring topological attributes of nodes and hyperedges. To address these challenges, we propose a Topology-guided Hypergraph Transformer Network (THTN). In this model, we first formulate a hypergraph from a graph while retaining its structural essence to learn higher-order relations within the graph. Then, we design a simple yet effective structural and spatial encoding module to incorporate the topological and spatial information of the nodes into their representation. Further, we present a structure-aware self-attention mechanism that discovers the important nodes and hyperedges from both semantic and structural viewpoints. By leveraging these two modules, THTN crafts an improved node representation, capturing both local and global topological expressions. Extensive experiments conducted on node classification tasks demonstrate that the performance of the proposed model consistently exceeds that of the existing approaches.
Authors: Ruijie Jiang, Thuan Nguyen, Shuchin Aeron, Prakash Ishwar
Abstract: For a widely-studied data model and general loss and sample-hardening functions we prove that the losses of Supervised Contrastive Learning (SCL), Hard-SCL (HSCL), and Unsupervised Contrastive Learning (UCL) are minimized by representations that exhibit Neural-Collapse (NC), i.e., the class means form an Equiangular Tight Frame (ETF) and data from the same class are mapped to the same representation. We also prove that for any representation mapping, the HSCL and Hard-UCL (HUCL) losses are lower bounded by the corresponding SCL and UCL losses. In contrast to existing literature, our theoretical results for SCL do not require class-conditional independence of augmented views and work for a general loss function class that includes the widely used InfoNCE loss function. Moreover, our proofs are simpler, compact, and transparent. Similar to existing literature, our theoretical claims also hold for the practical scenario where batching is used for optimization. We empirically demonstrate, for the first time, that Adam optimization (with batching) of HSCL and HUCL losses with random initialization and suitable hardness levels can indeed converge to the NC-geometry if we incorporate unit-ball or unit-sphere feature normalization. Without incorporating hard-negatives or feature normalization, however, the representations learned via Adam suffer from Dimensional-Collapse (DC) and fail to attain the NC-geometry. These results exemplify the role of hard-negative sampling in contrastive representation learning and we conclude with several open theoretical problems for future work. The code can be found at \url{https://github.com/rjiang03/HCL/tree/main}
Authors: Hao Sun, Alex J. Chan, Nabeel Seedat, Alihan H\"uy\"uk, Mihaela van der Schaar
Abstract: Evaluating the value of a hypothetical target policy with only a logged dataset is important but challenging. On the one hand, it brings opportunities for safe policy improvement under high-stakes scenarios like clinical guidelines. On the other hand, such opportunities raise a need for precise off-policy evaluation (OPE). While previous work on OPE focused on improving the algorithm in value estimation, in this work, we emphasize the importance of the offline dataset, hence putting forward a data-centric framework for evaluating OPE problems. We propose DataCOPE, a data-centric framework for evaluating OPE, that answers the questions of whether and to what extent we can evaluate a target policy given a dataset. DataCOPE (1) forecasts the overall performance of OPE algorithms without access to the environment, which is especially useful before real-world deployment where evaluating OPE is impossible; (2) identifies the sub-group in the dataset where OPE can be inaccurate; (3) permits evaluations of datasets or data-collection strategies for OPE problems. Our empirical analysis of DataCOPE in the logged contextual bandit settings using healthcare datasets confirms its ability to evaluate both machine-learning and human expert policies like clinical guidelines. Finally, we apply DataCOPE to the task of reward modeling in Large Language Model alignment to demonstrate its scalability in real-world applications.
Authors: Quentin Delfosse, Sebastian Sztwiertnia, Mark Rothermel, Wolfgang Stammer, Kristian Kersting
Abstract: Goal misalignment, reward sparsity and difficult credit assignment are only a few of the many issues that make it difficult for deep reinforcement learning (RL) agents to learn optimal policies. Unfortunately, the black-box nature of deep neural networks impedes the inclusion of domain experts for inspecting the model and revising suboptimal policies. To this end, we introduce *Successive Concept Bottleneck Agents* (SCoBots), that integrate consecutive concept bottleneck (CB) layers. In contrast to current CB models, SCoBots do not just represent concepts as properties of individual objects, but also as relations between objects which is crucial for many RL tasks. Our experimental results provide evidence of SCoBots' competitive performances, but also of their potential for domain experts to understand and regularize their behavior. Among other things, SCoBots enabled us to identify a previously unknown misalignment problem in the iconic video game, Pong, and resolve it. Overall, SCoBots thus result in more human-aligned RL agents. Our code is available at https://github.com/k4ntz/SCoBots .
Authors: Luca Manneschi, Ian T. Vidamour, Kilian D. Stenning, Charles Swindells, Guru Venkat, David Griffin, Lai Gui, Daanish Sonawala, Denis Donskikh, Dana Hariga, Susan Stepney, Will R. Branford, Jack C. Gartside, Thomas Hayward, Matthew O. A. Ellis, Eleni Vasilaki
Abstract: Physical computing has the potential to enable widespread embodied intelligence by leveraging the intrinsic dynamics of complex systems for efficient sensing, processing, and interaction. While individual devices provide basic data processing capabilities, networks of interconnected devices can perform more complex and varied tasks. However, designing networks to perform dynamic tasks is challenging without physical models and accurate quantification of device noise. We propose a novel, noise-aware methodology for training device networks using Neural Stochastic Differential Equations (Neural-SDEs) as differentiable digital twins, accurately capturing the dynamics and associated stochasticity of devices with intrinsic memory. Our approach employs backpropagation through time and cascade learning, allowing networks to effectively exploit the temporal properties of physical devices. We validate our method on diverse networks of spintronic devices across temporal classification and regression benchmarks. By decoupling the training of individual device models from network training, our method reduces the required training data and provides a robust framework for programming dynamical devices without relying on analytical descriptions of their dynamics.
Authors: Tiansheng Huang, Sihao Hu, Ling Liu
Abstract: The new paradigm of finetuning-as-a-service introduces a new attack surface for Large Language Models (LLMs): a few harmful data uploaded by users can easily trick the finetuning to produce an alignment-broken model. We conduct an empirical analysis and uncover a \textit{harmful embedding drift} phenomenon, showing a probable cause of the alignment-broken effect. Inspired by our findings, we propose Vaccine, a perturbation-aware alignment technique to mitigate the security risk of users finetuning. The core idea of Vaccine is to produce invariant hidden embeddings by progressively adding crafted perturbation to them in the alignment phase. This enables the embeddings to withstand harmful perturbation from un-sanitized user data in the finetuning phase. Our results on open source mainstream LLMs (e.g., Llama2, Opt, Vicuna) demonstrate that Vaccine can boost the robustness of alignment against harmful prompts induced embedding drift while reserving reasoning ability towards benign prompts. Our code is available at \url{https://github.com/git-disl/Vaccine}.
Authors: Qilong Ma, Haixu Wu, Lanxiang Xing, Shangchen Miao, Mingsheng Long
Abstract: Accurately predicting the future fluid is vital to extensive areas such as meteorology, oceanology, and aerodynamics. However, since the fluid is usually observed from the Eulerian perspective, its moving and intricate dynamics are seriously obscured and confounded in static grids, bringing thorny challenges to the prediction. This paper introduces a new Lagrangian-Eulerian combined paradigm to tackle the tanglesome fluid dynamics. Instead of solely predicting the future based on Eulerian observations, we propose DeepLag to discover hidden Lagrangian dynamics within the fluid by tracking the movements of adaptively sampled key particles. Further, DeepLag presents a new paradigm for fluid prediction, where the Lagrangian movement of the tracked particles is inferred from Eulerian observations, and their accumulated Lagrangian dynamics information is incorporated into global Eulerian evolving features to guide future prediction respectively. Tracking key particles not only provides a transparent and interpretable clue for fluid dynamics but also makes our model free from modeling complex correlations among massive grids for better efficiency. Experimentally, DeepLag excels in three challenging fluid prediction tasks covering 2D and 3D, simulated and real-world fluids. Code is available at this repository: https://github.com/thuml/DeepLag.
Authors: Yao Shu, Jiongfeng Fang, Ying Tiffany He, Fei Richard Yu
Abstract: First-order optimization (FOO) algorithms are pivotal in numerous computational domains such as machine learning and signal denoising. However, their application to complex tasks like neural network training often entails significant inefficiencies due to the need for many sequential iterations for convergence. In response, we introduce first-order optimization expedited with approximately parallelized iterations (OptEx), the first framework that enhances the efficiency of FOO by leveraging parallel computing to mitigate its iterative bottleneck. OptEx employs kernelized gradient estimation to make use of gradient history for future gradient prediction, enabling parallelization of iterations -- a strategy once considered impractical because of the inherent iterative dependency in FOO. We provide theoretical guarantees for the reliability of our kernelized gradient estimation and the iteration complexity of SGD-based OptEx, confirming that estimation errors diminish to zero as historical gradients accumulate and that SGD-based OptEx enjoys an effective acceleration rate of $\Omega(\sqrt{N})$ over standard SGD given parallelism of N. We also use extensive empirical studies, including synthetic functions, reinforcement learning tasks, and neural network training across various datasets, to underscore the substantial efficiency improvements achieved by OptEx.
Authors: Jonas H\"ubotter, Bhavya Sukhija, Lenart Treven, Yarden As, Andreas Krause
Abstract: We study a generalization of classical active learning to real-world settings with concrete prediction targets where sampling is restricted to an accessible region of the domain, while prediction targets may lie outside this region. We analyze a family of decision rules that sample adaptively to minimize uncertainty about prediction targets. We are the first to show, under general regularity assumptions, that such decision rules converge uniformly to the smallest possible uncertainty obtainable from the accessible data. We demonstrate their strong sample efficiency in two key applications: active fine-tuning of large neural networks and safe Bayesian optimization, where they achieve state-of-the-art performance.
Authors: Tianjiao Luo, Tim Pearce, Huayu Chen, Jianfei Chen, Jun Zhu
Abstract: Generative Adversarial Imitation Learning (GAIL) trains a generative policy to mimic a demonstrator. It uses on-policy Reinforcement Learning (RL) to optimize a reward signal derived from a GAN-like discriminator. A major drawback of GAIL is its training instability - it inherits the complex training dynamics of GANs, and the distribution shift introduced by RL. This can cause oscillations during training, harming its sample efficiency and final policy performance. Recent work has shown that control theory can help with the convergence of a GAN's training. This paper extends this line of work, conducting a control-theoretic analysis of GAIL and deriving a novel controller that not only pushes GAIL to the desired equilibrium but also achieves asymptotic stability in a 'one-step' setting. Based on this, we propose a practical algorithm 'Controlled-GAIL' (C-GAIL). On MuJoCo tasks, our controlled variant is able to speed up the rate of convergence, reduce the range of oscillation and match the expert's distribution more closely both for vanilla GAIL and GAIL-DAC.
Authors: Zhiqi Bu, Xinwei Zhang, Mingyi Hong, Sheng Zha, George Karypis
Abstract: The superior performance of large foundation models relies on the use of massive amounts of high-quality data, which often contain sensitive, private and copyrighted material that requires formal protection. While differential privacy (DP) is a prominent method to gauge the degree of security provided to the models, its application is commonly limited to the model fine-tuning stage, due to the performance degradation when applying DP during the pre-training stage. Consequently, DP is yet not capable of protecting a substantial portion of the data used during the initial pre-training process. In this work, we first provide a theoretical understanding of the efficacy of DP training by analyzing the per-iteration loss improvement. We make a key observation that DP optimizers' performance degradation can be significantly mitigated by the use of limited public data, which leads to a novel DP continual pre-training strategy. Empirically, using only 10\% of public data, our strategy can achieve DP accuracy of 41.5\% on ImageNet-21k (with $\epsilon=8$), as well as non-DP accuracy of 55.7\% and and 60.0\% on downstream tasks Places365 and iNaturalist-2021, respectively, on par with state-of-the-art standard pre-training and substantially outperforming existing DP pre-trained models. Our DP pre-trained models are released in fastDP library (https://github.com/awslabs/fast-differential-privacy/releases/tag/v2.1)
URLs: https://github.com/awslabs/fast-differential-privacy/releases/tag/v2.1)
Authors: Zeheng Wang, James Scott Cooper, Muhammad Usman, Timothy van der Laan
Abstract: The rapid advancement of Internet of Things (IoT) necessitates the development of optimized nanoparticle-based Chemiresistive Sensor (CRS) arrays that are energy-efficient, specific, and sensitive. This study introduces an optimization strategy that employs a rapid ensemble learning-based model committee approach to achieve these goals. Utilizing machine learning models such as Elastic Net Regression, Random Forests, and XGBoost, among others, the strategy identifies the most impactful sensors in a CRS array for accurate classification. A weighted voting mechanism is introduced to aggregate the models' opinions in sensor selection, thereby setting up two distinct working modes, termed "Blue" and "Green". The Blue mode operates with all sensors for maximum detection capability, while the Green mode selectively activates only key sensors, significantly reducing energy consumption without compromising detection accuracy. The strategy is validated through theoretical calculations and Monte Carlo simulations, demonstrating its effectiveness and accuracy. The employed optimization strategy elevates the detection capability of CRS arrays while also pushing it closer to theoretical limits, promising significant implications for the development of low-cost, easily fabricable next-generation IoT sensor terminals.
Authors: Silvia Corbara, Alejandro Moreo
Abstract: Authorship Verification (AV) is a text classification task concerned with inferring whether a candidate text has been written by one specific author or by someone else. It has been shown that many AV systems are vulnerable to adversarial attacks, where a malicious author actively tries to fool the classifier by either concealing their writing style, or by imitating the style of another author. In this paper, we investigate the potential benefits of augmenting the classifier training set with (negative) synthetic examples. These synthetic examples are generated to imitate the style of the author of interest. We analyze the improvements in classifier prediction that this augmentation brings to bear in the task of AV in an adversarial setting. In particular, we experiment with three different generator architectures (one based on Recurrent Neural Networks, another based on small-scale transformers, and another based on the popular GPT model) and with two training strategies (one inspired by standard Language Models, and another inspired by Wasserstein Generative Adversarial Networks). We evaluate our hypothesis on five datasets (three of which have been specifically collected to represent an adversarial setting) and using two learning algorithms for the AV classifier (Support Vector Machines and Convolutional Neural Networks). This experimentation has yielded negative results, revealing that, although our methodology proves effective in many adversarial settings, its benefits are too sporadic for a pragmatical application.
Authors: Zhicong Tang, Tiankai Hang, Shuyang Gu, Dong Chen, Baining Guo
Abstract: This paper introduces a novel theoretical simplification of the Diffusion Schr\"odinger Bridge (DSB) that facilitates its unification with Score-based Generative Models (SGMs), addressing the limitations of DSB in complex data generation and enabling faster convergence and enhanced performance. By employing SGMs as an initial solution for DSB, our approach capitalizes on the strengths of both frameworks, ensuring a more efficient training process and improving the performance of SGM. We also propose a reparameterization technique that, despite theoretical approximations, practically improves the network's fitting capabilities. Our extensive experimental evaluations confirm the effectiveness of the simplified DSB, demonstrating its significant improvements. We believe the contributions of this work pave the way for advanced generative modeling.
Authors: Akshay Krishnamurthy, Keegan Harris, Dylan J. Foster, Cyril Zhang, Aleksandrs Slivkins
Abstract: We investigate the extent to which contemporary Large Language Models (LLMs) can engage in exploration, a core capability in reinforcement learning and decision making. We focus on native performance of existing LLMs, without training interventions. We deploy LLMs as agents in simple multi-armed bandit environments, specifying the environment description and interaction history entirely in-context, i.e., within the LLM prompt. We experiment with GPT-3.5, GPT-4, and Llama2, using a variety of prompt designs, and find that the models do not robustly engage in exploration without substantial interventions: i) Across all of our experiments, only one configuration resulted in satisfactory exploratory behavior: GPT-4 with chain-of-thought reasoning and an externally summarized interaction history, presented as sufficient statistics; ii) All other configurations did not result in robust exploratory behavior, including those with chain-of-thought reasoning but unsummarized history. Although these findings can be interpreted positively, they suggest that external summarization -- which may not be possible in more complex settings -- is important for obtaining desirable behavior from LLM agents. We conclude that non-trivial algorithmic interventions, such as fine-tuning or dataset curation, may be required to empower LLM-based decision making agents in complex settings.
Authors: Chanwoo Park, Xiangyu Liu, Asuman Ozdaglar, Kaiqing Zhang
Abstract: Large language models (LLMs) have been increasingly employed for (interactive) decision-making, via the development of LLM-based autonomous agents. Despite their emerging successes, the performance of LLM agents in decision-making has not been fully investigated through quantitative metrics, especially in the multi-agent setting when they interact with each other, a typical scenario in real-world LLM-agent applications. To better understand the limits of LLM agents in these interactive environments, we propose to study their interactions in benchmark decision-making settings in online learning and game theory, through the performance metric of \emph{regret}. We first empirically study the {no-regret} behaviors of LLMs in canonical (non-stationary) online learning problems, as well as the emergence of equilibria when LLM agents interact through playing repeated games. We then provide some theoretical insights into the no-regret behaviors of LLM agents, under certain assumptions on the supervised pre-training and the rationality model of human decision-makers who generate the data. Notably, we also identify (simple) cases where advanced LLMs such as GPT-4 fail to be no-regret. To promote the no-regret behaviors, we propose a novel \emph{unsupervised} training loss of \emph{regret-loss}, which, in contrast to the supervised pre-training loss, does not require the labels of (optimal) actions. We then establish the statistical guarantee of generalization bound for regret-loss minimization, followed by the optimization guarantee that minimizing such a loss may automatically lead to known no-regret learning algorithms. Our further experiments demonstrate the effectiveness of our regret-loss, especially in addressing the above ``regrettable'' cases.
Authors: Saleh Ashkboos, Amirkeivan Mohtashami, Maximilian L. Croci, Bo Li, Pashmina Cameron, Martin Jaggi, Dan Alistarh, Torsten Hoefler, James Hensman
Abstract: We introduce QuaRot, a new Quantization scheme based on Rotations, which is able to quantize LLMs end-to-end, including all weights, activations, and KV cache in 4 bits. QuaRot rotates LLMs in a way that removes outliers from the hidden state without changing the output, making quantization easier. This computational invariance is applied to the hidden state (residual) of the LLM, as well as to the activations of the feed-forward components, aspects of the attention mechanism, and to the KV cache. The result is a quantized model where all matrix multiplications are performed in 4 bits, without any channels identified for retention in higher precision. Our 4-bit quantized LLaMa2-70B model has losses of at most 0.47 WikiText-2 perplexity and retains 99% of the zero-shot performance. We also show that QuaRot can provide lossless 6 and 8 bit LLaMa2 models without any calibration data using round-to-nearest quantization. Code is available at: https://github.com/spcl/QuaRot.
Authors: Johnny Xi, Jana Osea, Zuheng Xu, Jason Hartford
Abstract: Multimodal representation learning techniques typically rely on paired samples to learn common representations, but paired samples are challenging to collect in fields such as biology where measurement devices often destroy the samples. This paper presents an approach to address the challenge of aligning unpaired samples across disparate modalities in multimodal representation learning. We draw an analogy between potential outcomes in causal inference and potential views in multimodal observations, which allows us to use Rubin's framework to estimate a common space in which to match samples. Our approach assumes we collect samples that are experimentally perturbed by treatments, and uses this to estimate a propensity score from each modality, which encapsulates all shared information between a latent state and treatment and can be used to define a distance between samples. We experiment with two alignment techniques that leverage this distance -- shared nearest neighbours (SNN) and optimal transport (OT) matching -- and find that OT matching results in significant improvements over state-of-the-art alignment approaches in both a synthetic multi-modal setting and in real-world data from NeurIPS Multimodal Single-Cell Integration Challenge.
Authors: Yuezhu Xu, S. Sivaranjani
Abstract: The Lipschitz constant plays a crucial role in certifying the robustness of neural networks to input perturbations. Since calculating the exact Lipschitz constant is NP-hard, efforts have been made to obtain tight upper bounds on the Lipschitz constant. Typically, this involves solving a large matrix verification problem, the computational cost of which grows significantly for both deeper and wider networks. In this paper, we provide a compositional approach to estimate Lipschitz constants for deep feed-forward neural networks. We first obtain an exact decomposition of the large matrix verification problem into smaller sub-problems. Then, leveraging the underlying cascade structure of the network, we develop two algorithms. The first algorithm explores the geometric features of the problem and enables us to provide Lipschitz estimates that are comparable to existing methods by solving small semidefinite programs (SDPs) that are only as large as the size of each layer. The second algorithm relaxes these sub-problems and provides a closed-form solution to each sub-problem for extremely fast estimation, altogether eliminating the need to solve SDPs. The two algorithms represent different levels of trade-offs between efficiency and accuracy. Finally, we demonstrate that our approach provides a steep reduction in computation time (as much as several thousand times faster, depending on the algorithm for deeper networks) while yielding Lipschitz bounds that are very close to or even better than those achieved by state-of-the-art approaches in a broad range of experiments. In summary, our approach considerably advances the scalability and efficiency of certifying neural network robustness, making it particularly attractive for online learning tasks.
Authors: Jiayi Shen, Cheems Wang, Zehao Xiao, Nanne Van Noord, Marcel Worring
Abstract: This paper proposes \textit{GO4Align}, a multi-task optimization approach that tackles task imbalance by explicitly aligning the optimization across tasks. To achieve this, we design an adaptive group risk minimization strategy, comprising two techniques in implementation: (i) dynamical group assignment, which clusters similar tasks based on task interactions; (ii) risk-guided group indicators, which exploit consistent task correlations with risk information from previous iterations. Comprehensive experimental results on diverse benchmarks demonstrate our method's performance superiority with even lower computational costs.
