Authors: Cyrile Delestre, Yoann Sola
Abstract: Banking Transaction Flow (BTF) is a sequential data found in a number of banking activities such as marketing, credit risk or banking fraud. It is a multimodal data composed of three modalities: a date, a numerical value and a wording. We propose in this work an application of self-attention mechanism to the processing of BTFs. We trained two general models on a large amount of BTFs in a self-supervised way: one RNN-based model and one Transformer-based model. We proposed a specific tokenization in order to be able to process BTFs. The performance of these two models was evaluated on two banking downstream tasks: a transaction categorization task and a credit risk task. The results show that fine-tuning these two pre-trained models allowed to perform better than the state-of-the-art approaches for both tasks.
Authors: Sukwon Yun, Inyoung Choi, Jie Peng, Yangfan Wu, Jingxuan Bao, Qiyiwen Zhang, Jiayi Xin, Qi Long, Tianlong Chen
Abstract: Multimodal learning has gained increasing importance across various fields, offering the ability to integrate data from diverse sources such as images, text, and personalized records, which are frequently observed in medical domains. However, in scenarios where some modalities are missing, many existing frameworks struggle to accommodate arbitrary modality combinations, often relying heavily on a single modality or complete data. This oversight of potential modality combinations limits their applicability in real-world situations. To address this challenge, we propose Flex-MoE (Flexible Mixture-of-Experts), a new framework designed to flexibly incorporate arbitrary modality combinations while maintaining robustness to missing data. The core idea of Flex-MoE is to first address missing modalities using a new missing modality bank that integrates observed modality combinations with the corresponding missing ones. This is followed by a uniquely designed Sparse MoE framework. Specifically, Flex-MoE first trains experts using samples with all modalities to inject generalized knowledge through the generalized router ($\mathcal{G}$-Router). The $\mathcal{S}$-Router then specializes in handling fewer modality combinations by assigning the top-1 gate to the expert corresponding to the observed modality combination. We evaluate Flex-MoE on the ADNI dataset, which encompasses four modalities in the Alzheimer's Disease domain, as well as on the MIMIC-IV dataset. The results demonstrate the effectiveness of Flex-MoE highlighting its ability to model arbitrary modality combinations in diverse missing modality scenarios. Code is available at https://github.com/UNITES-Lab/flex-moe.
Authors: Jalmari Nevanlinna, Anna Eidst{\o}, Jari Yl\"a-Mattila, Teemu Koivistoinen, Niku Oksala, Juho Kanniainen, Ari Palom\"aki, Antti Roine
Abstract: Emergency department (ED) crowding is a global public health issue that has been repeatedly associated with increased mortality. Predicting future service demand would enable preventative measures aiming to eliminate crowding along with it's detrimental effects. Recent findings in our ED indicate that occupancy ratios exceeding 90% are associated with increased 10-day mortality. In this paper, we aim to predict these crisis periods using retrospective data from a large Nordic ED with a LightGBM model. We provide predictions for the whole ED and individually for it's different operational sections. We demonstrate that afternoon crowding can be predicted at 11 a.m. with an AUC of 0.82 (95% CI 0.78-0.86) and at 8 a.m. with an AUC up to 0.79 (95% CI 0.75-0.83). Consequently we show that forecasting mortality-associated crowding using anonymous administrative data is feasible.
Authors: Ziqi Yang, Zhaopeng Peng, Zihui Wang, Jianzhong Qi, Chaochao Chen, Weike Pan, Chenglu Wen, Cheng Wang, Xiaoliang Fan
Abstract: Cross-domain recommendation (CDR) offers a promising solution to the data sparsity problem by enabling knowledge transfer across source and target domains. However, many recent CDR models overlook crucial issues such as privacy as well as the risk of negative transfer (which negatively impact model performance), especially in multi-domain settings. To address these challenges, we propose FedGCDR, a novel federated graph learning framework that securely and effectively leverages positive knowledge from multiple source domains. First, we design a positive knowledge transfer module that ensures privacy during inter-domain knowledge transmission. This module employs differential privacy-based knowledge extraction combined with a feature mapping mechanism, transforming source domain embeddings from federated graph attention networks into reliable domain knowledge. Second, we design a knowledge activation module to filter out potential harmful or conflicting knowledge from source domains, addressing the issues of negative transfer. This module enhances target domain training by expanding the graph of the target domain to generate reliable domain attentions and fine-tunes the target model for improved negative knowledge filtering and more accurate predictions. We conduct extensive experiments on 16 popular domains of the Amazon dataset, demonstrating that FedGCDR significantly outperforms state-of-the-art methods.