Authors: Roussel Desmond Nzoyem, David A. W. Barton, Tom Deakin
Abstract: Neural Ordinary Differential Equations (NODEs) often struggle to adapt to new dynamic behaviors caused by parameter changes in the underlying system, even when these dynamics are similar to previously observed behaviors. This problem becomes more challenging when the changing parameters are unobserved, meaning their value or influence cannot be directly measured when collecting data. To address this issue, we introduce Neural Context Flow (NCF), a robust and interpretable Meta-Learning framework that includes uncertainty estimation. NCF uses higher-order Taylor expansion to enable contextual self-modulation, allowing context vectors to influence dynamics from other domains while also modulating themselves. After establishing convergence guarantees, we empirically test NCF and compare it to related adaptation methods. Our results show that NCF achieves state-of-the-art Out-of-Distribution performance on 5 out of 6 linear and non-linear benchmark problems. Through extensive experiments, we explore the flexible model architecture of NCF and the encoded representations within the learned context vectors. Our findings highlight the potential implications of NCF for foundational models in the physical sciences, offering a promising approach to improving the adaptability and generalization of NODEs in various scientific applications. Our code is openly available at \url{https://github.com/ddrous/ncflow}.
Authors: Simon Kristoffersson Lind, Ziliang Xiong, Per-Erik Forss\'en, Volker Kr\"uger
Abstract: When deploying deep neural networks on robots or other physical systems, the learned model should reliably quantify predictive uncertainty. A reliable uncertainty allows downstream modules to reason about the safety of its actions. In this work, we address metrics for evaluating such an uncertainty. Specifically, we focus on regression tasks, and investigate Area Under Sparsification Error (AUSE), Calibration Error, Spearman's Rank Correlation, and Negative Log-Likelihood (NLL). Using synthetic regression datasets, we look into how those metrics behave under four typical types of uncertainty, their stability regarding the size of the test set, and reveal their strengths and weaknesses. Our results indicate that Calibration Error is the most stable and interpretable metric, but AUSE and NLL also have their respective use cases. We discourage the usage of Spearman's Rank Correlation for evaluating uncertainties and recommend replacing it with AUSE.
Authors: Nikita P. Kalinin, Christoph Lampert
Abstract: Current state-of-the-art methods for differentially private model training are based on matrix factorization techniques. However, these methods suffer from high computational overhead because they require numerically solving a demanding optimization problem to determine an approximately optimal factorization prior to the actual model training. In this work, we present a new matrix factorization approach, BSR, which overcomes this computational bottleneck. By exploiting properties of the standard matrix square root, BSR allows to efficiently handle also large-scale problems. For the key scenario of stochastic gradient descent with momentum and weight decay, we even derive analytical expressions for BSR that render the computational overhead negligible. We prove bounds on the approximation quality that hold both in the centralized and in the federated learning setting. Our numerical experiments demonstrate that models trained using BSR perform on par with the best existing methods, while completely avoiding their computational overhead.
Authors: Amir Mohammad Abouei, Ehsan Mokhtarian, Negar Kiyavash, Matthias Grossglauser
Abstract: The s-ID problem seeks to compute a causal effect in a specific sub-population from the observational data pertaining to the same sub population (Abouei et al., 2023). This problem has been addressed when all the variables in the system are observable. In this paper, we consider an extension of the s-ID problem that allows for the presence of latent variables. To tackle the challenges induced by the presence of latent variables in a sub-population, we first extend the classical relevant graphical definitions, such as c-components and Hedges, initially defined for the so-called ID problem (Pearl, 1995; Tian & Pearl, 2002), to their new counterparts. Subsequently, we propose a sound algorithm for the s-ID problem with latent variables.
Authors: Hyosoon Jang, Yunhui Jang, Minsu Kim, Jinkyoo Park, Sungsoo Ahn
Abstract: This paper studies Generative Flow Networks (GFlowNets), which learn to sample objects proportionally to a given reward function through the trajectory of state transitions. In this work, we observe that GFlowNets tend to under-exploit the high-reward objects due to training on insufficient number of trajectories, which may lead to a large gap between the estimated flow and the (known) reward value. In response to this challenge, we propose a pessimistic backward policy for GFlowNets (PBP-GFN), which maximizes the observed flow to align closely with the true reward for the object. We extensively evaluate PBP-GFN across eight benchmarks, including hyper-grid environment, bag generation, structured set generation, molecular generation, and four RNA sequence generation tasks. In particular, PBP-GFN enhances the discovery of high-reward objects, maintains the diversity of the objects, and consistently outperforms existing methods.
Authors: Zekun Cai, Guangji Bai, Renhe Jiang, Xuan Song, Liang Zhao
Abstract: Temporal Domain Generalization (TDG) addresses the challenge of training predictive models under temporally varying data distributions. Traditional TDG approaches typically focus on domain data collected at fixed, discrete time intervals, which limits their capability to capture the inherent dynamics within continuous-evolving and irregularly-observed temporal domains. To overcome this, this work formalizes the concept of Continuous Temporal Domain Generalization (CTDG), where domain data are derived from continuous times and are collected at arbitrary times. CTDG tackles critical challenges including: 1) Characterizing the continuous dynamics of both data and models, 2) Learning complex high-dimensional nonlinear dynamics, and 3) Optimizing and controlling the generalization across continuous temporal domains. To address them, we propose a Koopman operator-driven continuous temporal domain generalization (Koodos) framework. We formulate the problem within a continuous dynamic system and leverage the Koopman theory to learn the underlying dynamics; the framework is further enhanced with a comprehensive optimization strategy equipped with analysis and control driven by prior knowledge of the dynamics patterns. Extensive experiments demonstrate the effectiveness and efficiency of our approach. The code can be found at: https://github.com/Zekun-Cai/Koodos.
Authors: Saravanan Kandasamy, Dheeraj Nagaraj
Abstract: Langevin Dynamics is a Stochastic Differential Equation (SDE) central to sampling and generative modeling and is implemented via time discretization. Langevin Monte Carlo (LMC), based on the Euler-Maruyama discretization, is the simplest and most studied algorithm. LMC can suffer from slow convergence - requiring a large number of steps of small step-size to obtain good quality samples. This becomes stark in the case of diffusion models where a large number of steps gives the best samples, but the quality degrades rapidly with smaller number of steps. Randomized Midpoint Method has been recently proposed as a better discretization of Langevin dynamics for sampling from strongly log-concave distributions. However, important applications such as diffusion models involve non-log concave densities and contain time varying drift. We propose its variant, the Poisson Midpoint Method, which approximates a small step-size LMC with large step-sizes. We prove that this can obtain a quadratic speed up of LMC under very weak assumptions. We apply our method to diffusion models for image generation and show that it maintains the quality of DDPM with 1000 neural network calls with just 50-80 neural network calls and outperforms ODE based methods with similar compute.
Authors: Tiansheng Huang, Sihao Hu, Fatih Ilhan, Selim Furkan Tekin, Ling Liu
Abstract: Recent studies show that Large Language Models (LLMs) with safety alignment can be jail-broken by fine-tuning on a dataset mixed with harmful data. First time in the literature, we show that the jail-broken effect can be mitigated by separating states in the finetuning stage to optimize the alignment and user datasets. Unfortunately, our subsequent study shows that this simple Bi-State Optimization (BSO) solution experiences convergence instability when steps invested in its alignment state is too small, leading to downgraded alignment performance. By statistical analysis, we show that the \textit{excess drift} towards consensus could be a probable reason for the instability. To remedy this issue, we propose \textbf{L}azy(\textbf{i}) \textbf{s}afety \textbf{a}lignment (\textbf{Lisa}), which introduces a proximal term to constraint the drift of each state. Theoretically, the benefit of the proximal term is supported by the convergence analysis, wherein we show that a sufficient large proximal factor is necessary to guarantee Lisa's convergence. Empirically, our results on four downstream finetuning tasks show that Lisa with a proximal term can significantly increase alignment performance while maintaining the LLM's accuracy on the user tasks. Code is available at \url{https://github.com/git-disl/Lisa}.
Authors: Mouad El Bouchattaoui, Myriam Tami, Benoit Lepetit, Paul-Henry Courn\`ede
Abstract: Estimating treatment effects over time holds significance in various domains, including precision medicine, epidemiology, economy, and marketing. This paper introduces a unique approach to counterfactual regression over time, emphasizing long-term predictions. Distinguishing itself from existing models like Causal Transformer, our approach highlights the efficacy of employing RNNs for long-term forecasting, complemented by Contrastive Predictive Coding (CPC) and Information Maximization (InfoMax). Emphasizing efficiency, we avoid the need for computationally expensive transformers. Leveraging CPC, our method captures long-term dependencies in the presence of time-varying confounders. Notably, recent models have disregarded the importance of invertible representation, compromising identification assumptions. To remedy this, we employ the InfoMax principle, maximizing a lower bound of mutual information between sequence data and its representation. Our method achieves state-of-the-art counterfactual estimation results using both synthetic and real-world data, marking the pioneering incorporation of Contrastive Predictive Encoding in causal inference.
Authors: Jintang Li, Ruofan Wu, Xinzhou Jin, Boqun Ma, Liang Chen, Zibin Zheng
Abstract: Over the past few years, research on deep graph learning has shifted from static graphs to temporal graphs in response to real-world complex systems that exhibit dynamic behaviors. In practice, temporal graphs are formalized as an ordered sequence of static graph snapshots observed at discrete time points. Sequence models such as RNNs or Transformers have long been the predominant backbone networks for modeling such temporal graphs. Yet, despite the promising results, RNNs struggle with long-range dependencies, while transformers are burdened by quadratic computational complexity. Recently, state space models (SSMs), which are framed as discretized representations of an underlying continuous-time linear dynamical system, have garnered substantial attention and achieved breakthrough advancements in independent sequence modeling. In this work, we undertake a principled investigation that extends SSM theory to temporal graphs by integrating structural information into the online approximation objective via the adoption of a Laplacian regularization term. The emergent continuous-time system introduces novel algorithmic challenges, thereby necessitating our development of GraphSSM, a graph state space model for modeling the dynamics of temporal graphs. Extensive experimental results demonstrate the effectiveness of our GraphSSM framework across various temporal graph benchmarks.
Authors: Shay Deutsch, Lionel Yelibi, Alex Tong Lin, Arjun Ravi Kannan
Abstract: Deriving meaningful representations from complex, high-dimensional data in unsupervised settings is crucial across diverse machine learning applications. This paper introduces a framework for multi-scale graph network embedding based on spectral graph wavelets that employs a contrastive learning approach. We theoretically show that in Paley-Wiener spaces on combinatorial graphs, the spectral graph wavelets operator provides greater flexibility and control over smoothness compared to the Laplacian operator, motivating our approach. An additional key advantage of the proposed embedding is its ability to establish a correspondence between the embedding and input feature spaces, enabling the derivation of feature importance. We validate the effectiveness of our graph embedding framework on multiple public datasets across various downstream tasks, including clustering and unsupervised feature importance.
Authors: Tian Wang, Chuang Wang
Abstract: Neural operators effectively solve PDE problems from data without knowing the explicit equations, which learn the map from the input sequences of observed samples to the predicted values. Most existing works build the model in the original geometric space, leading to high computational costs when the number of sample points is large. We present the Latent Neural Operator (LNO) solving PDEs in the latent space. In particular, we first propose Physics-Cross-Attention (PhCA) transforming representation from the geometric space to the latent space, then learn the operator in the latent space, and finally recover the real-world geometric space via the inverse PhCA map. Our model retains flexibility that can decode values in any position not limited to locations defined in the training set, and therefore can naturally perform interpolation and extrapolation tasks particularly useful for inverse problems. Moreover, the proposed LNO improves both prediction accuracy and computational efficiency. Experiments show that LNO reduces the GPU memory by 50%, speeds up training 1.8 times, and reaches state-of-the-art accuracy on four out of six benchmarks for forward problems and a benchmark for inverse problem. Code is available at https://github.com/L-I-M-I-T/LatentNeuralOperator.
Authors: Zhihao Li, Zhilu Lai, Xiaobo Zhang, Wei Wang
Abstract: Solving partial differential equations (PDEs) effectively necessitates a multi-scale approach, particularly critical in high-dimensional scenarios characterized by increasing grid points or resolution. Traditional methods often fail to capture the detailed features necessary for accurate modeling, presenting a significant challenge in scientific computing. In response, we introduce the Multiwavelet-based Algebraic Multigrid Neural Operator (M2NO), a novel deep learning framework that synergistically combines multiwavelet transformations and algebraic multigrid (AMG) techniques. By exploiting the inherent similarities between these two approaches, M2NO overcomes their individual limitations and enhances precision and flexibility across various PDE benchmarks. Employing Multiresolution Analysis (MRA) with high-pass and low-pass filters, the model executes hierarchical decomposition to accurately delineate both global trends and localized details within PDE solutions, supporting adaptive data representation at multiple scales. M2NO also automates node selection and adeptly manages complex boundary conditions through its multiwavelet-based operators. Extensive evaluations on a diverse array of PDE datasets with different boundary conditions confirm M2NO's superior performance. Furthermore, M2NO excels in handling high-resolution and super-resolution tasks, consistently outperforming competing models and demonstrating robust adaptability in complex computational scenarios.
Authors: Seohong Park, Kevin Frans, Sergey Levine, Aviral Kumar
Abstract: While imitation learning requires access to high-quality data, offline reinforcement learning (RL) should, in principle, perform similarly or better with substantially lower data quality by using a value function. However, current results indicate that offline RL often performs worse than imitation learning, and it is often unclear what holds back the performance of offline RL. Motivated by this observation, we aim to understand the bottlenecks in current offline RL algorithms. While poor performance of offline RL is typically attributed to an imperfect value function, we ask: is the main bottleneck of offline RL indeed in learning the value function, or something else? To answer this question, we perform a systematic empirical study of (1) value learning, (2) policy extraction, and (3) policy generalization in offline RL problems, analyzing how these components affect performance. We make two surprising observations. First, we find that the choice of a policy extraction algorithm significantly affects the performance and scalability of offline RL, often more so than the value learning objective. For instance, we show that common value-weighted behavioral cloning objectives (e.g., AWR) do not fully leverage the learned value function, and switching to behavior-constrained policy gradient objectives (e.g., DDPG+BC) often leads to substantial improvements in performance and scalability. Second, we find that a big barrier to improving offline RL performance is often imperfect policy generalization on test-time states out of the support of the training data, rather than policy learning on in-distribution states. We then show that the use of suboptimal but high-coverage data or test-time policy training techniques can address this generalization issue in practice. Specifically, we propose two simple test-time policy improvement methods and show that these methods lead to better performance.
Authors: Luckeciano C. Melo, Panagiotis Tigas, Alessandro Abate, Yarin Gal
Abstract: Leveraging human preferences for steering the behavior of Large Language Models (LLMs) has demonstrated notable success in recent years. Nonetheless, data selection and labeling are still a bottleneck for these systems, particularly at large scale. Hence, selecting the most informative points for acquiring human feedback may considerably reduce the cost of preference labeling and unleash the further development of LLMs. Bayesian Active Learning provides a principled framework for addressing this challenge and has demonstrated remarkable success in diverse settings. However, previous attempts to employ it for Preference Modeling did not meet such expectations. In this work, we identify that naive epistemic uncertainty estimation leads to the acquisition of redundant samples. We address this by proposing the Bayesian Active Learner for Preference Modeling (BAL-PM), a novel stochastic acquisition policy that not only targets points of high epistemic uncertainty according to the preference model but also seeks to maximize the entropy of the acquired prompt distribution in the feature space spanned by the employed LLM. Notably, our experiments demonstrate that BAL-PM requires 33% to 68% fewer preference labels in two popular human preference datasets and exceeds previous stochastic Bayesian acquisition policies.
Authors: Samuele Bortolotti, Emanuele Marconato, Tommaso Carraro, Paolo Morettin, Emile van Krieken, Antonio Vergari, Stefano Teso, Andrea Passerini
Abstract: The advent of powerful neural classifiers has increased interest in problems that require both learning and reasoning. These problems are critical for understanding important properties of models, such as trustworthiness, generalization, interpretability, and compliance to safety and structural constraints. However, recent research observed that tasks requiring both learning and reasoning on background knowledge often suffer from reasoning shortcuts (RSs): predictors can solve the downstream reasoning task without associating the correct concepts to the high-dimensional data. To address this issue, we introduce rsbench, a comprehensive benchmark suite designed to systematically evaluate the impact of RSs on models by providing easy access to highly customizable tasks affected by RSs. Furthermore, rsbench implements common metrics for evaluating concept quality and introduces novel formal verification procedures for assessing the presence of RSs in learning tasks. Using rsbench, we highlight that obtaining high quality concepts in both purely neural and neuro-symbolic models is a far-from-solved problem. rsbench is available at: https://unitn-sml.github.io/rsbench.
Authors: Parand A. Alamdari, Yanshuai Cao, Kevin H. Wilson
Abstract: We present substantial evidence demonstrating the benefits of integrating Large Language Models (LLMs) with a Contextual Multi-Armed Bandit framework. Contextual bandits have been widely used in recommendation systems to generate personalized suggestions based on user-specific contexts. We show that LLMs, pre-trained on extensive corpora rich in human knowledge and preferences, can simulate human behaviours well enough to jump-start contextual multi-armed bandits to reduce online learning regret. We propose an initialization algorithm for contextual bandits by prompting LLMs to produce a pre-training dataset of approximate human preferences for the bandit. This significantly reduces online learning regret and data-gathering costs for training such models. Our approach is validated empirically through two sets of experiments with different bandit setups: one which utilizes LLMs to serve as an oracle and a real-world experiment utilizing data from a conjoint survey experiment.
Authors: Wenchen Han, Shay Vargaftik, Michael Mitzenmacher, Brad Karp, Ran Ben Basat
Abstract: Gradient aggregation has long been identified as a major bottleneck in today's large-scale distributed machine learning training systems. One promising solution to mitigate such bottlenecks is gradient compression, directly reducing communicated gradient data volume. However, in practice, many gradient compression schemes do not achieve acceleration of the training process while also preserving accuracy. In this work, we identify common issues in previous gradient compression systems and evaluation methodologies. These include excessive computational overheads; incompatibility with all-reduce; and insufficient evaluation methods, such as not using an end-to-end metric or using a 32-bit baseline instead of the stronger 16-bit baseline. We revisit common compression approaches (sparsification, quantization, and low-rank decomposition) and demonstrate how considering the above issues can lead to minor but strategic design changes, resulting in notably better performance. Our goal is to raise awareness of the need for design and evaluation standards that naturally translate to the end-to-end utility of gradient compression.
Authors: Frederic Z. Zhang, Paul Albert, Cristian Rodriguez-Opazo, Anton van den Hengel, Ehsan Abbasnejad
Abstract: Pre-trained models produce strong generic representations that can be adapted via fine-tuning. The learned weight difference relative to the pre-trained model, known as a task vector, characterises the direction and stride of fine-tuning. The significance of task vectors is such that simple arithmetic operations on them can be used to combine diverse representations from different domains. This paper builds on these properties of task vectors and aims to answer (1) whether components of task vectors, particularly parameter blocks, exhibit similar characteristics, and (2) how such blocks can be used to enhance knowledge composition and transfer. To this end, we introduce aTLAS, an algorithm that linearly combines parameter blocks with different learned coefficients, resulting in anisotropic scaling at the task vector level. We show that such linear combinations explicitly exploit the low intrinsic dimensionality of pre-trained models, with only a few coefficients being the learnable parameters. Furthermore, composition of parameter blocks leverages the already learned representations, thereby reducing the dependency on large amounts of data. We demonstrate the effectiveness of our method in task arithmetic, few-shot recognition and test-time adaptation, with supervised or unsupervised objectives. In particular, we show that (1) learned anisotropic scaling allows task vectors to be more disentangled, causing less interference in composition; (2) task vector composition excels with scarce or no labeled data and is less prone to domain shift, thus leading to better generalisability; (3) mixing the most informative parameter blocks across different task vectors prior to training can reduce the memory footprint and improve the flexibility of knowledge transfer. Moreover, we show the potential of aTLAS as a PEFT method, particularly with less data, and demonstrate its scalibility.
Authors: Amit Sinha, Matthieu Geist, Aditya Mahajan
Abstract: The standard approach for Partially Observable Markov Decision Processes (POMDPs) is to convert them to a fully observed belief-state MDP. However, the belief state depends on the system model and is therefore not viable in reinforcement learning (RL) settings. A widely used alternative is to use an agent state, which is a model-free, recursively updateable function of the observation history. Examples include frame stacking and recurrent neural networks. Since the agent state is model-free, it is used to adapt standard RL algorithms to POMDPs. However, standard RL algorithms like Q-learning learn a stationary policy. Our main thesis that we illustrate via examples is that because the agent state does not satisfy the Markov property, non-stationary agent-state based policies can outperform stationary ones. To leverage this feature, we propose PASQL (periodic agent-state based Q-learning), which is a variant of agent-state-based Q-learning that learns periodic policies. By combining ideas from periodic Markov chains and stochastic approximation, we rigorously establish that PASQL converges to a cyclic limit and characterize the approximation error of the converged periodic policy. Finally, we present a numerical experiment to highlight the salient features of PASQL and demonstrate the benefit of learning periodic policies over stationary policies.
Authors: Yuhan Li, Peisong Wang, Xiao Zhu, Aochuan Chen, Haiyun Jiang, Deng Cai, Victor Wai Kin Chan, Jia Li
Abstract: The emergence of large language models (LLMs) has revolutionized the way we interact with graphs, leading to a new paradigm called GraphLLM. Despite the rapid development of GraphLLM methods in recent years, the progress and understanding of this field remain unclear due to the lack of a benchmark with consistent experimental protocols. To bridge this gap, we introduce GLBench, the first comprehensive benchmark for evaluating GraphLLM methods in both supervised and zero-shot scenarios. GLBench provides a fair and thorough evaluation of different categories of GraphLLM methods, along with traditional baselines such as graph neural networks. Through extensive experiments on a collection of real-world datasets with consistent data processing and splitting strategies, we have uncovered several key findings. Firstly, GraphLLM methods outperform traditional baselines in supervised settings, with LLM-as-enhancers showing the most robust performance. However, using LLMs as predictors is less effective and often leads to uncontrollable output issues. We also notice that no clear scaling laws exist for current GraphLLM methods. In addition, both structures and semantics are crucial for effective zero-shot transfer, and our proposed simple baseline can even outperform several models tailored for zero-shot scenarios. The data and code of the benchmark can be found at https://github.com/NineAbyss/GLBench.
Authors: Jiahao Wu, Ning Lu, Zeiyu Dai, Wenqi Fan, Shengcai Liu, Qing Li, Ke Tang
Abstract: Recently, graph condensation has emerged as a prevalent technique to improve the training efficiency for graph neural networks (GNNs). It condenses a large graph into a small one such that a GNN trained on this small synthetic graph can achieve comparable performance to a GNN trained on the large graph. However, while existing graph condensation studies mainly focus on the best trade-off between graph size and the GNNs' performance (model utility), the security issues of graph condensation have not been studied. To bridge this research gap, we propose the task of backdoor graph condensation. Effective backdoor attacks on graph condensation aim to (1) maintain the quality and utility of condensed graphs despite trigger injections and (2) ensure trigger effectiveness through the condensation process, yielding a high attack success rate. To pursue the objectives, we devise the first backdoor attack against graph condensation, denoted as BGC, where effective attack is launched by consistently updating triggers throughout condensation and focusing on poisoning representative nodes. The extensive experiments demonstrate the effectiveness of our attack. BGC achieves a high attack success rate (close to 1.0) and good model utility in all cases. Furthermore, the results against multiple defense methods demonstrate BGC's resilience under their defenses. Finally, we conduct studies to analyze the factors that influence the attack performance.
Authors: Dengke Han, Mingyu Yan, Xiaochun Ye, Dongrui Fan
Abstract: Owing to their remarkable representation capabilities for heterogeneous graph data, Heterogeneous Graph Neural Networks (HGNNs) have been widely adopted in many critical real-world domains such as recommendation systems and medical analysis. Prior to their practical application, identifying the optimal HGNN model parameters tailored to specific tasks through extensive training is a time-consuming and costly process. To enhance the efficiency of HGNN training, it is essential to characterize and analyze the execution semantics and patterns within the training process to identify performance bottlenecks. In this study, we conduct an in-depth quantification and analysis of two mainstream HGNN training scenarios, including single-GPU and multi-GPU distributed training. Based on the characterization results, we disclose the performance bottlenecks and their underlying causes in different HGNN training scenarios and provide optimization guidelines from both software and hardware perspectives.
Authors: Linh Trinh, Ali Anwar, Siegfried Mercelis
Abstract: As the popularity of autonomous vehicles has grown, many standards and regulators, such as ISO, NHTSA, and Euro NCAP, require safety validation to ensure a sufficient level of safety before deploying them in the real world. Manufacturers gather a large amount of public road data for this purpose. However, the majority of these validation activities are done manually by humans. Furthermore, the data used to validate each driving feature may differ. As a result, it is essential to have an efficient data selection method that can be used flexibly and dynamically for verification and validation while also accelerating the validation process. In this paper, we present a data selection method that is practical, flexible, and efficient for assessment of autonomous vehicles. Our idea is to optimize the similarity between the metadata distribution of the selected data and a predefined metadata distribution that is expected for validation. Our experiments on the large dataset BDD100K show that our method can perform data selection tasks efficiently. These results demonstrate that our methods are highly reliable and can be used to select appropriate data for the validation of various safety functions.
Authors: Jerry Huang, Prasanna Parthasarathi, Mehdi Rezagholizadeh, Sarath Chandar
Abstract: Despite their widespread adoption, large language models (LLMs) remain prohibitive to use under resource constraints, with their ever growing sizes only increasing the barrier for use. One noted issue is the high latency associated with auto-regressive generation, rendering large LLMs use dependent on advanced computing infrastructure. Assisted decoding, where a smaller draft model guides a larger target model's generation, has helped alleviate this, but remains dependent on alignment between the two models. Thus if the draft model is insufficiently capable on some domain relative to the target model, performance can degrade. Alternatively, one can leverage multiple draft models to better cover the expertise of the target, but when multiple black-box draft models are available, selecting an assistant without details about its construction can be difficult. To better understand this decision making problem, we observe it as a contextual bandit, where a policy must choose a draft model based on a context. We show that even without prior knowledge of the draft models, creating an offline dataset from only outputs of independent draft/target models and training a policy over the alignment of these outputs can accelerate performance on multiple domains provided the candidates are effective. Further results show this to hold on various settings with multiple assisted decoding candidates, highlighting its flexibility and the advantageous role that such decision making can play.
Authors: Zhiyong Wang, Dongruo Zhou, John C. S. Lui, Wen Sun
Abstract: Learning a transition model via Maximum Likelihood Estimation (MLE) followed by planning inside the learned model is perhaps the most standard and simplest Model-based Reinforcement Learning (RL) framework. In this work, we show that such a simple Model-based RL scheme, when equipped with optimistic and pessimistic planning procedures, achieves strong regret and sample complexity bounds in online and offline RL settings. Particularly, we demonstrate that under the conditions where the trajectory-wise reward is normalized between zero and one and the transition is time-homogenous, it achieves nearly horizon-free and second-order bounds. Nearly horizon-free means that our bounds have no polynomial dependence on the horizon of the Markov Decision Process. A second-order bound is a type of instance-dependent bound that scales with respect to the variances of the returns of the policies which can be small when the system is nearly deterministic and (or) the optimal policy has small values. We highlight that our algorithms are simple, fairly standard, and indeed have been extensively studied in the RL literature: they learn a model via MLE, build a version space around the MLE solution, and perform optimistic or pessimistic planning depending on whether operating in the online or offline mode. These algorithms do not rely on additional specialized algorithmic designs such as learning variances and performing variance-weighted learning and thus can easily leverage non-linear function approximations. The simplicity of the algorithms also implies that our horizon-free and second-order regret analysis is actually standard and mainly follows the general framework of optimism/pessimism in the face of uncertainty.
Authors: Youdong Guo, Timothy E. Holy
Abstract: Non-negative matrix factorization (NMF) is a key technique for feature extraction and widely used in source separation. However, existing algorithms may converge to poor local minima, or to one of several minima with similar objective value but differing feature parametrizations. Here we show that some of these weaknesses may be mitigated by performing NMF in a higher-dimensional feature space and then iteratively combining components with an analytically-solvable pairwise merge strategy. Experimental results demonstrate our method helps non-ideal NMF solutions escape to better local optima and achieve greater consistency of the solutions. Despite these extra steps, our approach exhibits similar computational performance to established methods by reducing the occurrence of "plateau phenomenon" near saddle points. Moreover, the results also illustrate that our method is compatible with different NMF algorithms. Thus, this can be recommended as a preferred approach for most applications of NMF.
Authors: Xiaochen Wang, Jiaqi Wang, Houping Xiao, Jinghui Chen, Fenglong Ma
Abstract: Foundation models have demonstrated remarkable capabilities in handling diverse modalities and tasks, outperforming conventional artificial intelligence (AI) approaches that are highly task-specific and modality-reliant. In the medical domain, however, the development of comprehensive foundation models is constrained by limited access to diverse modalities and stringent privacy regulations. To address these constraints, this study introduces a novel knowledge injection approach, FedKIM, designed to scale the medical foundation model within a federated learning framework. FedKIM leverages lightweight local models to extract healthcare knowledge from private data and integrates this knowledge into a centralized foundation model using a designed adaptive Multitask Multimodal Mixture Of Experts (M3OE) module. This method not only preserves privacy but also enhances the model's ability to handle complex medical tasks involving multiple modalities. Our extensive experiments across twelve tasks in seven modalities demonstrate the effectiveness of FedKIM in various settings, highlighting its potential to scale medical foundation models without direct access to sensitive data.
Authors: Christian Schmid, James M. Murray
Abstract: The ability of a brain or a neural network to efficiently learn depends crucially on both the task structure and the learning rule. Previous works have analyzed the dynamical equations describing learning in the relatively simplified context of the perceptron under assumptions of a student-teacher framework or a linearized output. While these assumptions have facilitated theoretical understanding, they have precluded a detailed understanding of the roles of the nonlinearity and input-data distribution in determining the learning dynamics, limiting the applicability of the theories to real biological or artificial neural networks. Here, we use a stochastic-process approach to derive flow equations describing learning, applying this framework to the case of a nonlinear perceptron performing binary classification. We characterize the effects of the learning rule (supervised or reinforcement learning, SL/RL) and input-data distribution on the perceptron's learning curve and the forgetting curve as subsequent tasks are learned. In particular, we find that the input-data noise differently affects the learning speed under SL vs. RL, as well as determines how quickly learning of a task is overwritten by subsequent learning. Additionally, we verify our approach with real data using the MNIST dataset. This approach points a way toward analyzing learning dynamics for more-complex circuit architectures.
Authors: Shen Li, Yuyang Zhang, Zhaolin Ren, Claire Liang, Na Li, Julie A. Shah
Abstract: Interactive preference learning systems present humans with queries as pairs of options; humans then select their preferred choice, allowing the system to infer preferences from these binary choices. While binary choice feedback is simple and widely used, it offers limited information about preference strength. To address this, we leverage human response times, which inversely correlate with preference strength, as complementary information. We introduce a computationally efficient method based on the EZ-diffusion model, combining choices and response times to estimate the underlying human utility function. Theoretical and empirical comparisons with traditional choice-only estimators show that for queries where humans have strong preferences (i.e., "easy" queries), response times provide valuable complementary information and enhance utility estimates. We integrate this estimator into preference-based linear bandits for fixed-budget best-arm identification. Simulations on three real-world datasets demonstrate that incorporating response times significantly accelerates preference learning.
Authors: Sourav Chatterjee, Timothy Sudijono
Abstract: We show that feedforward neural networks with ReLU activation generalize on low complexity data, suitably defined. Given i.i.d. data generated from a simple programming language, the minimum description length (MDL) feedforward neural network which interpolates the data generalizes with high probability. We define this simple programming language, along with a notion of description length of such networks. We provide several examples on basic computational tasks, such as checking primality of a natural number, and more. For primality testing, our theorem shows the following. Suppose that we draw an i.i.d. sample of $\Theta(N^{\delta}\ln N)$ numbers uniformly at random from $1$ to $N$, where $\delta\in (0,1)$. For each number $x_i$, let $y_i = 1$ if $x_i$ is a prime and $0$ if it is not. Then with high probability, the MDL network fitted to this data accurately answers whether a newly drawn number between $1$ and $N$ is a prime or not, with test error $\leq O(N^{-\delta})$. Note that the network is not designed to detect primes; minimum description learning discovers a network which does so.
Authors: Vitaliy Kinakh, Slava Voloshynovskiy
Abstract: Generating synthetic tabular data is critical in machine learning, especially when real data is limited or sensitive. Traditional generative models often face challenges due to the unique characteristics of tabular data, such as mixed data types and varied distributions, and require complex preprocessing or large pretrained models. In this paper, we introduce a novel, lossless binary transformation method that converts any tabular data into fixed-size binary representations, and a corresponding new generative model called Binary Diffusion, specifically designed for binary data. Binary Diffusion leverages the simplicity of XOR operations for noise addition and removal and employs binary cross-entropy loss for training. Our approach eliminates the need for extensive preprocessing, complex noise parameter tuning, and pretraining on large datasets. We evaluate our model on several popular tabular benchmark datasets, demonstrating that Binary Diffusion outperforms existing state-of-the-art models on Travel, Adult Income, and Diabetes datasets while being significantly smaller in size. Code and models are available at: https://github.com/vkinakh/binary-diffusion-tabular
Authors: Zesen Zhao, Shuowei Jin, Z. Morley Mao
Abstract: The proliferation of Large Language Models (LLMs) with varying capabilities and costs has created a need for efficient model selection in AI systems. LLM routers address this need by dynamically choosing the most suitable model for a given query based on task requirements and budget constraints. However, existing routers face challenges in scalability and real-time adaptation, particularly in high-volume online environments. We present Eagle, a novel LLM routing approach that combines global and local ELO ranking modules to overcome these limitations. By evaluating both general and specialized LLM abilities, Eagle provides a scalable, training-free solution that enhances model selection quality while reducing computational overhead. Our experiments across multiple datasets show Eagle consistently outperforms baseline methods, with improvements of up to 23.52 percent in Area Under Curve (AUC) scores. Moreover, Eagle demonstrates remarkable efficiency, requiring only 1/20 of baseline methods' time for initialization and 100 to 200 times faster incremental updates in online scenarios, making it well-suited for dynamic, high-volume online serving environments.
Authors: Jie Hu, Yi-Ting Ma, Do Young Eun
Abstract: Distributed learning is essential to train machine learning algorithms across heterogeneous agents while maintaining data privacy. We conduct an asymptotic analysis of Unified Distributed SGD (UD-SGD), exploring a variety of communication patterns, including decentralized SGD and local SGD within Federated Learning (FL), as well as the increasing communication interval in the FL setting. In this study, we assess how different sampling strategies, such as i.i.d. sampling, shuffling, and Markovian sampling, affect the convergence speed of UD-SGD by considering the impact of agent dynamics on the limiting covariance matrix as described in the Central Limit Theorem (CLT). Our findings not only support existing theories on linear speedup and asymptotic network independence, but also theoretically and empirically show how efficient sampling strategies employed by individual agents contribute to overall convergence in UD-SGD. Simulations reveal that a few agents using highly efficient sampling can achieve or surpass the performance of the majority employing moderately improved strategies, providing new insights beyond traditional analyses focusing on the worst-performing agent.
Authors: Haoran Li, Zhennan Jiang, Yuhui Chen, Dongbin Zhao
Abstract: With high-dimensional state spaces, visual reinforcement learning (RL) faces significant challenges in exploitation and exploration, resulting in low sample efficiency and training stability. As a time-efficient diffusion model, although consistency models have been validated in online state-based RL, it is still an open question whether it can be extended to visual RL. In this paper, we investigate the impact of non-stationary distribution and the actor-critic framework on consistency policy in online RL, and find that consistency policy was unstable during the training, especially in visual RL with the high-dimensional state space. To this end, we suggest sample-based entropy regularization to stabilize the policy training, and propose a consistency policy with prioritized proximal experience regularization (CP3ER) to improve sample efficiency. CP3ER achieves new state-of-the-art (SOTA) performance in 21 tasks across DeepMind control suite and Meta-world. To our knowledge, CP3ER is the first method to apply diffusion/consistency models to visual RL and demonstrates the potential of consistency models in visual RL. More visualization results are available at https://jzndd.github.io/CP3ER-Page/.
Authors: Gabriel Franco, Mark Crovella
Abstract: Many papers have shown that attention heads work in conjunction with each other to perform complex tasks. It's frequently assumed that communication between attention heads is via the addition of specific features to token residuals. In this work we seek to isolate and identify the features used to effect communication and coordination among attention heads in GPT-2 small. Our key leverage on the problem is to show that these features are very often sparsely coded in the singular vectors of attention head matrices. We characterize the dimensionality and occurrence of these signals across the attention heads in GPT-2 small when used for the Indirect Object Identification (IOI) task. The sparse encoding of signals, as provided by attention head singular vectors, allows for efficient separation of signals from the residual background and straightforward identification of communication paths between attention heads. We explore the effectiveness of this approach by tracing portions of the circuits used in the IOI task. Our traces reveal considerable detail not present in previous studies, shedding light on the nature of redundant paths present in GPT-2. And our traces go beyond previous work by identifying features used to communicate between attention heads when performing IOI.
Authors: Rapha\"el Sarfati, Toni J. B. Liu, Nicolas Boull\'e, Christopher J. Earls
Abstract: Large Language Models achieve next-token prediction by transporting a vectorized piece of text (prompt) across an accompanying embedding space under the action of successive transformer layers. The resulting high-dimensional trajectories realize different contextualization, or 'thinking', steps, and fully determine the output probability distribution. We aim to characterize the statistical properties of ensembles of these 'lines of thought.' We observe that independent trajectories cluster along a low-dimensional, non-Euclidean manifold, and that their path can be well approximated by a stochastic equation with few parameters extracted from data. We find it remarkable that the vast complexity of such large models can be reduced to a much simpler form, and we reflect on implications.
Authors: Yan Scholten, Stephan G\"unnemann, Leo Schwinn
Abstract: Comprehensive evaluation of Large Language Models (LLMs) is an open research problem. Existing evaluations rely on deterministic point estimates generated via greedy decoding. However, we find that deterministic evaluations fail to capture the whole output distribution of a model, yielding inaccurate estimations of model capabilities. This is particularly problematic in critical contexts such as unlearning and alignment, where precise model evaluations are crucial. To remedy this, we introduce the first formal probabilistic evaluation framework in LLMs. Namely, we derive novel metrics with high-probability guarantees concerning the output distribution of a model. Our metrics are application-independent and allow practitioners to make more reliable estimates about model capabilities before deployment. Through a case study focused on unlearning, we reveal that deterministic evaluations falsely indicate successful unlearning, whereas our probabilistic evaluations demonstrate that most if not all of the supposedly unlearned information remains accessible in these models. Additionally, we propose a novel unlearning loss based on entropy optimization and adaptive temperature scaling, which significantly improves unlearning in probabilistic settings on recent benchmarks. Our proposed shift from point estimates to probabilistic evaluations of output distributions represents an important step toward comprehensive evaluations of LLMs. Code available at https://github.com/yascho/probabilistic-unlearning
Authors: Michael Ibrahim, Heraldo Rozas, Nagi Gebraeel, Weijun Xie
Abstract: The training of classification models for fault diagnosis tasks using geographically dispersed data is a crucial task for original equipment manufacturers (OEMs) seeking to provide long-term service contracts (LTSCs) to their customers. Due to privacy and bandwidth constraints, such models must be trained in a federated fashion. Moreover, due to harsh industrial settings the data often suffers from feature and label uncertainty. Therefore, we study the problem of training a distributionally robust (DR) support vector machine (SVM) in a federated fashion over a network comprised of a central server and $G$ clients without sharing data. We consider the setting where the local data of each client $g$ is sampled from a unique true distribution $\mathbb{P}_g$, and the clients can only communicate with the central server. We propose a novel Mixture of Wasserstein Balls (MoWB) ambiguity set that relies on local Wasserstein balls centered at the empirical distribution of the data at each client. We study theoretical aspects of the proposed ambiguity set, deriving its out-of-sample performance guarantees and demonstrating that it naturally allows for the separability of the DR problem. Subsequently, we propose two distributed optimization algorithms for training the global FDR-SVM: i) a subgradient method-based algorithm, and ii) an alternating direction method of multipliers (ADMM)-based algorithm. We derive the optimization problems to be solved by each client and provide closed-form expressions for the computations performed by the central server during each iteration for both algorithms. Finally, we thoroughly examine the performance of the proposed algorithms in a series of numerical experiments utilizing both simulation data and popular real-world datasets.
Authors: Hengyi Wang, Shiwei Tan, Zhiqing Hong, Desheng Zhang, Hao Wang
Abstract: Foundation Language Models (FLMs) such as BERT and its variants have achieved remarkable success in natural language processing. To date, the interpretability of FLMs has primarily relied on the attention weights in their self-attention layers. However, these attention weights only provide word-level interpretations, failing to capture higher-level structures, and are therefore lacking in readability and intuitiveness. To address this challenge, we first provide a formal definition of conceptual interpretation and then propose a variational Bayesian framework, dubbed VAriational Language Concept (VALC), to go beyond word-level interpretations and provide concept-level interpretations. Our theoretical analysis shows that our VALC finds the optimal language concepts to interpret FLM predictions. Empirical results on several real-world datasets show that our method can successfully provide conceptual interpretation for FLMs.
Authors: Kaiyue Wen, Zhiyuan Li, Jason Wang, David Hall, Percy Liang, Tengyu Ma
Abstract: Training language models currently requires pre-determining a fixed compute budget because the typical cosine learning rate schedule depends on the total number of steps. In contrast, the Warmup-Stable-Decay (WSD) schedule uses a constant learning rate to produce a main branch of iterates that can in principle continue indefinitely without a pre-specified compute budget. Then, given any compute budget, one can branch out from the main branch at a proper at any time with a rapidly decaying learning rate to produce a strong model. Empirically, WSD generates a non-traditional loss curve: the loss remains elevated during the stable phase but sharply declines during the decay phase. Towards explaining this phenomenon, we conjecture that pretraining loss exhibits a river valley landscape, which resembles a deep valley with a river at its bottom. Under this assumption, we show that during the stable phase, the iterate undergoes large oscillations due to the high learning rate, yet it progresses swiftly along the river. During the decay phase, the rapidly dropping learning rate minimizes the iterate's oscillations, moving it closer to the river and revealing true optimization progress. Therefore, the sustained high learning rate phase and fast decaying phase are responsible for progress in the river and the mountain directions respectively, and are both critical. Our analysis predicts phenomenons consistent with empirical observations and shows that this landscape can emerge from pretraining on a simple bi-gram dataset. Inspired by the theory, we introduce WSD-S, a variant of WSD that reuses previous checkpoints' decay phases and keeps only one main branch, where we resume from a decayed checkpoint. WSD-S empirically outperforms WSD and Cyclic-Cosine in obtaining multiple language model checkpoints across various compute budgets in a single run for parameters scaling from 0.1B to 1.2B.