Authors: David D. Baek, Yuxiao Li, Max Tegmark
Abstract: Motivated by interpretability and reliability, we investigate how neural networks represent knowledge during graph learning, We find hints of universality, where equivalent representations are learned across a range of model sizes (from $10^2$ to $10^9$ parameters) and contexts (MLP toy models, LLM in-context learning and LLM training). We show that these attractor representations optimize generalization to unseen examples by exploiting properties of knowledge graph relations (e.g. symmetry and meta-transitivity). We find experimental support for such universality by showing that LLMs and simpler neural networks can be stitched, i.e., by stitching the first part of one model to the last part of another, mediated only by an affine or almost affine transformation. We hypothesize that this dynamic toward simplicity and generalization is driven by "intelligence from starvation": where overfitting is minimized by pressure to minimize the use of resources that are either scarce or competed for against other tasks.
Authors: Cheng Fang, Sicong Liu, Zimu Zhou, Bin Guo, Jiaqi Tang, Ke Ma, Zhiwen Yu
Abstract: On-device adapting to continual, unpredictable domain shifts is essential for mobile applications like autonomous driving and augmented reality to deliver seamless user experiences in evolving environments. Test-time adaptation (TTA) emerges as a promising solution by tuning model parameters with unlabeled live data immediately before prediction. However, TTA's unique forward-backward-reforward pipeline notably increases the latency over standard inference, undermining the responsiveness in time-sensitive mobile applications. This paper presents AdaShadow, a responsive test-time adaptation framework for non-stationary mobile data distribution and resource dynamics via selective updates of adaptation-critical layers. Although the tactic is recognized in generic on-device training, TTA's unsupervised and online context presents unique challenges in estimating layer importance and latency, as well as scheduling the optimal layer update plan. AdaShadow addresses these challenges with a backpropagation-free assessor to rapidly identify critical layers, a unit-based runtime predictor to account for resource dynamics in latency estimation, and an online scheduler for prompt layer update planning. Also, AdaShadow incorporates a memory I/O-aware computation reuse scheme to further reduce latency in the reforward pass. Results show that AdaShadow achieves the best accuracy-latency balance under continual shifts. At low memory and energy costs, Adashadow provides a 2x to 3.5x speedup (ms-level) over state-of-the-art TTA methods with comparable accuracy and a 14.8% to 25.4% accuracy boost over efficient supervised methods with similar latency.
Authors: Sirui Li, Janardhan Kulkarni, Ishai Menache, Cathy Wu, Beibin Li
Abstract: Mixed Integer Linear Programming (MILP) is essential for modeling complex decision-making problems but faces challenges in computational tractability and requires expert formulation. Current deep learning approaches for MILP focus on specific problem classes and do not generalize to unseen classes. To address this shortcoming, we take a foundation model training approach, where we train a single deep learning model on a diverse set of MILP problems to generalize across problem classes. As existing datasets for MILP lack diversity and volume, we introduce MILP-Evolve, a novel LLM-based evolutionary framework that is capable of generating a large set of diverse MILP classes with an unlimited amount of instances. We study our methodology on three key learning tasks that capture diverse aspects of MILP: (1) integrality gap prediction, (2) learning to branch, and (3) a new task of aligning MILP instances with natural language descriptions. Our empirical results show that models trained on the data generated by MILP-Evolve achieve significant improvements on unseen problems, including MIPLIB benchmarks. Our work highlights the potential of moving towards a foundation model approach for MILP that can generalize to a broad range of MILP applications. We are committed to fully open-sourcing our work to advance further research.