Authors: Yun Zhu, Haizhou Shi, Xiaotang Wang, Yongchao Liu, Yaoke Wang, Boci Peng, Chuntao Hong, Siliang Tang
Abstract: Recently, research on Text-Attributed Graphs (TAGs) has gained significant attention due to the prevalence of free-text node features in real-world applications and the advancements in Large Language Models (LLMs) that bolster TAG methodologies. However, current TAG approaches face two primary challenges: (i) Heavy reliance on label information and (ii) Limited cross-domain zero/few-shot transferability. These issues constrain the scaling of both data and model size, owing to high labor costs and scaling laws, complicating the development of graph foundation models with strong transferability. In this work, we propose the GraphCLIP framework to address these challenges by learning graph foundation models with strong cross-domain zero/few-shot transferability through a self-supervised contrastive graph-summary pretraining method. Specifically, we generate and curate large-scale graph-summary pair data with the assistance of LLMs, and introduce a novel graph-summary pretraining method, combined with invariant learning, to enhance graph foundation models with strong cross-domain zero-shot transferability. For few-shot learning, we propose a novel graph prompt tuning technique aligned with our pretraining objective to mitigate catastrophic forgetting and minimize learning costs. Extensive experiments show the superiority of GraphCLIP in both zero-shot and few-shot settings, while evaluations across various downstream tasks confirm the versatility of GraphCLIP. Our code is available at: https://github.com/ZhuYun97/GraphCLIP
Authors: Mathieu Alain, So Takao, Xiaowen Dong, Bastian Rieck, Emmanuel Noutahi
Abstract: The problem of classifying graphs is ubiquitous in machine learning. While it is standard to apply graph neural networks or graph kernel methods, Gaussian processes can be employed by transforming spatial features from the graph domain into spectral features in the Euclidean domain, and using them as input points. However, this approach only takes into account features on vertices, whereas some graph datasets also support features on edges. In this work, we present a Gaussian process-based classification algorithm that can leverage one or both vertex and edges features. Furthermore, we take advantage of the Hodge decomposition to better capture the intricate richness of vertex and edge features, which can be beneficial on diverse tasks.
Authors: Bernhard Bermeitinger, Tomas Hrycej, Massimo Pavone, Julianus Kath, Siegfried Handschuh
Abstract: Transformers are a widespread and successful model architecture, particularly in Natural Language Processing (NLP) and Computer Vision (CV). The essential innovation of this architecture is the Attention Mechanism, which solves the problem of extracting relevant context information from long sequences in NLP and realistic scenes in CV. A classical neural network component, a Multi-Layer Perceptron (MLP), complements the attention mechanism. Its necessity is frequently justified by its capability of modeling nonlinear relationships. However, the attention mechanism itself is nonlinear through its internal use of similarity measures. A possible hypothesis is that this nonlinearity is sufficient for modeling typical application problems. As the MLPs usually contain the most trainable parameters of the whole model, their omission would substantially reduce the parameter set size. Further components can also be reorganized to reduce the number of parameters. Under some conditions, query and key matrices can be collapsed into a single matrix of the same size. The same is true about value and projection matrices, which can also be omitted without eliminating the substance of the attention mechanism. Initially, the similarity measure was defined asymmetrically, with peculiar properties such as that a token is possibly dissimilar to itself. A possible symmetric definition requires only half of the parameters. We have laid the groundwork by testing widespread CV benchmarks: MNIST and CIFAR-10. The tests have shown that simplified transformer architectures (a) without MLP, (b) with collapsed matrices, and (c) symmetric similarity matrices exhibit similar performance as the original architecture, saving up to 90% of parameters without hurting the classification performance.
Authors: Kushagra Pandey, Jaideep Pathak, Yilun Xu, Stephan Mandt, Michael Pritchard, Arash Vahdat, Morteza Mardani
Abstract: Diffusion models achieve state-of-the-art generation quality across many applications, but their ability to capture rare or extreme events in heavy-tailed distributions remains unclear. In this work, we show that traditional diffusion and flow-matching models with standard Gaussian priors fail to capture heavy-tailed behavior. We address this by repurposing the diffusion framework for heavy-tail estimation using multivariate Student-t distributions. We develop a tailored perturbation kernel and derive the denoising posterior based on the conditional Student-t distribution for the backward process. Inspired by $\gamma$-divergence for heavy-tailed distributions, we derive a training objective for heavy-tailed denoisers. The resulting framework introduces controllable tail generation using only a single scalar hyperparameter, making it easily tunable for diverse real-world distributions. As specific instantiations of our framework, we introduce t-EDM and t-Flow, extensions of existing diffusion and flow models that employ a Student-t prior. Remarkably, our approach is readily compatible with standard Gaussian diffusion models and requires only minimal code changes. Empirically, we show that our t-EDM and t-Flow outperform standard diffusion models in heavy-tail estimation on high-resolution weather datasets in which generating rare and extreme events is crucial.
Authors: Rhui Dih Lee, Laura Wynter
Abstract: We address the question of how to successively add new knowledge to an LLM whilst retaining previously-added knowledge. We consider two settings, semi-cooperative and fully-cooperative. Overall, LoRA performs better in most cases than full-fine tuning of all parameters when both new knowledge acquisition and retention of old, including recent, knowledge are taken into account. In the semi-cooperative setting, where datasets are not available after training, MOE mixing, model merging, and LoRA-based orthogonal subspace sequential learning, using a small weight on the orthogonality term, perform well. In the fully-cooperative setting where datasets remain available, joint training and sequential training with replay are both effective approaches with LoRA training generally preferable to full fine-tuning. The codes needed to reproduce the results are provided in an open source repository.
Authors: Heshan Fernando, Han Shen, Parikshit Ram, Yi Zhou, Horst Samulowitz, Nathalie Baracaldo, Tianyi Chen
Abstract: Post-training of pre-trained LLMs, which typically consists of the supervised fine-tuning (SFT) stage and the preference learning (RLHF or DPO) stage, is crucial to effective and safe LLM applications. The widely adopted approach in post-training popular open-source LLMs is to sequentially perform SFT and RLHF/DPO. However, sequential training is sub-optimal in terms of SFT and RLHF/DPO trade-off: the LLM gradually forgets about the first stage's training when undergoing the second stage's training. We theoretically prove the sub-optimality of sequential post-training. Furthermore, we propose a practical joint post-training framework with theoretical convergence guarantees and empirically outperforms sequential post-training framework, while having similar computational cost. Our code is available at https://github.com/heshandevaka/XRIGHT.
Authors: Jan Sp\"orer, Bernhard Bermeitinger, Tomas Hrycej, Niklas Limacher, Siegfried Handschuh
Abstract: In training neural networks, it is common practice to use partial gradients computed over batches, mostly very small subsets of the training set. This approach is motivated by the argument that such a partial gradient is close to the true one, with precision growing only with the square root of the batch size. A theoretical justification is with the help of stochastic approximation theory. However, the conditions for the validity of this theory are not satisfied in the usual learning rate schedules. Batch processing is also difficult to combine with efficient second-order optimization methods. This proposal is based on another hypothesis: the loss minimum of the training set can be expected to be well-approximated by the minima of its subsets. Such subset minima can be computed in a fraction of the time necessary for optimizing over the whole training set. This hypothesis has been tested with the help of the MNIST, CIFAR-10, and CIFAR-100 image classification benchmarks, optionally extended by training data augmentation. The experiments have confirmed that results equivalent to conventional training can be reached. In summary, even small subsets are representative if the overdetermination ratio for the given model parameter set sufficiently exceeds unity. The computing expense can be reduced to a tenth or less.
Authors: Jiachen Yao, Mayank Goswami, Chao Chen
Abstract: A prevalent assumption regarding real-world data is that it lies on or close to a low-dimensional manifold. When deploying a neural network on data manifolds, the required size, i.e., the number of neurons of the network, heavily depends on the intricacy of the underlying latent manifold. While significant advancements have been made in understanding the geometric attributes of manifolds, it's essential to recognize that topology, too, is a fundamental characteristic of manifolds. In this study, we investigate network expressive power in terms of the latent data manifold. Integrating both topological and geometric facets of the data manifold, we present a size upper bound of ReLU neural networks.
Authors: Arjun Subramonian, Samuel J. Bell, Levent Sagun, Elvis Dohmatob
Abstract: Machine learning models may capture and amplify biases present in data, leading to disparate test performance across social groups. To better understand, evaluate, and mitigate these possible biases, a deeper theoretical understanding of how model design choices and data distribution properties could contribute to bias is needed. In this work, we contribute a precise analytical theory in the context of ridge regression, both with and without random projections, where the former models neural networks in a simplified regime. Our theory offers a unified and rigorous explanation of machine learning bias, providing insights into phenomena such as bias amplification and minority-group bias in various feature and parameter regimes. For example, we demonstrate that there may be an optimal regularization penalty or training time to avoid bias amplification, and there can be fundamental differences in test error between groups that do not vanish with increased parameterization. Importantly, our theoretical predictions align with several empirical observations reported in the literature. We extensively empirically validate our theory on diverse synthetic and semi-synthetic datasets.
Authors: Yann Bouteiller, Karthik Soma, Giovanni Beltrame
Abstract: The universe involves many independent co-learning agents as an ever-evolving part of our observed environment. Yet, in practice, Multi-Agent Reinforcement Learning (MARL) applications are usually constrained to small, homogeneous populations and remain computationally intensive. In this paper, we study how large heterogeneous populations of learning agents evolve in normal-form games. We show how, under assumptions commonly made in the multi-armed bandit literature, Multi-Agent Policy Gradient closely resembles the Replicator Dynamic, and we further derive a fast, parallelizable implementation of Opponent-Learning Awareness tailored for evolutionary simulations. This enables us to simulate the evolution of very large populations made of heterogeneous co-learning agents, under both naive and advanced learning strategies. We demonstrate our approach in simulations of 200,000 agents, evolving in the classic games of Hawk-Dove, Stag-Hunt, and Rock-Paper-Scissors. Each game highlights distinct ways in which Opponent-Learning Awareness affects evolution.
Authors: Zhenqian Shen, Mingyang Zhou, Yongqi Zhang, Quanming Yao
Abstract: Predicting drug-drug interaction (DDI) plays an important role in pharmacology and healthcare for identifying potential adverse interactions and beneficial combination therapies between drug pairs. Recently, a flurry of graph learning methods have been introduced to predict drug-drug interactions. However, evaluating existing methods has several limitations, such as the absence of a unified comparison framework for DDI prediction methods, lack of assessments in meaningful real-world scenarios, and insufficient exploration of side information usage. In order to address these unresolved limitations in the literature, we propose a DDI prediction benchmark on graph learning. We first conduct unified evaluation comparison among existing methods. To meet realistic scenarios, we further evaluate the performance of different methods in settings with new drugs involved and examine the performance across different DDI types. Component analysis is conducted on the biomedical network to better utilize side information. Through this work, we hope to provide more insights for the problem of DDI prediction. Our implementation and data is open-sourced at https://anonymous.4open.science/r/DDI-Benchmark-ACD9/.
URLs: https://anonymous.4open.science/r/DDI-Benchmark-ACD9/.
Authors: Tuowei Wang, Ruwen Fan, Minxing Huang, Zixu Hao, Kun Li, Ting Cao, Youyou Lu, Yaoxue Zhang, Ju Ren
Abstract: Large Language Models (LLMs) have achieved remarkable success across various domains, yet deploying them on mobile devices remains an arduous challenge due to their extensive computational and memory demands. While lightweight LLMs have been developed to fit mobile environments, they suffer from degraded model accuracy. In contrast, sparsity-based techniques minimize DRAM usage by selectively transferring only relevant neurons to DRAM while retaining the full model in external storage, such as flash. However, such approaches are critically limited by numerous I/O operations, particularly on smartphones with severe IOPS constraints. In this paper, we propose Ripple, a novel approach that accelerates LLM inference on smartphones by optimizing neuron placement in flash memory. Ripple leverages the concept of Neuron Co-Activation, where neurons frequently activated together are linked to facilitate continuous read access and optimize data transfer efficiency. Our approach incorporates a two-stage solution: an offline stage that reorganizes neuron placement based on co-activation patterns, and an online stage that employs tailored data access and caching strategies to align well with hardware characteristics. Evaluations conducted on a variety of smartphones and LLMs demonstrate that Ripple achieves up to 5.93x improvements in I/O latency compared to the state-of-the-art. As the first solution to optimize storage placement under sparsity, Ripple explores a new optimization space at the intersection of sparsity-driven algorithm and storage-level system co-design in LLM inference.
Authors: Zhiyuan Pei, Jianqi Yan, Jin Yan, Bailing Yang, Ziyuan Li, Lin Zhang, Xin Liu, Yang Zhang
Abstract: Recently, deep learning in stock prediction has become an important branch. Image-based methods show potential by capturing complex visual patterns and spatial correlations, offering advantages in interpretability over time series models. However, image-based approaches are more prone to overfitting, hindering robust predictive performance. To improve accuracy, this paper proposes a novel method, named Sequence-based Multi-scale Fusion Regression Convolutional Neural Network (SMSFR-CNN), for predicting stock price movements in the China A-share market. By utilizing CNN to learn sequential features and combining them with image features, we improve the accuracy of stock trend prediction on the A-share market stock dataset. This approach reduces the search space for image features, stabilizes, and accelerates the training process. Extensive comparative experiments on 4,454 A-share stocks show that the model achieves a 61.15% positive predictive value and a 63.37% negative predictive value for the next 5 days, resulting in a total profit of 165.09%.
Authors: Mohamed Salim Aissi, Clement Romac, Thomas Carta, Sylvain Lamprier, Pierre-Yves Oudeyer, Olivier Sigaud, Laure Soulier, Nicolas Thome
Abstract: Reinforcement learning (RL) is a promising approach for aligning large language models (LLMs) knowledge with sequential decision-making tasks. However, few studies have thoroughly investigated the impact on LLM agents capabilities of fine-tuning them with RL in a specific environment. In this paper, we propose a novel framework to analyze the sensitivity of LLMs to prompt formulations following RL training in a textual environment. Our findings reveal that the performance of LLMs degrades when faced with prompt formulations different from those used during the RL training phase. Besides, we analyze the source of this sensitivity by examining the model's internal representations and salient tokens. Finally, we propose to use a contrastive loss to mitigate this sensitivity and improve the robustness and generalization capabilities of LLMs.
Authors: Amir Joudaki, Thomas Hofmann
Abstract: Understanding how neural networks transform input data across layers is fundamental to unraveling their learning and generalization capabilities. Although prior work has used insights from kernel methods to study neural networks, a global analysis of how the similarity between hidden representations evolves across layers remains underexplored. In this paper, we introduce a theoretical framework for the evolution of the kernel sequence, which measures the similarity between the hidden representation for two different inputs. Operating under the mean-field regime, we show that the kernel sequence evolves deterministically via a kernel map, which only depends on the activation function. By expanding activation using Hermite polynomials and using their algebraic properties, we derive an explicit form for kernel map and fully characterize its fixed points. Our analysis reveals that for nonlinear activations, the kernel sequence converges globally to a unique fixed point, which can correspond to orthogonal or similar representations depending on the activation and network architecture. We further extend our results to networks with residual connections and normalization layers, demonstrating similar convergence behaviors. This work provides new insights into the implicit biases of deep neural networks and how architectural choices influence the evolution of representations across layers.
Authors: Song Xia, Wenhan Yang, Yi Yu, Xun Lin, Henghui Ding, Lingyu Duan, Xudong Jiang
Abstract: The utilization of large foundational models has a dilemma: while fine-tuning downstream tasks from them holds promise for making use of the well-generalized knowledge in practical applications, their open accessibility also poses threats of adverse usage. This paper, for the first time, explores the feasibility of adversarial attacking various downstream models fine-tuned from the segment anything model (SAM), by solely utilizing the information from the open-sourced SAM. In contrast to prevailing transfer-based adversarial attacks, we demonstrate the existence of adversarial dangers even without accessing the downstream task and dataset to train a similar surrogate model. To enhance the effectiveness of the adversarial attack towards models fine-tuned on unknown datasets, we propose a universal meta-initialization (UMI) algorithm to extract the intrinsic vulnerability inherent in the foundation model, which is then utilized as the prior knowledge to guide the generation of adversarial perturbations. Moreover, by formulating the gradient difference in the attacking process between the open-sourced SAM and its fine-tuned downstream models, we theoretically demonstrate that a deviation occurs in the adversarial update direction by directly maximizing the distance of encoded feature embeddings in the open-sourced SAM. Consequently, we propose a gradient robust loss that simulates the associated uncertainty with gradient-based noise augmentation to enhance the robustness of generated adversarial examples (AEs) towards this deviation, thus improving the transferability. Extensive experiments demonstrate the effectiveness of the proposed universal meta-initialized and gradient robust adversarial attack (UMI-GRAT) toward SAMs and their downstream models. Code is available at https://github.com/xiasong0501/GRAT.
Authors: Juntong Shi, Minkai Xu, Harper Hua, Hengrui Zhang, Stefano Ermon, Jure Leskovec
Abstract: Synthesizing high-quality tabular data is an important topic in many data science tasks, ranging from dataset augmentation to privacy protection. However, developing expressive generative models for tabular data is challenging due to its inherent heterogeneous data types, complex inter-correlations, and intricate column-wise distributions. In this paper, we introduce TabDiff, a joint diffusion framework that models all multi-modal distributions of tabular data in one model. Our key innovation is the development of a joint continuous-time diffusion process for numerical and categorical data, where we propose feature-wise learnable diffusion processes to counter the high disparity of different feature distributions. TabDiff is parameterized by a transformer handling different input types, and the entire framework can be efficiently optimized in an end-to-end fashion. We further introduce a multi-modal stochastic sampler to automatically correct the accumulated decoding error during sampling, and propose classifier-free guidance for conditional missing column value imputation. Comprehensive experiments on seven datasets demonstrate that TabDiff achieves superior average performance over existing competitive baselines across all eight metrics, with up to $22.5\%$ improvement over the state-of-the-art model on pair-wise column correlation estimations. Code is available at https://github.com/MinkaiXu/TabDiff.
Authors: Ruifeng Li, Wei Liu, Xiangxin Zhou, Mingqian Li, Qiang Zhang, Hongyang Chen, Xuemin Lin
Abstract: In the drug discovery process, the low success rate of drug candidate screening often leads to insufficient labeled data, causing the few-shot learning problem in molecular property prediction. Existing methods for few-shot molecular property prediction overlook the sample selection bias, which arises from non-random sample selection in chemical experiments. This bias in data representativeness leads to suboptimal performance. To overcome this challenge, we present a novel method named contextual representation anchor Network (CRA), where an anchor refers to a cluster center of the representations of molecules and serves as a bridge to transfer enriched contextual knowledge into molecular representations and enhance their expressiveness. CRA introduces a dual-augmentation mechanism that includes context augmentation, which dynamically retrieves analogous unlabeled molecules and captures their task-specific contextual knowledge to enhance the anchors, and anchor augmentation, which leverages the anchors to augment the molecular representations. We evaluate our approach on the MoleculeNet and FS-Mol benchmarks, as well as in domain transfer experiments. The results demonstrate that CRA outperforms the state-of-the-art by 2.60% and 3.28% in AUC and $\Delta$AUC-PR metrics, respectively, and exhibits superior generalization capabilities.
Authors: Azizjon Azimi, Bonu Boboeva, Ilyas Varshavskiy, Shuhrat Khalilbekov, Akhlitdin Nizamitdinov, Najima Noyoftova, Sergey Shulgin
Abstract: The phenomenon of "black swans" has posed a fundamental challenge to performance of classical machine learning models. The perceived rise in frequency of outlier conditions, especially in post-pandemic environment, has necessitated exploration of synthetic data as a complement to real data in model training. This article provides a general overview and experimental investigation of the zGAN model architecture developed for the purpose of generating synthetic tabular data with outlier characteristics. The model is put to test in binary classification environments and shows promising results on realistic synthetic data generation, as well as uplift capabilities vis-\`a-vis model performance. A distinctive feature of zGAN is its enhanced correlation capability between features in the generated data, replicating correlations of features in real training data. Furthermore, crucial is the ability of zGAN to generate outliers based on covariance of real data or synthetically generated covariances. This approach to outlier generation enables modeling of complex economic events and augmentation of outliers for tasks such as training predictive models and detecting, processing or removing outliers. Experiments and comparative analyses as part of this study were conducted on both private (credit risk in financial services) and public datasets.
Authors: Xiaoqi Ling, Cheng Cai, Demin Kong, Zhisheng Wei, Jing Wu, Lei Wang, Zhaohong Deng
Abstract: Computational protein design (CPD) refers to the use of computational methods to design proteins. Traditional methods relying on energy functions and heuristic algorithms for sequence design are inefficient and do not meet the demands of the big data era in biomolecules, with their accuracy limited by the energy functions and search algorithms. Existing deep learning methods are constrained by the learning capabilities of the networks, failing to extract effective information from sparse protein structures, which limits the accuracy of protein design. To address these shortcomings, we developed an Efficient attention-based Models for Computational Protein Design using amino acid microenvironment (EMOCPD). It aims to predict the category of each amino acid in a protein by analyzing the three-dimensional atomic environment surrounding the amino acids, and optimize the protein based on the predicted high-probability potential amino acid categories. EMOCPD employs a multi-head attention mechanism to focus on important features in the sparse protein microenvironment and utilizes an inverse residual structure to optimize the network architecture. The proposed EMOCPD achieves over 80% accuracy on the training set and 68.33% and 62.32% accuracy on two independent test sets, respectively, surpassing the best comparative methods by over 10%. In protein design, the thermal stability and protein expression of the predicted mutants from EMOCPD show significant improvements compared to the wild type, effectively validating EMOCPD's potential in designing superior proteins. Furthermore, the predictions of EMOCPD are influenced positively, negatively, or have minimal impact based on the content of the 20 amino acids, categorizing amino acids as positive, negative, or neutral. Research findings indicate that EMOCPD is more suitable for designing proteins with lower contents of negative amino acids.
Authors: Maya Ramchandran, Rajarshi Mukherjee, Giovanni Parmigiani
Abstract: Adapting machine learning algorithms to better handle the presence of clusters or batch effects within training datasets is important across a wide variety of biological applications. This article considers the effect of ensembling Random Forest learners trained on clusters within a single dataset with heterogeneity in the distribution of the features. We find that constructing ensembles of forests trained on clusters determined by algorithms such as k-means results in significant improvements in accuracy and generalizability over the traditional Random Forest algorithm. We begin with a theoretical exploration of the benefits of our novel approach, denoted as the Cross-Cluster Weighted Forest, and subsequently empirically examine its robustness to various data-generating scenarios and outcome models. Furthermore, we explore the influence of the data-partitioning and ensemble weighting strategies the benefits of our method over the existing paradigm. Finally, we apply our approach to cancer molecular profiling and gene expression datasets that are naturally divisible into clusters and illustrate that our approach outperforms the classic Random Forest. The code and supplementary material are available at https://github.com/m-ramchandran/cross-cluster.
Authors: Xinjie Xu, Zhuangzhi Chen, Dongwei Xu, Huaji Zhou, Shanqing Yu, Shilian Zheng, Qi Xuan, Xiaoniu Yang
Abstract: With the rapid development of deep learning, automatic modulation recognition (AMR), as an important task in cognitive radio, has gradually transformed from traditional feature extraction and classification to automatic classification by deep learning technology. However, deep learning models are data-driven methods, which often require a large amount of data as the training support. Data augmentation, as the strategy of expanding dataset, can improve the generalization of the deep learning models and thus improve the accuracy of the models to a certain extent. In this paper, for AMR of radio signals, we propose a data augmentation strategy based on mixing signals and consider four specific methods (Random Mixing, Maximum-Similarity-Mixing, $\theta-$Similarity Mixing and n-times Random Mixing) to achieve data augmentation. Experiments show that our proposed method can improve the classification accuracy of deep learning based AMR models in the full public dataset RML2016.10a. In particular, for the case of a single signal-to-noise ratio signal set, the classification accuracy can be significantly improved, which verifies the effectiveness of the methods.
Authors: Aleksandr Talitckii, Brendon K. Colbert, Matthew M. Peet
Abstract: The accuracy and complexity of machine learning algorithms based on kernel optimization are determined by the set of kernels over which they are able to optimize. An ideal set of kernels should: admit a linear parameterization (for tractability); be dense in the set of all kernels (for robustness); be universal (for accuracy). Recently, a framework was proposed for using positive matrices to parameterize a class of positive semi-separable kernels. Although this class can be shown to meet all three criteria, previous algorithms for optimization of such kernels were limited to classification and furthermore relied on computationally complex Semidefinite Programming (SDP) algorithms. In this paper, we pose the problem of learning semiseparable kernels as a minimax optimization problem and propose a SVD-QCQP primal-dual algorithm which dramatically reduces the computational complexity as compared with previous SDP-based approaches. Furthermore, we provide an efficient implementation of this algorithm for both classification and regression -- an implementation which enables us to solve problems with 100 features and up to 30,000 datums. Finally, when applied to benchmark data, the algorithm demonstrates the potential for significant improvement in accuracy over typical (but non-convex) approaches such as Neural Nets and Random Forest with similar or better computation time.
Authors: Hakjin Lee, Minki Song, Jamyoung Koo, Junghoon Seo
Abstract: Detection Transformers (DETR) have recently set new benchmarks in object detection. However, their performance in detecting rotated objects lags behind established oriented object detectors. Our analysis identifies a key observation: the boundary discontinuity and square-like problem in bipartite matching poses an issue with assigning appropriate ground truths to predictions, leading to duplicate low-confidence predictions. To address this, we introduce a Hausdorff distance-based cost for bipartite matching, which more accurately quantifies the discrepancy between predictions and ground truths. Additionally, we find that a static denoising approach impedes the training of rotated DETR, especially as the quality of the detector's predictions begins to exceed that of the noised ground truths. To overcome this, we propose an adaptive query denoising method that employs bipartite matching to selectively eliminate noised queries that detract from model improvement. When compared to models adopting a ResNet-50 backbone, our proposed model yields remarkable improvements, achieving $\textbf{+4.18}$ AP$_{50}$, $\textbf{+4.59}$ AP$_{50}$, and $\textbf{+4.99}$ AP$_{50}$ on DOTA-v2.0, DOTA-v1.5, and DIOR-R, respectively.
Authors: Marco Paul E. Apolinario, Kaushik Roy
Abstract: Spiking Neural Networks (SNNs) are biologically plausible models that have been identified as potentially apt for deploying energy-efficient intelligence at the edge, particularly for sequential learning tasks. However, training of SNNs poses significant challenges due to the necessity for precise temporal and spatial credit assignment. Back-propagation through time (BPTT) algorithm, whilst the most widely used method for addressing these issues, incurs high computational cost due to its temporal dependency. In this work, we propose S-TLLR, a novel three-factor temporal local learning rule inspired by the Spike-Timing Dependent Plasticity (STDP) mechanism, aimed at training deep SNNs on event-based learning tasks. Furthermore, S-TLLR is designed to have low memory and time complexities, which are independent of the number of time steps, rendering it suitable for online learning on low-power edge devices. To demonstrate the scalability of our proposed method, we have conducted extensive evaluations on event-based datasets spanning a wide range of applications, such as image and gesture recognition, audio classification, and optical flow estimation. In all the experiments, S-TLLR achieved high accuracy, comparable to BPTT, with a reduction in memory between $5-50\times$ and multiply-accumulate (MAC) operations between $1.3-6.6\times$.
Authors: Keiichi Tamai, Tsuyoshi Okubo, Truong Vinh Truong Duy, Naotake Natori, Synge Todo
Abstract: We demonstrate that conventional artificial deep neural networks operating near the phase boundary of the signal propagation dynamics, also known as the edge of chaos, exhibit universal scaling laws of absorbing phase transitions in non-equilibrium statistical mechanics. Our numerical results indicate that the multilayer perceptrons and the convolutional neural networks belong to the mean-field and the directed percolation universality classes, respectively. Also, the finite-size scaling is successfully applied, suggesting a potential connection to the depth-width trade-off in deep learning. Furthermore, our analysis of the training dynamics under gradient descent reveals that hyperparameter tuning to the phase boundary is necessary but insufficient for achieving optimal generalization in deep networks. Remarkably, nonuniversal metric factors associated with the scaling laws are shown to play a significant role in concretizing the above observations. These findings highlight the usefulness of the notion of criticality for analyzing the behavior of artificial deep neural networks and offer new insights toward a unified understanding of an essential relationship between criticality and intelligence.
Authors: Shoji Toyota, Kenji Fukumizu
Abstract: Deep Neural Networks often inherit spurious correlations embedded in training data and hence may fail to generalize to unseen domains, which have different distributions from the domain to provide training data. M. Arjovsky et al. (2019) introduced the concept out-of-distribution (o.o.d.) risk, which is the maximum risk among all domains, and formulated the issue caused by spurious correlations as a minimization problem of the o.o.d. risk. Invariant Risk Minimization (IRM) is considered to be a promising approach to minimize the o.o.d. risk: IRM estimates a minimum of the o.o.d. risk by solving a bi-level optimization problem. While IRM has attracted considerable attention with empirical success, it comes with few theoretical guarantees. Especially, a solid theoretical guarantee that the bi-level optimization problem gives the minimum of the o.o.d. risk has not yet been established. Aiming at providing a theoretical justification for IRM, this paper rigorously proves that a solution to the bi-level optimization problem minimizes the o.o.d. risk under certain conditions. The result also provides sufficient conditions on distributions providing training data and on a dimension of feature space for the bi-leveled optimization problem to minimize the o.o.d. risk.
Authors: Mihir Prabhudesai, Anirudh Goyal, Deepak Pathak, Katerina Fragkiadaki
Abstract: Text-to-image diffusion models have recently emerged at the forefront of image generation, powered by very large-scale unsupervised or weakly supervised text-to-image training datasets. Due to their unsupervised training, controlling their behavior in downstream tasks, such as maximizing human-perceived image quality, image-text alignment, or ethical image generation, is difficult. Recent works finetune diffusion models to downstream reward functions using vanilla reinforcement learning, notorious for the high variance of the gradient estimators. In this paper, we propose AlignProp, a method that aligns diffusion models to downstream reward functions using end-to-end backpropagation of the reward gradient through the denoising process. While naive implementation of such backpropagation would require prohibitive memory resources for storing the partial derivatives of modern text-to-image models, AlignProp finetunes low-rank adapter weight modules and uses gradient checkpointing, to render its memory usage viable. We test AlignProp in finetuning diffusion models to various objectives, such as image-text semantic alignment, aesthetics, compressibility and controllability of the number of objects present, as well as their combinations. We show AlignProp achieves higher rewards in fewer training steps than alternatives, while being conceptually simpler, making it a straightforward choice for optimizing diffusion models for differentiable reward functions of interest. Code and Visualization results are available at https://align-prop.github.io/.
Authors: Saptarshi Roy, Zehua Wang, Ambuj Tewari
Abstract: We consider the problem of model selection in a high-dimensional sparse linear regression model under privacy constraints. We propose a differentially private (DP) best subset selection method with strong statistical utility properties by adopting the well-known exponential mechanism for selecting the best model. To achieve computational expediency, we propose an efficient Metropolis-Hastings algorithm and under certain regularity conditions, we establish that it enjoys polynomial mixing time to its stationary distribution. As a result, we also establish both approximate differential privacy and statistical utility for the estimates of the mixed Metropolis-Hastings chain. Finally, we perform some illustrative experiments on simulated data showing that our algorithm can quickly identify active features under reasonable privacy budget constraints.
Authors: Sagi Shaier, Lawrence E. Hunter, Katharina von der Wense
Abstract: Both standalone language models (LMs) as well as LMs within downstream-task systems have been shown to generate statements which are factually untrue. This problem is especially severe for low-resource languages, where training data is scarce and of worse quality than for high-resource languages. In this opinion piece, we argue that LMs in their current state will never be fully trustworthy in critical settings and suggest a possible novel strategy to handle this issue: by building LMs such that can cite their sources - i.e., point a user to the parts of their training data that back up their outputs. We first discuss which current NLP tasks would or would not benefit from such models. We then highlight the expected benefits such models would bring, e.g., quick verifiability of statements. We end by outlining the individual tasks that would need to be solved on the way to developing LMs with the ability to cite. We hope to start a discussion about the field's current approach to building LMs, especially for low-resource languages, and the role of the training data in explaining model generations.
Authors: Dimitrios Bachtis
Abstract: We propose inverse renormalization group transformations to construct approximate configurations for lattice volumes that have not yet been accessed by supercomputers or large-scale simulations in the study of spin glasses. Specifically, starting from lattices of volume $V=8^{3}$ in the case of the three-dimensional Edwards-Anderson model we employ machine learning algorithms to construct rescaled lattices up to $V'=128^{3}$, which we utilize to extract two critical exponents. We conclude by discussing how to incorporate numerical exactness within inverse renormalization group methods of disordered systems, thus opening up the opportunity to explore a sustainable and energy-efficient generation of exact configurations for increasing lattice volumes without the use of dedicated supercomputers.
Authors: Sofia Ek, Dave Zachariah
Abstract: Randomized trials are widely considered as the gold standard for evaluating the effects of decision policies. Trial data is, however, drawn from a population which may differ from the intended target population and this raises a problem of external validity (aka. generalizability). In this paper we seek to use trial data to draw valid inferences about the outcome of a policy on the target population. Additional covariate data from the target population is used to model the sampling of individuals in the trial study. We develop a method that yields certifiably valid trial-based policy evaluations under any specified range of model miscalibrations. The method is nonparametric and the validity is assured even with finite samples. The certified policy evaluations are illustrated using both simulated and real data.
Authors: Alessandro Scagliotti, Sara Farinelli
Abstract: The term "Normalizing Flows" is related to the task of constructing invertible transport maps between probability measures by means of deep neural networks. In this paper, we consider the problem of recovering the $W_2$-optimal transport map $T$ between absolutely continuous measures $\mu,\nu\in\mathcal{P}(\mathbb{R}^n)$ as the flow of a linear-control neural ODE, where the control depends only on the time variable and takes values in a finite-dimensional space. We first show that, under suitable assumptions on $\mu,\nu$ and on the controlled vector fields, the optimal transport map is contained in the $C^0_c$-closure of the flows generated by the system. Assuming that discrete approximations $\mu_N,\nu_N$ of the original measures $\mu,\nu$ are available, we use a discrete optimal coupling $\gamma_N$ to define an optimal control problem. With a $\Gamma$-convergence argument, we prove that its solutions correspond to flows that approximate the optimal transport map $T$. Finally, taking advantage of the Pontryagin Maximum Principle, we propose an iterative numerical scheme for the resolution of the optimal control problem, resulting in an algorithm for the practical computation of the approximated optimal transport map.
Authors: Zengzhi Wang, Xuefeng Li, Rui Xia, Pengfei Liu
Abstract: High-quality, large-scale corpora are the cornerstone of building foundation models. In this work, we introduce MathPile, a diverse and high-quality math-centric corpus comprising about 9.5 billion tokens. Throughout its creation, we adhered to the principle of "less is more", firmly believing in the supremacy of data quality over quantity, even in the pre-training phase. Our meticulous data collection and processing efforts included a complex suite of preprocessing, prefiltering, language identification, cleaning, filtering, and deduplication, ensuring the high quality of our corpus. Furthermore, we performed data contamination detection on downstream benchmark test sets to eliminate duplicates and conducted continual pre-training experiments, booting the performance on common mathematical reasoning benchmarks. We aim for our MathPile to boost language models' mathematical reasoning abilities and open-source its different versions and processing scripts to advance the field.
Authors: Jacob M. Chen, Rohit Bhattacharya, Katherine A. Keith
Abstract: Recent text-based causal methods attempt to mitigate confounding bias by estimating proxies of confounding variables that are partially or imperfectly measured from unstructured text data. These approaches, however, assume analysts have supervised labels of the confounders given text for a subset of instances, a constraint that is sometimes infeasible due to data privacy or annotation costs. In this work, we address settings in which an important confounding variable is completely unobserved. We propose a new causal inference method that uses two instances of pre-treatment text data, infers two proxies using two zero-shot models on the separate instances, and applies these proxies in the proximal g-formula. We prove, under certain assumptions about the instances of text and accuracy of the zero-shot predictions, that our method of inferring text-based proxies satisfies identification conditions of the proximal g-formula while other seemingly reasonable proposals do not. To address untestable assumptions associated with our method and the proximal g-formula, we further propose an odds ratio falsification heuristic that flags when to proceed with downstream effect estimation using the inferred proxies. We evaluate our method in synthetic and semi-synthetic settings -- the latter with real-world clinical notes from MIMIC-III and open large language models for zero-shot prediction -- and find that our method produces estimates with low bias. We believe that this text-based design of proxies allows for the use of proximal causal inference in a wider range of scenarios, particularly those for which obtaining suitable proxies from structured data is difficult.
Authors: Bargav Jayaraman, Chuan Guo, Kamalika Chaudhuri
Abstract: Vision-Language Models (VLMs) have emerged as the state-of-the-art representation learning solution, with myriads of downstream applications such as image classification, retrieval and generation. A natural question is whether these models memorize their training data, which also has implications for generalization. We propose a new method for measuring memorization in VLMs, which we call d\'ej\`a vu memorization. For VLMs trained on image-caption pairs, we show that the model indeed retains information about individual objects in the training images beyond what can be inferred from correlations or the image caption. We evaluate d\'ej\`a vu memorization at both sample and population level, and show that it is significant for OpenCLIP trained on as many as 50M image-caption pairs. Finally, we show that text randomization considerably mitigates memorization while only moderately impacting the model's downstream task performance.
Authors: Yifan Zhang, Yifan Luo, Yang Yuan, Andrew Chi-Chih Yao
Abstract: To improve language models' proficiency in mathematical reasoning via continual pretraining, we introduce a novel strategy that leverages base language models for autonomous data selection. Departing from conventional supervised fine-tuning or trained classifiers with human-annotated data, our approach Autonomous Data Selection (AutoDS) utilizes meta-prompted language models as zero-shot verifiers to evaluate and select high-quality mathematical content autonomously. To demonstrate the efficacy of our method, we continuously pretrained a 7B-parameter language model on our curated dataset, achieving substantial improvements in downstream performance on the MATH, GSM8K, and BIG-Bench Hard (BBH) tasks with a token amount reduced by orders of magnitude compared to previous continual pretraining works. Our method showcases a 2 times increase in pretraining token efficiency compared to state-of-the-art baselines, underscoring the potential of our approach in enhancing models' mathematical reasoning capabilities. The AutoMathText dataset is available at https://huggingface.co/datasets/math-ai/AutoMathText. The code is available at https://github.com/yifanzhang-pro/AutoMathText.
URLs: https://huggingface.co/datasets/math-ai/AutoMathText., https://github.com/yifanzhang-pro/AutoMathText.
Authors: Cheng Li, Mengzhou Chen, Jindong Wang, Sunayana Sitaram, Xing Xie
Abstract: Large language models (LLMs) are reported to be partial to certain cultures owing to the training data dominance from the English corpora. Since multilingual cultural data are often expensive to collect, existing efforts handle this by prompt engineering or culture-specific pre-training. However, they might overlook the knowledge deficiency of low-resource culture and require extensive computing resources. In this paper, we propose CultureLLM, a cost-effective solution to incorporate cultural differences into LLMs. CultureLLM adopts World Value Survey (WVS) as seed data and generates semantically equivalent training data via the proposed semantic data augmentation. Using only 50 seed samples from WVS with augmented data, we fine-tune culture-specific LLMs and one unified model (CultureLLM-One) for 9 cultures covering rich and low-resource languages. Extensive experiments on 60 culture-related datasets demonstrate that CultureLLM significantly outperforms various counterparts such as GPT-3.5 (by 8.1%) and Gemini Pro (by 9.5%) with comparable performance to GPT-4 or even better. Our human study shows that the generated samples are semantically equivalent to the original samples, providing an effective solution for LLMs augmentation. Code is released at https://github.com/Scarelette/CultureLLM.
Authors: Alberto Abadie, Anish Agarwal, Raaz Dwivedi, Abhin Shah
Abstract: This article introduces a new estimator of average treatment effects under unobserved confounding in modern data-rich environments featuring large numbers of units and outcomes. The proposed estimator is doubly robust, combining outcome imputation, inverse probability weighting, and a novel cross-fitting procedure for matrix completion. We derive finite-sample and asymptotic guarantees, and show that the error of the new estimator converges to a mean-zero Gaussian distribution at a parametric rate. Simulation results demonstrate the relevance of the formal properties of the estimators analyzed in this article.
Authors: Marcel Torne, Anthony Simeonov, Zechu Li, April Chan, Tao Chen, Abhishek Gupta, Pulkit Agrawal
Abstract: Imitation learning methods need significant human supervision to learn policies robust to changes in object poses, physical disturbances, and visual distractors. Reinforcement learning, on the other hand, can explore the environment autonomously to learn robust behaviors but may require impractical amounts of unsafe real-world data collection. To learn performant, robust policies without the burden of unsafe real-world data collection or extensive human supervision, we propose RialTo, a system for robustifying real-world imitation learning policies via reinforcement learning in "digital twin" simulation environments constructed on the fly from small amounts of real-world data. To enable this real-to-sim-to-real pipeline, RialTo proposes an easy-to-use interface for quickly scanning and constructing digital twins of real-world environments. We also introduce a novel "inverse distillation" procedure for bringing real-world demonstrations into simulated environments for efficient fine-tuning, with minimal human intervention and engineering required. We evaluate RialTo across a variety of robotic manipulation problems in the real world, such as robustly stacking dishes on a rack, placing books on a shelf, and six other tasks. RialTo increases (over 67%) in policy robustness without requiring extensive human data collection. Project website and videos at https://real-to-sim-to-real.github.io/RialTo/
Authors: Xin Zhu, Hongyi Pan, Yury Velichko, Adam B. Murphy, Ashley Ross, Baris Turkbey, Ahmet Enis Cetin, Ulas Bagci
Abstract: Magnetic field inhomogeneity correction remains a challenging task in MRI analysis. Most established techniques are designed for brain MRI by supposing that image intensities in the identical tissue follow a uniform distribution. Such an assumption cannot be easily applied to other organs, especially those that are small in size and heterogeneous in texture (large variations in intensity), such as the prostate. To address this problem, this paper proposes a probabilistic Hadamard U-Net (PHU-Net) for prostate MRI bias field correction. First, a novel Hadamard U-Net (HU-Net) is introduced to extract the low-frequency scalar field, multiplied by the original input to obtain the prototypical corrected image. HU-Net converts the input image from the time domain into the frequency domain via Hadamard transform. In the frequency domain, high-frequency components are eliminated using the trainable filter (scaling layer), hard-thresholding layer, and sparsity penalty. Next, a conditional variational autoencoder is used to encode possible bias field-corrected variants into a low-dimensional latent space. Random samples drawn from latent space are then incorporated with a prototypical corrected image to generate multiple plausible images. Experimental results demonstrate the effectiveness of PHU-Net in correcting bias-field in prostate MRI with a fast inference speed. It has also been shown that prostate MRI segmentation accuracy improves with the high-quality corrected images from PHU-Net. The code will be available in the final version of this manuscript.