Authors: Khashayar Gatmiry, Nikunj Saunshi, Sashank J. Reddi, Stefanie Jegelka, Sanjiv Kumar
Abstract: The remarkable capability of Transformers to do reasoning and few-shot learning, without any fine-tuning, is widely conjectured to stem from their ability to implicitly simulate a multi-step algorithms -- such as gradient descent -- with their weights in a single forward pass. Recently, there has been progress in understanding this complex phenomenon from an expressivity point of view, by demonstrating that Transformers can express such multi-step algorithms. However, our knowledge about the more fundamental aspect of its learnability, beyond single layer models, is very limited. In particular, can training Transformers enable convergence to algorithmic solutions? In this work we resolve this for in-context linear regression with linear looped Transformers -- a multi-layer model with weight sharing that is conjectured to have an inductive bias to learn fix-point iterative algorithms. More specifically, for this setting we show that the global minimizer of the population training loss implements multi-step preconditioned gradient descent, with a preconditioner that adapts to the data distribution. Furthermore, we show a fast convergence for gradient flow on the regression loss, despite the non-convexity of the landscape, by proving a novel gradient dominance condition. To our knowledge, this is the first theoretical analysis for multi-layer Transformer in this setting. We further validate our theoretical findings through synthetic experiments.
Authors: Abu Fuad Ahmad, Md Shohel Sayeed, Khaznah Alshammari, Istiaque Ahmed
Abstract: Machine learning (ML) has become a ubiquitous tool across various domains of data mining and big data analysis. The efficacy of ML models depends heavily on high-quality datasets, which are often complicated by the presence of missing values. Consequently, the performance and generalization of ML models are at risk in the face of such datasets. This paper aims to examine the nuanced impact of missing values on ML workflows, including their types, causes, and consequences. Our analysis focuses on the challenges posed by missing values, including biased inferences, reduced predictive power, and increased computational burdens. The paper further explores strategies for handling missing values, including imputation techniques and removal strategies, and investigates how missing values affect model evaluation metrics and introduces complexities in cross-validation and model selection. The study employs case studies and real-world examples to illustrate the practical implications of addressing missing values. Finally, the discussion extends to future research directions, emphasizing the need for handling missing values ethically and transparently. The primary goal of this paper is to provide insights into the pervasive impact of missing values on ML models and guide practitioners toward effective strategies for achieving robust and reliable model outcomes.
Authors: Haoteng Yin, Rongzhe Wei, Eli Chien, Pan Li
Abstract: Graphs offer unique insights into relationships and interactions between entities, complementing data modalities like text, images, and videos. By incorporating relational information from graph data, AI models can extend their capabilities beyond traditional tasks. However, relational data in sensitive domains such as finance and healthcare often contain private information, making privacy preservation crucial. Existing privacy-preserving methods, such as DP-SGD, which rely on gradient decoupling assumptions, are not well-suited for relational learning due to the inherent dependencies between coupled training samples. To address this challenge, we propose a privacy-preserving relational learning pipeline that decouples dependencies in sampled relations during training, ensuring differential privacy through a tailored application of DP-SGD. We apply this method to fine-tune large language models (LLMs) on sensitive graph data, and tackle the associated computational complexities. Our approach is evaluated on LLMs of varying sizes (e.g., BERT, Llama2) using real-world relational data from four text-attributed graphs. The results demonstrate significant improvements in relational learning tasks, all while maintaining robust privacy guarantees during training. Additionally, we explore the trade-offs between privacy, utility, and computational efficiency, offering insights into the practical deployment of our approach. Code is available at https://github.com/Graph-COM/PvGaLM.
Authors: Timothy L. Cronin IV, Sanmukh Kuppannagari
Abstract: Deep learning technologies, particularly deep neural networks (DNNs), have demonstrated significant success across many domains. This success has been accompanied by substantial advancements and innovations in the algorithms behind the operations required by DNNs. These enhanced algorithms hold the potential to greatly increase the performance of DNNs. However, discovering the best performing algorithm for a DNN and altering the DNN to use such algorithm is a difficult and time consuming task. To address this, we introduce an open source framework which provides easy to use fine grain algorithmic control for DNNs, enabling algorithmic exploration and selection. Along with built-in high performance implementations of common deep learning operations, the framework enables users to implement and select their own algorithms to be utilized by the DNN. The framework's built-in accelerated implementations are shown to yield outputs equivalent to and exhibit similar performance as implementations in PyTorch, a popular DNN framework. Moreover, the framework incurs no additional performance overhead, meaning that performance depends solely on the algorithms chosen by the user.