Authors: Thomas Melistas, Nikos Spyrou, Nefeli Gkouti, Pedro Sanchez, Athanasios Vlontzos, Yannis Panagakis, Giorgos Papanastasiou, Sotirios A. Tsaftaris
Abstract: Generative AI has revolutionised visual content editing, empowering users to effortlessly modify images and videos. However, not all edits are equal. To perform realistic edits in domains such as natural image or medical imaging, modifications must respect causal relationships inherent to the data generation process. Such image editing falls into the counterfactual image generation regime. Evaluating counterfactual image generation is substantially complex: not only it lacks observable ground truths, but also requires adherence to causal constraints. Although several counterfactual image generation methods and evaluation metrics exist, a comprehensive comparison within a unified setting is lacking. We present a comparison framework to thoroughly benchmark counterfactual image generation methods. We integrate all models that have been used for the task at hand and expand them to novel datasets and causal graphs, demonstrating the superiority of Hierarchical VAEs across most datasets and metrics. Our framework is implemented in a user-friendly Python package that can be extended to incorporate additional SCMs, causal methods, generative models, and datasets for the community to build on. Code: https://github.com/gulnazaki/counterfactual-benchmark.
URLs: https://github.com/gulnazaki/counterfactual-benchmark.
Authors: Vishaal Udandarao, Ameya Prabhu, Adhiraj Ghosh, Yash Sharma, Philip H. S. Torr, Adel Bibi, Samuel Albanie, Matthias Bethge
Abstract: Web-crawled pretraining datasets underlie the impressive "zero-shot" evaluation performance of multimodal models, such as CLIP for classification/retrieval and Stable-Diffusion for image generation. However, it is unclear how meaningful the notion of "zero-shot" generalization is for such multimodal models, as it is not known to what extent their pretraining datasets encompass the downstream concepts targeted for during "zero-shot" evaluation. In this work, we ask: How is the performance of multimodal models on downstream concepts influenced by the frequency of these concepts in their pretraining datasets? We comprehensively investigate this question across 34 models and five standard pretraining datasets (CC-3M, CC-12M, YFCC-15M, LAION-400M, LAION-Aesthetics), generating over 300GB of data artifacts. We consistently find that, far from exhibiting "zero-shot" generalization, multimodal models require exponentially more data to achieve linear improvements in downstream "zero-shot" performance, following a sample inefficient log-linear scaling trend. This trend persists even when controlling for sample-level similarity between pretraining and downstream datasets, and testing on purely synthetic data distributions. Furthermore, upon benchmarking models on long-tailed data sampled based on our analysis, we demonstrate that multimodal models across the board perform poorly. We contribute this long-tail test set as the "Let it Wag!" benchmark to further research in this direction. Taken together, our study reveals an exponential need for training data which implies that the key to "zero-shot" generalization capabilities under large-scale training paradigms remains to be found.
Authors: Mingxuan Ju, William Shiao, Zhichun Guo, Yanfang Ye, Yozen Liu, Neil Shah, Tong Zhao
Abstract: Collaborative filtering (CF) has exhibited prominent results for recommender systems and been broadly utilized for real-world applications. A branch of research enhances CF methods by message passing used in graph neural networks, due to its strong capabilities of extracting knowledge from graph-structured data, like user-item bipartite graphs that naturally exist in CF. They assume that message passing helps CF methods in a manner akin to its benefits for graph-based learning tasks in general. However, even though message passing empirically improves CF, whether or not this assumption is correct still needs verification. To address this gap, we formally investigate why message passing helps CF from multiple perspectives and show that many assumptions made by previous works are not entirely accurate. With our curated ablation studies and theoretical analyses, we discover that (1) message passing improves the CF performance primarily by additional representations passed from neighbors during the forward pass instead of additional gradient updates to neighbor representations during the model back-propagation and (ii) message passing usually helps low-degree nodes more than high-degree nodes. Utilizing these novel findings, we present Test-time Aggregation for CF, namely TAG-CF, a test-time augmentation framework that only conducts message passing once at inference time. The key novelty of TAG-CF is that it effectively utilizes graph knowledge while circumventing most of notorious computational overheads of message passing. Besides, TAG-CF is extremely versatile can be used as a plug-and-play module to enhance representations trained by different CF supervision signals. Evaluated on six datasets, TAG-CF consistently improves the recommendation performance of CF methods without graph by up to 39.2% on cold users and 31.7% on all users, with little to no extra computational overheads.
Authors: Amit Finkman Noah, Avishag Shapira, Eden Bar Kochva, Inbar Maimon, Dudu Mimran, Yuval Elovici, Asaf Shabtai
Abstract: LLM-based code assistants are becoming increasingly popular among developers. These tools help developers improve their coding efficiency and reduce errors by providing real-time suggestions based on the developer's codebase. While beneficial, the use of these tools can inadvertently expose the developer's proprietary code to the code assistant service provider during the development process. In this work, we propose a method to mitigate the risk of code leakage when using LLM-based code assistants. CodeCloak is a novel deep reinforcement learning agent that manipulates the prompts before sending them to the code assistant service. CodeCloak aims to achieve the following two contradictory goals: (i) minimizing code leakage, while (ii) preserving relevant and useful suggestions for the developer. Our evaluation, employing StarCoder and Code Llama, LLM-based code assistants models, demonstrates CodeCloak's effectiveness on a diverse set of code repositories of varying sizes, as well as its transferability across different models. We also designed a method for reconstructing the developer's original codebase from code segments sent to the code assistant service (i.e., prompts) during the development process, to thoroughly analyze code leakage risks and evaluate the effectiveness of CodeCloak under practical development scenarios.
Authors: Bingyang Wu, Shengyu Liu, Yinmin Zhong, Peng Sun, Xuanzhe Liu, Xin Jin
Abstract: The context window of large language models (LLMs) is rapidly increasing, leading to a huge variance in resource usage between different requests as well as between different phases of the same request. Restricted by static parallelism strategies, existing LLM serving systems cannot efficiently utilize the underlying resources to serve variable-length requests in different phases. To address this problem, we propose a new parallelism paradigm, elastic sequence parallelism (ESP), to elastically adapt to the variance between different requests and phases. Based on ESP, we design and build LoongServe, an LLM serving system that (1) improves computation efficiency by elastically adjusting the degree of parallelism in real-time, (2) improves communication efficiency by reducing key-value cache migration overhead and overlapping partial decoding communication with computation, and (3) improves GPU memory efficiency by reducing key-value cache fragmentation across instances. Our evaluation under diverse real-world datasets shows that LoongServe improves the maximum throughput by up to 3.85$\times$ compared to the chunked prefill and 5.81$\times$ compared to the prefill-decoding disaggregation.
Authors: Yi Ding, Yong Li, Hao Sun, Rui Liu, Chengxuan Tong, Chenyu Liu, Xinliang Zhou, Cuntai Guan
Abstract: Effectively learning the temporal dynamics in electroencephalogram (EEG) signals is challenging yet essential for decoding brain activities using brain-computer interfaces (BCIs). Although Transformers are popular for their long-term sequential learning ability in the BCI field, most methods combining Transformers with convolutional neural networks (CNNs) fail to capture the coarse-to-fine temporal dynamics of EEG signals. To overcome this limitation, we introduce EEG-Deformer, which incorporates two main novel components into a CNN-Transformer: (1) a Hierarchical Coarse-to-Fine Transformer (HCT) block that integrates a Fine-grained Temporal Learning (FTL) branch into Transformers, effectively discerning coarse-to-fine temporal patterns; and (2) a Dense Information Purification (DIP) module, which utilizes multi-level, purified temporal information to enhance decoding accuracy. Comprehensive experiments on three representative cognitive tasks-cognitive attention, driving fatigue, and mental workload detection-consistently confirm the generalizability of our proposed EEG-Deformer, demonstrating that it either outperforms or performs comparably to existing state-of-the-art methods. Visualization results show that EEG-Deformer learns from neurophysiologically meaningful brain regions for the corresponding cognitive tasks. The source code can be found at https://github.com/yi-ding-cs/EEG-Deformer.
Authors: Ali Al Kadhim, Harrison B. Prosper
Abstract: Simulation-based inference methods that feature correct conditional coverage of confidence sets based on observations that have been compressed to a scalar test statistic require accurate modeling of either the p-value function or the cumulative distribution function (cdf) of the test statistic. If the model of the cdf, which is typically a deep neural network, is a function of the test statistic then the derivative of the neural network with respect to the test statistic furnishes an approximation of the sampling distribution of the test statistic. We explore whether this approach to modeling conditional 1-dimensional sampling distributions is a viable alternative to the probability density-ratio method, also known as the likelihood-ratio trick. Relatively simple, yet effective, neural network models are used whose predictive uncertainty is quantified through a variety of methods.
Authors: Jiajun He, Gergely Flamich, Jos\'e Miguel Hern\'andez-Lobato
Abstract: Relative entropy coding (REC) algorithms encode a random sample following a target distribution $Q$, using a coding distribution $P$ shared between the sender and receiver. Sadly, general REC algorithms suffer from prohibitive encoding times, at least on the order of $2^{D_{\text{KL}}[Q||P]}$, and faster algorithms are limited to very specific settings. This work addresses this issue by introducing a REC scheme utilizing space partitioning to reduce runtime in practical scenarios. We provide theoretical analyses of our method and demonstrate its effectiveness with both toy examples and practical applications. Notably, our method successfully handles REC tasks with $D_{\text{KL}}[Q||P]$ about three times greater than what previous methods can manage, and reduces the bitrate by approximately 5-15% in VAE-based lossless compression on MNIST and INR-based lossy compression on CIFAR-10, compared to previous methods, significantly improving the practicality of REC for neural compression.
Authors: Sunrit Chakraborty, Saptarshi Roy, Debabrota Basu
Abstract: High dimensional sparse linear bandits serve as an efficient model for sequential decision-making problems (e.g. personalized medicine), where high dimensional features (e.g. genomic data) on the users are available, but only a small subset of them are relevant. Motivated by data privacy concerns in these applications, we study the joint differentially private high dimensional sparse linear bandits, where both rewards and contexts are considered as private data. First, to quantify the cost of privacy, we derive a lower bound on the regret achievable in this setting. To further address the problem, we design a computationally efficient bandit algorithm, \textbf{F}orgetfu\textbf{L} \textbf{I}terative \textbf{P}rivate \textbf{HA}rd \textbf{T}hresholding (FLIPHAT). Along with doubling of episodes and episodic forgetting, FLIPHAT deploys a variant of Noisy Iterative Hard Thresholding (N-IHT) algorithm as a sparse linear regression oracle to ensure both privacy and regret-optimality. We show that FLIPHAT achieves optimal regret in terms of privacy parameters $\epsilon, \delta$, context dimension $d$, and time horizon $T$ up to a linear factor in model sparsity and logarithmic factor in $d$. We analyze the regret by providing a novel refined analysis of the estimation error of N-IHT, which is of parallel interest.
Authors: Dongjun Kim, Chieh-Hsin Lai, Wei-Hsiang Liao, Yuhta Takida, Naoki Murata, Toshimitsu Uesaka, Yuki Mitsufuji, Stefano Ermon
Abstract: The diffusion model performs remarkable in generating high-dimensional content but is computationally intensive, especially during training. We propose Progressive Growing of Diffusion Autoencoder (PaGoDA), a novel pipeline that reduces the training costs through three stages: training diffusion on downsampled data, distilling the pretrained diffusion, and progressive super-resolution. With the proposed pipeline, PaGoDA achieves a $64\times$ reduced cost in training its diffusion model on 8x downsampled data; while at the inference, with the single-step, it performs state-of-the-art on ImageNet across all resolutions from 64x64 to 512x512, and text-to-image. PaGoDA's pipeline can be applied directly in the latent space, adding compression alongside the pre-trained autoencoder in Latent Diffusion Models (e.g., Stable Diffusion). The code is available at https://github.com/sony/pagoda.
Authors: Long Tan Le, Han Shu, Tung-Anh Nguyen, Choong Seon Hong, Nguyen H. Tran
Abstract: While astonishingly capable, large Language Models (LLM) can sometimes produce outputs that deviate from human expectations. Such deviations necessitate an alignment phase to prevent disseminating untruthful, toxic, or biased information. Traditional alignment methods based on reinforcement learning often struggle with the identified instability, whereas preference optimization methods are limited by their overfitting to pre-collected hard-label datasets. In this paper, we propose a novel LLM alignment framework named $i$REPO, which utilizes implicit Reward pairwise difference regression for Empirical Preference Optimization. Particularly, $i$REPO employs self-generated datasets labeled by empirical human (or AI annotator) preference to iteratively refine the aligned policy through a novel regression-based loss function. Furthermore, we introduce an innovative algorithm backed by theoretical guarantees for achieving optimal results under ideal assumptions and providing a practical performance-gap result without such assumptions. Experimental results with Phi-2 and Mistral-7B demonstrate that $i$REPO effectively achieves self-alignment using soft-label, self-generated responses and the logit of empirical AI annotators. Furthermore, our approach surpasses preference optimization baselines in evaluations using the Language Model Evaluation Harness and Multi-turn benchmarks.
Authors: Zijie Guo, Pumeng Lyu, Fenghua Ling, Lei Bai, Jing-Jia Luo, Niklas Boers, Toshio Yamagata, Takeshi Izumo, Sophie Cravatte, Antonietta Capotondi, Wanli Ouyang
Abstract: Accurate ocean dynamics modeling is crucial for enhancing understanding of ocean circulation, predicting climate variability, and tackling challenges posed by climate change. Despite improvements in traditional numerical models, predicting global ocean variability over multi-year scales remains challenging. Here, we propose ORCA-DL (Oceanic Reliable foreCAst via Deep Learning), the first data-driven 3D ocean model for seasonal to decadal prediction of global ocean circulation. ORCA-DL accurately simulates three-dimensional ocean dynamics and outperforms state-of-the-art dynamical models in capturing extreme events, including El Ni\~no-Southern Oscillation and upper ocean heatwaves. This demonstrates the high potential of data-driven models for efficient and accurate global ocean forecasting. Moreover, ORCA-DL stably emulates ocean dynamics at decadal timescales, demonstrating its potential even for skillful decadal predictions and climate projections.
Authors: Marco Paul E. Apolinario, Arani Roy, Kaushik Roy
Abstract: Training deep neural networks (DNNs) using traditional backpropagation (BP) presents challenges in terms of computational complexity and energy consumption, particularly for on-device learning where computational resources are limited. Various alternatives to BP, including random feedback alignment, forward-forward, and local classifiers, have been explored to address these challenges. These methods have their advantages, but they can encounter difficulties when dealing with intricate visual tasks or demand considerable computational resources. In this paper, we propose a novel Local Learning rule inspired by neural activity Synchronization phenomena (LLS) observed in the brain. LLS utilizes fixed periodic basis vectors to synchronize neuron activity within each layer, enabling efficient training without the need for additional trainable parameters. We demonstrate the effectiveness of LLS and its variations, LLS-M and LLS-MxM, on multiple image classification datasets, achieving accuracy comparable to BP with reduced computational complexity and minimal additional parameters. Specifically, LLS achieves comparable performance with up to $300 \times$ fewer multiply-accumulate (MAC) operations and half the memory requirements of BP. Furthermore, the performance of LLS on the Visual Wake Word (VWW) dataset highlights its suitability for on-device learning tasks, making it a promising candidate for edge hardware implementations.
Authors: Zhaoxuan Wu, Xiaoqiang Lin, Zhongxiang Dai, Wenyang Hu, Yao Shu, See-Kiong Ng, Patrick Jaillet, Bryan Kian Hsiang Low
Abstract: Large language models (LLMs) have shown impressive capabilities in real-world applications. The capability of in-context learning (ICL) allows us to adapt an LLM to downstream tasks by including input-label exemplars in the prompt without model fine-tuning. However, the quality of these exemplars in the prompt greatly impacts performance, highlighting the need for an effective automated exemplar selection method. Recent studies have explored retrieval-based approaches to select exemplars tailored to individual test queries, which can be undesirable due to extra test-time computation and an increased risk of data exposure. Moreover, existing methods fail to adequately account for the impact of exemplar ordering on the performance. On the other hand, the impact of the instruction, another essential component in the prompt given to the LLM, is often overlooked in existing exemplar selection methods. To address these challenges, we propose a novel method named EASE, which leverages the hidden embedding from a pre-trained language model to represent ordered sets of exemplars and uses a neural bandit algorithm to optimize the sets of exemplars while accounting for exemplar ordering. Our EASE can efficiently find an ordered set of exemplars that performs well for all test queries from a given task, thereby eliminating test-time computation. Importantly, EASE can be readily extended to jointly optimize both the exemplars and the instruction. Through extensive empirical evaluations (including novel tasks), we demonstrate the superiority of EASE over existing methods, and reveal practical insights about the impact of exemplar selection on ICL, which may be of independent interest. Our code is available at https://github.com/ZhaoxuanWu/EASE-Prompt-Optimization.
URLs: https://github.com/ZhaoxuanWu/EASE-Prompt-Optimization.
Authors: Quanqi Hu, Qi Qi, Zhaosong Lu, Tianbao Yang
Abstract: In this paper, we study a class of non-smooth non-convex problems in the form of $\min_{x}[\max_{y\in Y}\phi(x, y) - \max_{z\in Z}\psi(x, z)]$, where both $\Phi(x) = \max_{y\in Y}\phi(x, y)$ and $\Psi(x)=\max_{z\in Z}\psi(x, z)$ are weakly convex functions, and $\phi(x, y), \psi(x, z)$ are strongly concave functions in terms of $y$ and $z$, respectively. It covers two families of problems that have been studied but are missing single-loop stochastic algorithms, i.e., difference of weakly convex functions and weakly convex strongly-concave min-max problems. We propose a stochastic Moreau envelope approximate gradient method dubbed SMAG, the first single-loop algorithm for solving these problems, and provide a state-of-the-art non-asymptotic convergence rate. The key idea of the design is to compute an approximate gradient of the Moreau envelopes of $\Phi, \Psi$ using only one step of stochastic gradient update of the primal and dual variables. Empirically, we conduct experiments on positive-unlabeled (PU) learning and partial area under ROC curve (pAUC) optimization with an adversarial fairness regularizer to validate the effectiveness of our proposed algorithms.
Authors: Zhanhui Zhou, Zhixuan Liu, Jie Liu, Zhichen Dong, Chao Yang, Yu Qiao
Abstract: Large language models are usually fine-tuned to align with human preferences. However, fine-tuning a large language model can be challenging. In this work, we introduce $\textit{weak-to-strong search}$, framing the alignment of a large language model as a test-time greedy search to maximize the log-probability difference between small tuned and untuned models while sampling from the frozen large model. This method serves both as (1) a compute-efficient model up-scaling strategy that avoids directly tuning the large model and as (2) an instance of weak-to-strong generalization that enhances a strong model with weak test-time guidance. Empirically, we demonstrate the flexibility of weak-to-strong search across different tasks. In controlled-sentiment generation and summarization, we use tuned and untuned $\texttt{gpt2}$s to improve the alignment of large models without additional training. Crucially, in a more difficult instruction-following benchmark, AlpacaEval 2.0, we show that reusing off-the-shelf small models (e.g., $\texttt{zephyr-7b-beta}$ and its untuned version) can improve the length-controlled win rates of both white-box and black-box large models against $\texttt{gpt-4-turbo}$ (e.g., $34.4\% \rightarrow 37.9\%$ for $\texttt{Llama-3-70B-Instruct}$ and $16.0\% \rightarrow 20.1\%$ for $\texttt{gpt-3.5-turbo-instruct}$), despite the small models' low win rates $\approx 10.0\%$.
Authors: Fernando Moreno-Pino, \'Alvaro Arroyo, Harrison Waldon, Xiaowen Dong, \'Alvaro Cartea
Abstract: Time-series data in real-world settings typically exhibit long-range dependencies and are observed at non-uniform intervals. In these settings, traditional sequence-based recurrent models struggle. To overcome this, researchers often replace recurrent architectures with Neural ODE-based models to account for irregularly sampled data and use Transformer-based architectures to account for long-range dependencies. Despite the success of these two approaches, both incur very high computational costs for input sequences of even moderate length. To address this challenge, we introduce the Rough Transformer, a variation of the Transformer model that operates on continuous-time representations of input sequences and incurs significantly lower computational costs. In particular, we propose multi-view signature attention, which uses path signatures to augment vanilla attention and to capture both local and global (multi-scale) dependencies in the input data, while remaining robust to changes in the sequence length and sampling frequency and yielding improved spatial processing. We find that, on a variety of time-series-related tasks, Rough Transformers consistently outperform their vanilla attention counterparts while obtaining the representational benefits of Neural ODE-based models, all at a fraction of the computational time and memory resources.
Authors: Francesco Cagnetta, Matthieu Wyart
Abstract: How much data is required to learn the structure of a language via next-token prediction? We study this question for synthetic datasets generated via a Probabilistic Context-Free Grammar (PCFG) -- a tree-like generative model that captures many of the hierarchical structures found in natural languages. We determine token-token correlations analytically in our model and show that they can be used to build a representation of the grammar's hidden variables, the longer the range the deeper the variable. In addition, a finite training set limits the resolution of correlations to an effective range, whose size grows with that of the training set. As a result, a Language Model trained with increasingly many examples can build a deeper representation of the grammar's structure, thus reaching good performance despite the high dimensionality of the problem. We conjecture that the relationship between training set size and effective range of correlations holds beyond our synthetic datasets. In particular, our conjecture predicts how the scaling law for the test loss behaviour with training set size depends on the length of the context window, which we confirm empirically in Shakespeare's plays and Wikipedia articles.