Authors: Alberto Alfarano, Fran\c{c}ois Charton, Amaury Hayat
Abstract: Despite their spectacular progress, language models still struggle on complex reasoning tasks, such as advanced mathematics. We consider a long-standing open problem in mathematics: discovering a Lyapunov function that ensures the global stability of a dynamical system. This problem has no known general solution, and algorithmic solvers only exist for some small polynomial systems. We propose a new method for generating synthetic training samples from random solutions, and show that sequence-to-sequence transformers trained on such datasets perform better than algorithmic solvers and humans on polynomial systems, and can discover new Lyapunov functions for non-polynomial systems.
Authors: Grigory Malinovsky, Umberto Michieli, Hasan Abed Al Kader Hammoud, Taha Ceritli, Hayder Elesedy, Mete Ozay, Peter Richt\'arik
Abstract: Fine-tuning has become a popular approach to adapting large foundational models to specific tasks. As the size of models and datasets grows, parameter-efficient fine-tuning techniques are increasingly important. One of the most widely used methods is Low-Rank Adaptation (LoRA), with adaptation update expressed as the product of two low-rank matrices. While LoRA was shown to possess strong performance in fine-tuning, it often under-performs when compared to full-parameter fine-tuning (FPFT). Although many variants of LoRA have been extensively studied empirically, their theoretical optimization analysis is heavily under-explored. The starting point of our work is a demonstration that LoRA and its two extensions, Asymmetric LoRA and Chain of LoRA, indeed encounter convergence issues. To address these issues, we propose Randomized Asymmetric Chain of LoRA (RAC-LoRA) -- a general optimization framework that rigorously analyzes the convergence rates of LoRA-based methods. Our approach inherits the empirical benefits of LoRA-style heuristics, but introduces several small but important algorithmic modifications which turn it into a provably convergent method. Our framework serves as a bridge between FPFT and low-rank adaptation. We provide provable guarantees of convergence to the same solution as FPFT, along with the rate of convergence. Additionally, we present a convergence analysis for smooth, non-convex loss functions, covering gradient descent, stochastic gradient descent, and federated learning settings. Our theoretical findings are supported by experimental results.
Authors: Huy Hoang, Tien Mai, Pradeep Varakantham
Abstract: We address the problem of offline learning a policy that avoids undesirable demonstrations. Unlike conventional offline imitation learning approaches that aim to imitate expert or near-optimal demonstrations, our setting involves avoiding undesirable behavior (specified using undesirable demonstrations). To tackle this problem, unlike standard imitation learning where the aim is to minimize the distance between learning policy and expert demonstrations, we formulate the learning task as maximizing a statistical distance, in the space of state-action stationary distributions, between the learning policy and the undesirable policy. This significantly different approach results in a novel training objective that necessitates a new algorithm to address it. Our algorithm, UNIQ, tackles these challenges by building on the inverse Q-learning framework, framing the learning problem as a cooperative (non-adversarial) task. We then demonstrate how to efficiently leverage unlabeled data for practical training. Our method is evaluated on standard benchmark environments, where it consistently outperforms state-of-the-art baselines. The code implementation can be accessed at: https://github.com/hmhuy0/UNIQ.
Authors: Abu Fuad Ahmad, Khaznah Alshammari, Istiaque Ahmed, MD Shohel Sayed
Abstract: In recent times, a considerable number of research studies have been carried out to address the issue of Missing Value Imputation (MVI). MVI aims to provide a primary solution for datasets that have one or more missing attribute values. The advancements in Artificial Intelligence (AI) drive the development of new and improved machine learning (ML) algorithms and methods. The advancements in ML have opened up significant opportunities for effectively imputing these missing values. The main objective of this article is to conduct a comprehensive and rigorous review, as well as analysis, of the state-of-the-art ML applications in MVI methods. This analysis seeks to enhance researchers' understanding of the subject and facilitate the development of robust and impactful interventions in data preprocessing for Data Analytics. The review is performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) technique. More than 100 articles published between 2014 and 2023 are critically reviewed, considering the methods and findings. Furthermore, the latest literature is examined to scrutinize the trends in MVI methods and their evaluation. The accomplishments and limitations of the existing literature are discussed in detail. The survey concludes by identifying the current gaps in research and providing suggestions for future research directions and emerging trends in related fields of interest.