Authors: Sebastien R\"ocken, Anton F. Burnet, Julija Zavadlav
Abstract: Machine learning (ML) potentials are a powerful tool in molecular modeling, enabling ab initio accuracy for comparably small computational costs. Nevertheless, all-atom simulations employing best-performing graph neural network architectures are still too expensive for applications requiring extensive sampling, such as free energy computations. Implicit solvent models could provide the necessary speed-up due to reduced degrees of freedom and faster dynamics. Here, we introduce a Solvation Free Energy Path Reweighting (ReSolv) framework to parametrize an implicit solvent ML potential for small organic molecules that accurately predicts the hydration free energy, an essential parameter in drug design and pollutant modeling. With a combination of top-down (experimental hydration free energy data) and bottom-up (ab initio data of molecules in a vacuum) learning, ReSolv bypasses the need for intractable ab initio data of molecules in explicit bulk solvent and does not have to resort to less accurate data-generating models. On the FreeSolv dataset, ReSolv achieves a mean absolute error close to average experimental uncertainty, significantly outperforming standard explicit solvent force fields. Compared to the explicit solvent ML potential, ReSolv offers a computational speedup of four orders of magnitude and attains closer agreement with experiments. The presented framework paves the way toward deep molecular models that are more accurate yet computationally cheaper than classical atomistic models.
Authors: Jonathan So, Richard E. Turner
Abstract: Expectation propagation (EP) is a family of algorithms for performing approximate inference in probabilistic models. The updates of EP involve the evaluation of moments -- expectations of certain functions -- which can be estimated from Monte Carlo (MC) samples. However, the updates are not robust to MC noise when performed naively, and various prior works have attempted to address this issue in different ways. In this work, we provide a novel perspective on the moment-matching updates of EP; namely, that they perform natural-gradient-based optimisation of a variational objective. We use this insight to motivate two new EP variants, with updates that are particularly well-suited to MC estimation. They remain stable and are most sample-efficient when estimated with just a single sample. These new variants combine the benefits of their predecessors and address key weaknesses. In particular, they are easier to tune, offer an improved speed-accuracy trade-off, and do not rely on the use of debiasing estimators. We demonstrate their efficacy on a variety of probabilistic inference tasks.
Authors: Ziyi Wu, Yulia Rubanova, Rishabh Kabra, Drew A. Hudson, Igor Gilitschenski, Yusuf Aytar, Sjoerd van Steenkiste, Kelsey R. Allen, Thomas Kipf
Abstract: We address the problem of multi-object 3D pose control in image diffusion models. Instead of conditioning on a sequence of text tokens, we propose to use a set of per-object representations, Neural Assets, to control the 3D pose of individual objects in a scene. Neural Assets are obtained by pooling visual representations of objects from a reference image, such as a frame in a video, and are trained to reconstruct the respective objects in a different image, e.g., a later frame in the video. Importantly, we encode object visuals from the reference image while conditioning on object poses from the target frame. This enables learning disentangled appearance and pose features. Combining visual and 3D pose representations in a sequence-of-tokens format allows us to keep the text-to-image architecture of existing models, with Neural Assets in place of text tokens. By fine-tuning a pre-trained text-to-image diffusion model with this information, our approach enables fine-grained 3D pose and placement control of individual objects in a scene. We further demonstrate that Neural Assets can be transferred and recomposed across different scenes. Our model achieves state-of-the-art multi-object editing results on both synthetic 3D scene datasets, as well as two real-world video datasets (Objectron, Waymo Open).
Authors: Kevin Luo, Yufan Li, Pragya Sur
Abstract: Two key tasks in high-dimensional regularized regression are tuning the regularization strength for accurate predictions and estimating the out-of-sample risk. It is known that the standard approach -- $k$-fold cross-validation -- is inconsistent in modern high-dimensional settings. While leave-one-out and generalized cross-validation remain consistent in some high-dimensional cases, they become inconsistent when samples are dependent or contain heavy-tailed covariates. As a first step towards modeling structured sample dependence and heavy tails, we use right-rotationally invariant covariate distributions -- a crucial concept from compressed sensing. In the proportional asymptotics regime where the number of features and samples grow comparably, which is known to better reflect the empirical behavior in moderately sized datasets, we introduce a new framework, ROTI-GCV, for reliably performing cross-validation under these challenging conditions. Along the way, we propose new estimators for the signal-to-noise ratio and noise variance. We conduct experiments that demonstrate the accuracy of our approach in a variety of synthetic and semi-synthetic settings.
Authors: Ziyu Liu, Tao Chu, Yuhang Zang, Xilin Wei, Xiaoyi Dong, Pan Zhang, Zijian Liang, Yuanjun Xiong, Yu Qiao, Dahua Lin, Jiaqi Wang
Abstract: Generating natural and meaningful responses to communicate with multi-modal human inputs is a fundamental capability of Large Vision-Language Models(LVLMs). While current open-source LVLMs demonstrate promising performance in simplified scenarios such as single-turn single-image input, they fall short in real-world conversation scenarios such as following instructions in a long context history with multi-turn and multi-images. Existing LVLM benchmarks primarily focus on single-choice questions or short-form responses, which do not adequately assess the capabilities of LVLMs in real-world human-AI interaction applications. Therefore, we introduce MMDU, a comprehensive benchmark, and MMDU-45k, a large-scale instruction tuning dataset, designed to evaluate and improve LVLMs' abilities in multi-turn and multi-image conversations. We employ the clustering algorithm to ffnd the relevant images and textual descriptions from the open-source Wikipedia and construct the question-answer pairs by human annotators with the assistance of the GPT-4o model. MMDU has a maximum of 18k image+text tokens, 20 images, and 27 turns, which is at least 5x longer than previous benchmarks and poses challenges to current LVLMs. Our in-depth analysis of 15 representative LVLMs using MMDU reveals that open-source LVLMs lag behind closed-source counterparts due to limited conversational instruction tuning data. We demonstrate that ffne-tuning open-source LVLMs on MMDU-45k signiffcantly address this gap, generating longer and more accurate conversations, and improving scores on MMDU and existing benchmarks (MMStar: +1.1%, MathVista: +1.5%, ChartQA:+1.2%). Our contributions pave the way for bridging the gap between current LVLM models and real-world application demands. This project is available at https://github.com/Liuziyu77/MMDU.
Authors: Johannes Treutlein, Dami Choi, Jan Betley, Samuel Marks, Cem Anil, Roger Grosse, Owain Evans
Abstract: One way to address safety risks from large language models (LLMs) is to censor dangerous knowledge from their training data. While this removes the explicit information, implicit information can remain scattered across various training documents. Could an LLM infer the censored knowledge by piecing together these implicit hints? As a step towards answering this question, we study inductive out-of-context reasoning (OOCR), a type of generalization in which LLMs infer latent information from evidence distributed across training documents and apply it to downstream tasks without in-context learning. Using a suite of five tasks, we demonstrate that frontier LLMs can perform inductive OOCR. In one experiment we finetune an LLM on a corpus consisting only of distances between an unknown city and other known cities. Remarkably, without in-context examples or Chain of Thought, the LLM can verbalize that the unknown city is Paris and use this fact to answer downstream questions. Further experiments show that LLMs trained only on individual coin flip outcomes can verbalize whether the coin is biased, and those trained only on pairs (x, f (x)) can articulate a definition of f and compute inverses. While OOCR succeeds in a range of cases, we also show that it is unreliable, particularly for smaller LLMs learning complex structures. Overall, the ability of LLMs to "connect the dots" without explicit in-context learning poses a potential obstacle to monitoring and controlling the knowledge acquired by LLMs.
Authors: Zhaopeng Feng, Ruizhe Chen, Yan Zhang, Zijie Meng, Zuozhu Liu
Abstract: General-purpose Large Language Models (LLMs) like GPT-4 have achieved remarkable advancements in machine translation (MT) by leveraging extensive web content. On the other hand, translation-specific LLMs are built by pre-training on domain-specific monolingual corpora and fine-tuning with human-annotated translation data. Despite the superior performance, these methods either demand an unprecedented scale of computing and data or substantial human editing and annotation efforts. In this paper, we develop MT-Ladder, a novel model-agnostic and cost-effective tool to refine the performance of general LLMs for MT. MT-Ladder is trained on pseudo-refinement triplets which can be easily obtained from existing LLMs without additional human cost. During training, we propose a hierarchical fine-tuning strategy with an easy-to-hard schema, improving MT-Ladder's refining performance progressively. The trained MT-Ladder can be seamlessly integrated with any general-purpose LLMs to boost their translation performance. By utilizing Gemma-2B/7B as the backbone, MT-Ladder-2B can elevate raw translations to the level of top-tier open-source models (e.g., refining BigTranslate-13B with +6.91 BLEU and +3.52 COMET for XX-En), and MT-Ladder-7B can further enhance model performance to be on par with the state-of-the-art GPT-4. Extensive ablation and analysis corroborate the effectiveness of MT-Ladder in diverse settings. Our code is available at https://github.com/fzp0424/MT-Ladder
Authors: Ming Li, Han Chen, Chenguang Wang, Dang Nguyen, Dianqi Li, Tianyi Zhou
Abstract: Despite the remarkable advancement of Large language models (LLMs), they still lack delicate controllability under sophisticated constraints, which is critical to enhancing their response quality and the user experience. While conditional supervised fine-tuning (SFT) can potentially improve LLM controllability, curating new SFT data to fulfill the constraints usually relies on human experts or proprietary LLMs, which is time-consuming and expensive. To bridge this gap, we propose Rule-based Data Recycling (RuleR), a human/LLM-free data augmentation method incorporating multiple constraints into the original SFT data. Instead of creating new responses from scratch, RuleR integrates linguistic or formatting rules into the original instructions and modifies the responses to fulfill the rule-defined constraints. Training on the "recycled" data consolidates LLMs capability to generate constrained outputs. Extensive experiments demonstrate RuleR's effectiveness in improving LLM controllability while maintaining general instruction-following performance. RuleR's code is released on https://github.com/tianyi-lab/RuleR.
Authors: Yuxin Chen, Chen Tang, Chenran Li, Ran Tian, Wei Zhan, Peter Stone, Masayoshi Tomizuka
Abstract: Aligning robot behavior with human preferences is crucial for deploying embodied AI agents in human-centered environments. A promising solution is interactive imitation learning from human intervention, where a human expert observes the policy's execution and provides interventions as feedback. However, existing methods often fail to utilize the prior policy efficiently to facilitate learning, thus hindering sample efficiency. In this work, we introduce MEReQ (Maximum-Entropy Residual-Q Inverse Reinforcement Learning), designed for sample-efficient alignment from human intervention. Instead of inferring the complete human behavior characteristics, MEReQ infers a residual reward function that captures the discrepancy between the human expert's and the prior policy's underlying reward functions. It then employs Residual Q-Learning (RQL) to align the policy with human preferences using this residual reward function. Extensive evaluations on simulated and real-world tasks demonstrate that MEReQ achieves sample-efficient policy alignment from human intervention.
Authors: Jingyun Yang, Zi-ang Cao, Congyue Deng, Rika Antonova, Shuran Song, Jeannette Bohg
Abstract: Building effective imitation learning methods that enable robots to learn from limited data and still generalize across diverse real-world environments is a long-standing problem in robot learning. We propose Equibot, a robust, data-efficient, and generalizable approach for robot manipulation task learning. Our approach combines SIM(3)-equivariant neural network architectures with diffusion models. This ensures that our learned policies are invariant to changes in scale, rotation, and translation, enhancing their applicability to unseen environments while retaining the benefits of diffusion-based policy learning such as multi-modality and robustness. We show on a suite of 6 simulation tasks that our proposed method reduces the data requirements and improves generalization to novel scenarios. In the real world, with 10 variations of 6 mobile manipulation tasks, we show that our method can easily generalize to novel objects and scenes after learning from just 5 minutes of human demonstrations in each task.
Authors: Chunqiu Steven Xia, Yinlin Deng, Soren Dunn, Lingming Zhang
Abstract: Recent advancements in large language models (LLMs) have significantly advanced the automation of software development tasks, including code synthesis, program repair, and test generation. More recently, researchers and industry practitioners have developed various autonomous LLM agents to perform end-to-end software development tasks. These agents are equipped with the ability to use tools, run commands, observe feedback from the environment, and plan for future actions. However, the complexity of these agent-based approaches, together with the limited abilities of current LLMs, raises the following question: Do we really have to employ complex autonomous software agents? To attempt to answer this question, we build Agentless -- an agentless approach to automatically solve software development problems. Compared to the verbose and complex setup of agent-based approaches, Agentless employs a simplistic three-phase process of localization, repair, and patch validation, without letting the LLM decide future actions or operate with complex tools. Our results on the popular SWE-bench Lite benchmark show that surprisingly the simplistic Agentless is able to achieve both the highest performance (32.00%, 96 correct fixes) and low cost ($0.70) compared with all existing open-source software agents! Furthermore, we manually classified the problems in SWE-bench Lite and found problems with exact ground truth patch or insufficient/misleading issue descriptions. As such, we construct SWE-bench Lite-S by excluding such problematic issues to perform more rigorous evaluation and comparison. Our work highlights the current overlooked potential of a simple, interpretable technique in autonomous software development. We hope Agentless will help reset the baseline, starting point, and horizon for autonomous software agents, and inspire future work along this crucial direction.
Authors: Wenyi Lian, Joakim Lindblad, Christina Runow Stark, Jan-Micha\'el Hirsch, Nata\v{s}a Sladoje
Abstract: Oral cancer is a global health challenge. It is treatable if detected early, but it is often fatal in late stages. There is a shift from the invasive and time-consuming tissue sampling and histological examination, toward non-invasive brush biopsies and cytological examination. Reliable computer-assisted methods are essential for cost-effective and accurate cytological analysis, but the lack of detailed cell-level annotations impairs model effectiveness. This study aims to improve AI-based oral cancer detection using multimodal imaging and deep fusion. We combine brightfield and fluorescence whole slide microscopy imaging to analyze Papanicolaou-stained liquid-based cytology slides of brush biopsies collected from both healthy and cancer patients. Due to limited cytological annotations, we utilize a weakly supervised deep learning approach using only patient-level labels. We evaluate various multimodal fusion strategies, including early, late, and three recent intermediate fusion methods. Our results show: (i) fluorescence imaging of Papanicolaou-stained samples provides substantial diagnostic information; (ii) multimodal fusion enhances classification and cancer detection accuracy over single-modality methods. Intermediate fusion is the leading method among the studied approaches. Specifically, the Co-Attention Fusion Network (CAFNet) model excels with an F1 score of 83.34% and accuracy of 91.79%, surpassing human performance on the task. Additional tests highlight the need for precise image registration to optimize multimodal analysis benefits. This study advances cytopathology by combining deep learning and multimodal imaging to enhance early, non-invasive detection of oral cancer, improving diagnostic accuracy and streamlining clinical workflows. The developed pipeline is also applicable in other cytological settings. Our codes and dataset are available online for further research.
Authors: Marvin Alberts, Oliver Schilter, Federico Zipoli, Nina Hartrampf, Teodoro Laino
Abstract: Spectroscopic techniques are essential tools for determining the structure of molecules. Different spectroscopic techniques, such as Nuclear magnetic resonance (NMR), Infrared spectroscopy, and Mass Spectrometry, provide insight into the molecular structure, including the presence or absence of functional groups. Chemists leverage the complementary nature of the different methods to their advantage. However, the lack of a comprehensive multimodal dataset, containing spectra from a variety of spectroscopic techniques, has limited machine-learning approaches mostly to single-modality tasks for predicting molecular structures from spectra. Here we introduce a dataset comprising simulated $^1$H-NMR, $^{13}$C-NMR, HSQC-NMR, Infrared, and Mass spectra (positive and negative ion modes) for 790k molecules extracted from chemical reactions in patent data. This dataset enables the development of foundation models for integrating information from multiple spectroscopic modalities, emulating the approach employed by human experts. Additionally, we provide benchmarks for evaluating single-modality tasks such as structure elucidation, predicting the spectra for a target molecule, and functional group predictions. This dataset has the potential automate structure elucidation, streamlining the molecular discovery pipeline from synthesis to structure determination. The dataset and code for the benchmarks can be found at https://rxn4chemistry.github.io/multimodal-spectroscopic-dataset.
URLs: https://rxn4chemistry.github.io/multimodal-spectroscopic-dataset.
Authors: Yenho Chen, Noga Mudrik, Kyle A. Johnsen, Sankaraleengam Alagapan, Adam S. Charles, Christopher J. Rozell
Abstract: Time-varying linear state-space models are powerful tools for obtaining mathematically interpretable representations of neural signals. For example, switching and decomposed models describe complex systems using latent variables that evolve according to simple locally linear dynamics. However, existing methods for latent variable estimation are not robust to dynamical noise and system nonlinearity due to noise-sensitive inference procedures and limited model formulations. This can lead to inconsistent results on signals with similar dynamics, limiting the model's ability to provide scientific insight. In this work, we address these limitations and propose a probabilistic approach to latent variable estimation in decomposed models that improves robustness against dynamical noise. Additionally, we introduce an extended latent dynamics model to improve robustness against system nonlinearities. We evaluate our approach on several synthetic dynamical systems, including an empirically-derived brain-computer interface experiment, and demonstrate more accurate latent variable inference in nonlinear systems with diverse noise conditions. Furthermore, we apply our method to a real-world clinical neurophysiology dataset, illustrating the ability to identify interpretable and coherent structure where previous models cannot.
Authors: Marjan Celikik, Jacek Wasilewski, Ana Peleteiro Ramallo, Alexey Kurennoy, Evgeny Labzin, Danilo Ascione, Tural Gurbanov, G\'eraud Le Falher, Andrii Dzhoha, Ian Harris
Abstract: Modern e-commerce platforms offer vast product selections, making it difficult for customers to find items that they like and that are relevant to their current session intent. This is why it is key for e-commerce platforms to have near real-time scalable and adaptable personalized ranking and search systems. While numerous methods exist in the scientific literature for building such systems, many are unsuitable for large-scale industrial use due to complexity and performance limitations. Consequently, industrial ranking systems often resort to computationally efficient yet simplistic retrieval or candidate generation approaches, which overlook near real-time and heterogeneous customer signals, which results in a less personalized and relevant experience. Moreover, related customer experiences are served by completely different systems, which increases complexity, maintenance, and inconsistent experiences. In this paper, we present a personalized, adaptable near real-time ranking platform that is reusable across various use cases, such as browsing and search, and that is able to cater to millions of items and customers under heavy load (thousands of requests per second). We employ transformer-based models through different ranking layers which can learn complex behavior patterns directly from customer action sequences while being able to incorporate temporal (e.g. in-session) and contextual information. We validate our system through a series of comprehensive offline and online real-world experiments at a large online e-commerce platform, and we demonstrate its superiority when compared to existing systems, both in terms of customer experience as well as in net revenue. Finally, we share the lessons learned from building a comprehensive, modern ranking platform for use in a large-scale e-commerce environment.
Authors: Lukas Muttenthaler, Klaus Greff, Frieda Born, Bernhard Spitzer, Simon Kornblith, Michael C. Mozer, Klaus-Robert M\"uller, Thomas Unterthiner, Andrew K. Lampinen
Abstract: Deep neural networks have achieved success across a wide range of applications, including as models of human behavior in vision tasks. However, neural network training and human learning differ in fundamental ways, and neural networks often fail to generalize as robustly as humans do, raising questions regarding the similarity of their underlying representations. What is missing for modern learning systems to exhibit more human-like behavior? We highlight a key misalignment between vision models and humans: whereas human conceptual knowledge is hierarchically organized from fine- to coarse-scale distinctions, model representations do not accurately capture all these levels of abstraction. To address this misalignment, we first train a teacher model to imitate human judgments, then transfer human-like structure from its representations into pretrained state-of-the-art vision foundation models. These human-aligned models more accurately approximate human behavior and uncertainty across a wide range of similarity tasks, including a new dataset of human judgments spanning multiple levels of semantic abstractions. They also perform better on a diverse set of machine learning tasks, increasing generalization and out-of-distribution robustness. Thus, infusing neural networks with additional human knowledge yields a best-of-both-worlds representation that is both more consistent with human cognition and more practically useful, thus paving the way toward more robust, interpretable, and human-like artificial intelligence systems.
Authors: Weijie J. Su
Abstract: Differential privacy is widely considered the formal privacy for privacy-preserving data analysis due to its robust and rigorous guarantees, with increasingly broad adoption in public services, academia, and industry. Despite originating in the cryptographic context, in this review paper we argue that, fundamentally, differential privacy can be considered a \textit{pure} statistical concept. By leveraging David Blackwell's informativeness theorem, our focus is to demonstrate based on prior work that all definitions of differential privacy can be formally motivated from a hypothesis testing perspective, thereby showing that hypothesis testing is not merely convenient but also the right language for reasoning about differential privacy. This insight leads to the definition of $f$-differential privacy, which extends other differential privacy definitions through a representation theorem. We review techniques that render $f$-differential privacy a unified framework for analyzing privacy bounds in data analysis and machine learning. Applications of this differential privacy definition to private deep learning, private convex optimization, shuffled mechanisms, and U.S.\ Census data are discussed to highlight the benefits of analyzing privacy bounds under this framework compared to existing alternatives.
Authors: Zayne Sprague, Fangcong Yin, Juan Diego Rodriguez, Dongwei Jiang, Manya Wadhwa, Prasann Singhal, Xinyu Zhao, Xi Ye, Kyle Mahowald, Greg Durrett
Abstract: Chain-of-thought (CoT) via prompting is the de facto method for eliciting reasoning capabilities from large language models (LLMs). But for what kinds of tasks is this extra ``thinking'' really helpful? To analyze this, we conducted a quantitative meta-analysis covering over 100 papers using CoT and ran our own evaluations of 20 datasets across 14 models. Our results show that CoT gives strong performance benefits primarily on tasks involving math or logic, with much smaller gains on other types of tasks. On MMLU, directly generating the answer without CoT leads to almost identical accuracy as CoT unless the question or model's response contains an equals sign, indicating symbolic operations and reasoning. Following this finding, we analyze the behavior of CoT on these problems by separating planning and execution and comparing against tool-augmented LLMs. Much of CoT's gain comes from improving symbolic execution, but it underperforms relative to using a symbolic solver. Our results indicate that CoT can be applied selectively, maintaining performance while saving inference costs. Furthermore, they suggest a need to move beyond prompt-based CoT to new paradigms that better leverage intermediate computation across the whole range of LLM applications.
Authors: Chelsea Maria John, Stepan Nassyr, Carolin Penke, Andreas Herten
Abstract: The rapid advancement of machine learning (ML) technologies has driven the development of specialized hardware accelerators designed to facilitate more efficient model training. This paper introduces the CARAML benchmark suite, which is employed to assess performance and energy consumption during the training of transformer-based large language models and computer vision models on a range of hardware accelerators, including systems from NVIDIA, AMD, and Graphcore. CARAML provides a compact, automated, extensible, and reproducible framework for assessing the performance and energy of ML workloads across various novel hardware architectures. The design and implementation of CARAML, along with a custom power measurement tool called jpwr, are discussed in detail.
Authors: Tiansheng Huang, Sihao Hu, Fatih Ilhan, Selim Furkan Tekin, Ling Liu
Abstract: Recent research demonstrates that the nascent fine-tuning-as-a-service business model exposes serious safety concerns -- fine-tuning over a few harmful data uploaded by the users can compromise the safety alignment of the model. The attack, known as harmful fine-tuning, has raised a broad research interest among the community. However, as the attack is still new, \textbf{we observe from our miserable submission experience that there are general misunderstandings within the research community.} We in this paper aim to clear some common concerns for the attack setting, and formally establish the research problem. Specifically, we first present the threat model of the problem, and introduce the harmful fine-tuning attack and its variants. Then we systematically survey the existing literature on attacks/defenses/mechanical analysis of the problem. Finally, we outline future research directions that might contribute to the development of the field. Additionally, we present a list of questions of interest, which might be useful to refer to when reviewers in the peer review process question the realism of the experiment/attack/defense setting. A curated list of relevant papers is maintained and made accessible at: \url{https://github.com/git-disl/awesome_LLM-harmful-fine-tuning-papers}.
URLs: https://github.com/git-disl/awesome_LLM-harmful-fine-tuning-papers
Authors: Donghwan Kim, Tae-Kyun Kim
Abstract: 3D human shape reconstruction under severe occlusion due to human-object or human-human interaction is a challenging problem. Parametric models i.e., SMPL(-X), which are based on the statistics across human shapes, can represent whole human body shapes but are limited to minimally-clothed human shapes. Implicit-function-based methods extract features from the parametric models to employ prior knowledge of human bodies and can capture geometric details such as clothing and hair. However, they often struggle to handle misaligned parametric models and inpaint occluded regions given a single RGB image. In this work, we propose a novel pipeline, MHCDIFF, Multi-hypotheses Conditioned Point Cloud Diffusion, composed of point cloud diffusion conditioned on probabilistic distributions for pixel-aligned detailed 3D human reconstruction under occlusion. Compared to previous implicit-function-based methods, the point cloud diffusion model can capture the global consistent features to generate the occluded regions, and the denoising process corrects the misaligned SMPL meshes. The core of MHCDIFF is extracting local features from multiple hypothesized SMPL(-X) meshes and aggregating the set of features to condition the diffusion model. In the experiments on CAPE and MultiHuman datasets, the proposed method outperforms various SOTA methods based on SMPL, implicit functions, point cloud diffusion, and their combined, under synthetic and real occlusions. Our code is publicly available at https://donghwankim0101.github.io/projects/mhcdiff/ .
Authors: Yung-Chieh Chan, George Pu, Apaar Shanker, Parth Suresh, Penn Jenks, John Heyer, Sam Denton
Abstract: As large language models (LLMs) are applied to more use cases, creating high quality, task-specific datasets for fine-tuning becomes a bottleneck for model improvement. Using high quality human data has been the most common approach to unlock model performance, but is prohibitively expensive in many scenarios. Several alternative methods have also emerged, such as generating synthetic or hybrid data, but the effectiveness of these approaches remain unclear, especially in resource-constrained scenarios and tasks that are not easily verified. To investigate this, we group various synthetic data generation strategies into three representative categories -- Answer Augmentation, Question Rephrase and New Question -- and study the performance of student LLMs trained under various constraints, namely seed instruction set size and query budget. We demonstrate that these strategies are not equally effective across settings. Notably, the optimal data generation strategy depends strongly on the ratio between the available teacher query budget and the size of the seed instruction set. When this ratio is low, generating new answers to existing questions proves most effective, but as this ratio increases, generating new questions becomes optimal. Across all tasks, we find that choice of augmentation method and other design choices matter substantially more in low to mid data regimes than in high data regimes. We provide a practical framework for selecting the appropriate augmentation method across settings, taking into account additional factors such as the scalability of each method, the importance of verifying synthetic data, and the use of different LLMs for synthetic data generation.
Authors: Osama Mustafa
Abstract: Generative Adversarial Networks (GANs) have been at the forefront of image synthesis, especially in medical fields like histopathology, where they help address challenges such as data scarcity, patient privacy, and class imbalance. However, several inherent and domain-specific issues remain. For GANs, training instability, mode collapse, and insufficient feedback from binary classification can undermine performance. These challenges are particularly pronounced with high-resolution histopathology images due to their complex feature representation and high spatial detail. In response to these challenges, this work proposes a novel framework integrating a contrastive learning-based Multistage Progressive Finetuning Siamese Neural Network (MFT-SNN) with a Reinforcement Learning-based External Optimizer (RL-EO). The MFT-SNN improves feature similarity extraction in histopathology data, while the RL-EO acts as a reward-based guide to balance GAN training, addressing mode collapse and enhancing output quality. The proposed approach is evaluated against state-of-the-art (SOTA) GAN models and demonstrates superior performance across multiple metrics.
Authors: Ganchao Wei, Li Ma
Abstract: Flow matching (FM) is a family of training algorithms for fitting continuous normalizing flows (CNFs). A standard approach to FM, called conditional flow matching (CFM), exploits the fact that the marginal vector field of a CNF can be learned by fitting least-square regression to the so-called conditional vector field specified given one or both ends of the flow path. We show that viewing CFM training from a Bayesian decision theoretic perspective on parameter estimation opens the door to generalizations of CFM algorithms. We propose one such extension by introducing a CFM algorithm based on defining conditional probability paths given what we refer to as ``streams'', instances of latent stochastic paths that connect pairs of noise and observed data. Further, we advocate the modeling of these latent streams using Gaussian processes (GPs). The unique distributional properties of GPs, and in particular the fact that the velocity of a GP is still a GP, allows drawing samples from the resulting stream-augmented conditional probability path without simulating the actual streams, and hence the ``simulation-free" nature of CFM training is preserved. We show that this generalization of the CFM can substantially reduce the variance in the estimated marginal vector field at a moderate computational cost, thereby improving the quality of the generated samples under common metrics. Additionally, we show that adopting the GP on the streams allows for flexibly linking multiple related training data points (e.g., time series) and incorporating additional prior information. We empirically validate our claim through both simulations and applications to two hand-written image datasets.
Authors: Bhrij Patel, Souradip Chakraborty, Wesley A. Suttle, Mengdi Wang, Amrit Singh Bedi, Dinesh Manocha
Abstract: Text-based AI system optimization typically involves a feedback loop scheme where a single LLM generates an evaluation in natural language of the current output to improve the next iteration's output. However, in this work, we empirically demonstrate that for a practical and complex task (code generation) with multiple criteria to evaluate, utilizing only one LLM evaluator tends to let errors in generated code go undetected, thus leading to incorrect evaluations and ultimately suboptimal test case performance. Motivated by this failure case, we assume there exists an optimal evaluation policy that samples an evaluation between response and ground truth. We then theoretically prove that a linear combination of multiple evaluators can approximate this optimal policy. From this insight, we propose AI system optimization via Multiple LLM Evaluators (AIME). AIME is an evaluation protocol that utilizes multiple LLMs that each independently generate an evaluation on separate criteria and then combine them via concatenation. We provide an extensive empirical study showing AIME outperforming baseline methods in code generation tasks, with up to $62\%$ higher error detection rate and up to $16\%$ higher success rate than a single LLM evaluation protocol on LeetCodeHard and HumanEval datasets. We also show that the selection of the number of evaluators and which criteria to utilize is non-trivial as it can impact pact success rate by up to $12\%$.
Authors: Chongyu Fan, Jiancheng Liu, Licong Lin, Jinghan Jia, Ruiqi Zhang, Song Mei, Sijia Liu
Abstract: In this work, we address the problem of large language model (LLM) unlearning, aiming to remove unwanted data influences and associated model capabilities (e.g., copyrighted data or harmful content generation) while preserving essential model utilities, without the need for retraining from scratch. Despite the growing need for LLM unlearning, a principled optimization framework remains lacking. To this end, we revisit the state-of-the-art approach, negative preference optimization (NPO), and identify the issue of reference model bias, which could undermine NPO's effectiveness, particularly when unlearning forget data of varying difficulty. Given that, we propose a simple yet effective unlearning optimization framework, called SimNPO, showing that 'simplicity' in removing the reliance on a reference model (through the lens of simple preference optimization) benefits unlearning. We also provide deeper insights into SimNPO's advantages, supported by analysis using mixtures of Markov chains. Furthermore, we present extensive experiments validating SimNPO's superiority over existing unlearning baselines in benchmarks like TOFU and MUSE, and robustness against relearning attacks. Codes are available at https://github.com/OPTML-Group/Unlearn-Simple.
Authors: Matheus Farias, H. T. Kung
Abstract: We introduce $\textit{sorted weight sectioning}$ (SWS): a weight allocation algorithm that places sorted deep neural network (DNN) weight sections on bit-sliced compute-in-memory (CIM) crossbars to reduce analog-to-digital converter (ADC) energy consumption. Data conversions are the most energy-intensive process in crossbar operation. SWS effectively reduces this cost leveraging (1) small weights and (2) zero weights (weight sparsity). DNN weights follow bell-shaped distributions, with most weights near zero. Using SWS, we only need low-order crossbar columns for sections with low-magnitude weights. This reduces the quantity and resolution of ADCs used, exponentially decreasing ADC energy costs without significantly degrading DNN accuracy. Unstructured sparsification further sharpens the weight distribution with small accuracy loss. However, it presents challenges in hardware tracking of zeros: we cannot switch zero rows to other layer weights in unsorted crossbars without index matching. SWS efficiently addresses unstructured sparse models using offline remapping of zeros into earlier sections, which reveals full sparsity potential and maximizes energy efficiency. Our method reduces ADC energy use by 89.5% on unstructured sparse BERT models. Overall, this paper introduces a novel algorithm to promote energy-efficient CIM crossbars for unstructured sparse DNN workloads.
Authors: Maksuda Akter, Rabea Khatun, Md. Alamin Talukder, Md. Manowarul Islam, Md. Ashraf Uddin
Abstract: Skin cancer is a serious and potentially fatal disease caused by DNA damage. Early detection significantly increases survival rates, making accurate diagnosis crucial. In this groundbreaking study, we present a hybrid framework based on Deep Learning (DL) that achieves precise classification of benign and malignant skin lesions. Our approach begins with dataset preprocessing to enhance classification accuracy, followed by training two separate pre-trained DL models, InceptionV3 and DenseNet121. By fusing the results of each model using the weighted sum rule, our system achieves exceptional accuracy rates. Specifically, we achieve a 92.27% detection accuracy rate, 92.33% sensitivity, 92.22% specificity, 90.81% precision, and 91.57% F1-score, outperforming existing models and demonstrating the robustness and trustworthiness of our hybrid approach. Our study represents a significant advance in skin cancer diagnosis and provides a promising foundation for further research in the field. With the potential to save countless lives through earlier detection, our hybrid deep-learning approach is a game-changer in the fight against skin cancer.
Authors: Srishti Gureja, Lester James V. Miranda, Shayekh Bin Islam, Rishabh Maheshwary, Drishti Sharma, Gusti Winata, Nathan Lambert, Sebastian Ruder, Sara Hooker, Marzieh Fadaee
Abstract: Reward models (RMs) have driven the state-of-the-art performance of LLMs today by enabling the integration of human feedback into the language modeling process. However, RMs are primarily trained and evaluated in English, and their capabilities in multilingual settings remain largely understudied. In this work, we conduct a systematic evaluation of several reward models in multilingual settings. We first construct the first-of-its-kind multilingual RM evaluation benchmark, M-RewardBench, consisting of 2.87k preference instances for 23 typologically diverse languages, that tests the chat, safety, reasoning, and translation capabilities of RMs. We then rigorously evaluate a wide range of reward models on M-RewardBench, offering fresh insights into their performance across diverse languages. We identify a significant gap in RMs' performances between English and non-English languages and show that RM preferences can change substantially from one language to another. We also present several findings on how different multilingual aspects impact RM performance. Specifically, we show that the performance of RMs is improved with improved translation quality. Similarly, we demonstrate that the models exhibit better performance for high-resource languages. We release M-RewardBench dataset and the codebase in this study to facilitate a better understanding of RM evaluation in multilingual settings.
Authors: Arushi Prakash, Dimitrios Bermperidis, Srivas Chennu
Abstract: Large-scale industrial recommendation models predict the most relevant items from catalogs containing millions or billions of options. To train these models efficiently, a small set of irrelevant items (negative samples) is selected from the vast catalog for each relevant item (positive example), helping the model distinguish between relevant and irrelevant items. Choosing the right negative sampling method is a common challenge. We address this by implementing and comparing various negative sampling methods - random, popularity-based, in-batch, mixed, adaptive, and adaptive with mixed variants - on modern sequential recommendation models. Our experiments, including hyperparameter optimization and 20x repeats on three benchmark datasets with varying popularity biases, show how the choice of method and dataset characteristics impact key model performance metrics. We also reveal that average performance metrics often hide imbalances across popularity bands (head, mid, tail). We find that commonly used random negative sampling reinforces popularity bias and performs best for head items. Popularity-based methods (in-batch and global popularity negative sampling) can offer balanced performance at the cost of lower overall model performance results. Our study serves as a practical guide to the trade-offs in selecting a negative sampling method for large-scale sequential recommendation models. Code, datasets, experimental results and hyperparameters are available at: https://github.com/apple/ml-negative-sampling.
Authors: Jerry Huang, Prasanna Parthasarathi, Mehdi Rezagholizadeh, Boxing Chen, Sarath Chandar
Abstract: The growth in prominence of large language models (LLMs) in everyday life can be largely attributed to their generative abilities, yet some of this is also owed to the risks and costs associated with their use. On one front is their tendency to \textit{hallucinate} false or misleading information, limiting their reliability. On another is the increasing focus on the computational limitations associated with traditional self-attention based LLMs, which has brought about new alternatives, in particular recurrent models, meant to overcome them. Yet it remains uncommon to consider these two concerns simultaneously. Do changes in architecture exacerbate/alleviate existing concerns about hallucinations? Do they affect how and where they occur? Through an extensive evaluation, we study how these architecture-based inductive biases affect the propensity to hallucinate. While hallucination remains a general phenomenon not limited to specific architectures, the situations in which they occur and the ease with which specific types of hallucinations can be induced can significantly differ based on the model architecture. These findings highlight the need for better understanding both these problems in conjunction with each other, as well as consider how to design more universal techniques for handling hallucinations.
Authors: Yuzhi Xu, Haowei Ni, Qinhui Gao, Chia-Hua Chang, Yanran Huo, Fanyu Zhao, Shiyu Hu, Wei Xia, Yike Zhang, Radu Grovu, Min He, John. Z. H. Zhang, Yuanqing Wang
Abstract: Computational molecular design -- the endeavor to design molecules, with various missions, aided by machine learning and molecular dynamics approaches, has been widely applied to create valuable new molecular entities, from small molecule therapeutics to protein biologics. In the small data regime, physics-based approaches model the interaction between the molecule being designed and proteins of key physiological functions, providing structural insights into the mechanism. When abundant data has been collected, a quantitative structure-activity relationship (QSAR) can be more directly constructed from experimental data, from which machine learning can distill key insights to guide the design of the next round of experiment design. Machine learning methodologies can also facilitate physical modeling, from improving the accuracy of force fields and extending them to unseen chemical spaces, to more directly enhancing the sampling on the conformational spaces. We argue that these techniques are mature enough to be applied to not just extend the longevity of life, but the beauty it manifests. In this perspective, we review the current frontiers in the research \& development of skin care products, as well as the statistical and physical toolbox applicable to addressing the challenges in this industry. Feasible interdisciplinary research projects are proposed to harness the power of machine learning tools to design innovative, effective, and inexpensive skin care products.
Authors: Lingxiao Li, Kaixiong Gong, Weihong Li, Xili Dai, Tao Chen, Xiaojun Yuan, Xiangyu Yue
Abstract: This paper introduces Bifr\"ost, a novel 3D-aware framework that is built upon diffusion models to perform instruction-based image composition. Previous methods concentrate on image compositing at the 2D level, which fall short in handling complex spatial relationships ($\textit{e.g.}$, occlusion). Bifr\"ost addresses these issues by training MLLM as a 2.5D location predictor and integrating depth maps as an extra condition during the generation process to bridge the gap between 2D and 3D, which enhances spatial comprehension and supports sophisticated spatial interactions. Our method begins by fine-tuning MLLM with a custom counterfactual dataset to predict 2.5D object locations in complex backgrounds from language instructions. Then, the image-compositing model is uniquely designed to process multiple types of input features, enabling it to perform high-fidelity image compositions that consider occlusion, depth blur, and image harmonization. Extensive qualitative and quantitative evaluations demonstrate that Bifr\"ost significantly outperforms existing methods, providing a robust solution for generating realistically composited images in scenarios demanding intricate spatial understanding. This work not only pushes the boundaries of generative image compositing but also reduces reliance on expensive annotated datasets by effectively utilizing existing resources in innovative ways.
Authors: Eric Cai, Octavian Donca, Ben Eisner, David Held
Abstract: The task of "relative placement" is to predict the placement of one object in relation to another, e.g. placing a mug onto a mug rack. Through explicit object-centric geometric reasoning, recent methods for relative placement have made tremendous progress towards data-efficient learning for robot manipulation while generalizing to unseen task variations. However, they have yet to represent deformable transformations, despite the ubiquity of non-rigid bodies in real world settings. As a first step towards bridging this gap, we propose ``cross-displacement" - an extension of the principles of relative placement to geometric relationships between deformable objects - and present a novel vision-based method to learn cross-displacement through dense diffusion. To this end, we demonstrate our method's ability to generalize to unseen object instances, out-of-distribution scene configurations, and multimodal goals on multiple highly deformable tasks (both in simulation and in the real world) beyond the scope of prior works. Supplementary information and videos can be found at https://sites.google.com/view/tax3d-corl-2024 .
Authors: Junwei Ma, Bo Li, Olufemi A. Omitaomu, Ali Mostafavi
Abstract: Power outages have become increasingly frequent, intense, and prolonged in the US due to climate change, aging electrical grids, and rising energy demand. However, largely due to the absence of granular spatiotemporal outage data, we lack data-driven evidence and analytics-based metrics to quantify power system vulnerability. This limitation has hindered the ability to effectively evaluate and address vulnerability to power outages in US communities. Here, we collected ~179 million power outage records at 15-minute intervals across 3022 US contiguous counties (96.15% of the area) from 2014 to 2023. We developed a power system vulnerability assessment framework based on three dimensions (intensity, frequency, and duration) and applied interpretable machine learning models (XGBoost and SHAP) to compute Power System Vulnerability Index (PSVI) at the county level. Our analysis reveals a consistent increase in power system vulnerability over the past decade. We identified 318 counties across 45 states as hotspots for high power system vulnerability, particularly in the West Coast (California and Washington), the East Coast (Florida and the Northeast area), the Great Lakes megalopolis (Chicago-Detroit metropolitan areas), and the Gulf of Mexico (Texas). Heterogeneity analysis indicates that urban counties, counties with interconnected grids, and states with high solar generation exhibit significantly higher vulnerability. Our results highlight the significance of the proposed PSVI for evaluating the vulnerability of communities to power outages. The findings underscore the widespread and pervasive impact of power outages across the country and offer crucial insights to support infrastructure operators, policymakers, and emergency managers in formulating policies and programs aimed at enhancing the resilience of the US power infrastructure.
Authors: Javid Dadashkarimi, Valeria Pena Trujillo, Camilo Jaimes, Lilla Z\"ollei, Malte Hoffmann
Abstract: Automated fetal brain extraction from full-uterus MRI is a challenging task due to variable head sizes, orientations, complex anatomy, and prevalent artifacts. While deep-learning (DL) models trained on synthetic images have been successful in adult brain extraction, adapting these networks for fetal MRI is difficult due to the sparsity of labeled data, leading to increased false-positive predictions. To address this challenge, we propose a test-time strategy that reduces false positives in networks trained on sparse, synthetic labels. The approach uses a breadth-fine search (BFS) to identify a subvolume likely to contain the fetal brain, followed by a deep-focused sliding window (DFS) search to refine the extraction, pooling predictions to minimize false positives. We train models at different window sizes using synthetic images derived from a small number of fetal brain label maps, augmented with random geometric shapes. Each model is trained on diverse head positions and scales, including cases with partial or no brain tissue. Our framework matches state-of-the-art brain extraction methods on clinical HASTE scans of third-trimester fetuses and exceeds them by up to 5\% in terms of Dice in the second trimester as well as EPI scans across both trimesters. Our results demonstrate the utility of a sliding-window approach and combining predictions from several models trained on synthetic images, for improving brain-extraction accuracy by progressively refining regions of interest and minimizing the risk of missing brain mask slices or misidentifying other tissues as brain.