Authors: Yongyi Yang, Core Francisco Park, Ekdeep Singh Lubana, Maya Okawa, Wei Hu, Hidenori Tanaka
Abstract: Prior work has shown that text-conditioned diffusion models can learn to identify and manipulate primitive concepts underlying a compositional data-generating process, enabling generalization to entirely novel, out-of-distribution compositions. Beyond performance evaluations, these studies develop a rich empirical phenomenology of learning dynamics, showing that models generalize sequentially, respecting the compositional hierarchy of the data-generating process. Moreover, concept-centric structures within the data significantly influence a model's speed of learning the ability to manipulate a concept. In this paper, we aim to better characterize these empirical results from a theoretical standpoint. Specifically, we propose an abstraction of prior work's compositional generalization problem by introducing a structured identity mapping (SIM) task, where a model is trained to learn the identity mapping on a Gaussian mixture with structurally organized centroids. We mathematically analyze the learning dynamics of neural networks trained on this SIM task and show that, despite its simplicity, SIM's learning dynamics capture and help explain key empirical observations on compositional generalization with diffusion models identified in prior work. Our theory also offers several new insights -- e.g., we find a novel mechanism for non-monotonic learning dynamics of test loss in early phases of training. We validate our new predictions by training a text-conditioned diffusion model, bridging our simplified framework and complex generative models. Overall, this work establishes the SIM task as a meaningful theoretical abstraction of concept learning dynamics in modern generative models.
Authors: Yinuo Ren, Tesi Xiao, Michael Shavlovsky, Lexing Ying, Holakou Rahmanian
Abstract: In LLM alignment and many other ML applications, one often faces the Multi-Objective Fine-Tuning (MOFT) problem, i.e. fine-tuning an existing model with datasets labeled w.r.t. different objectives simultaneously. To address the challenge, we propose the HyperDPO framework, a hypernetwork-based approach that extends the Direct Preference Optimization (DPO) technique, originally developed for efficient LLM alignment with preference data, to accommodate the MOFT settings. By substituting the Bradley-Terry-Luce model in DPO with the Plackett-Luce model, our framework is capable of handling a wide range of MOFT tasks that involve listwise ranking datasets. Compared with previous approaches, HyperDPO enjoys an efficient one-shot training process for profiling the Pareto front of auxiliary objectives, and offers flexible post-training control over trade-offs. Additionally, we propose a novel Hyper Prompt Tuning design, that conveys continuous weight across objectives to transformer-based models without altering their architecture. We demonstrate the effectiveness and efficiency of the HyperDPO framework through its applications to various tasks, including Learning-to-Rank (LTR) and LLM alignment, highlighting its viability for large-scale ML deployments.
Authors: Youzuo Lin, Shihang Feng, James Theiler, Yinpeng Chen, Umberto Villa, Jing Rao, John Greenhall, Cristian Pantea, Mark A. Anastasio, Brendt Wohlberg
Abstract: Computational wave imaging (CWI) extracts hidden structure and physical properties of a volume of material by analyzing wave signals that traverse that volume. Applications include seismic exploration of the Earth's subsurface, acoustic imaging and non-destructive testing in material science, and ultrasound computed tomography in medicine. Current approaches for solving CWI problems can be divided into two categories: those rooted in traditional physics, and those based on deep learning. Physics-based methods stand out for their ability to provide high-resolution and quantitatively accurate estimates of acoustic properties within the medium. However, they can be computationally intensive and are susceptible to ill-posedness and nonconvexity typical of CWI problems. Machine learning-based computational methods have recently emerged, offering a different perspective to address these challenges. Diverse scientific communities have independently pursued the integration of deep learning in CWI. This review delves into how contemporary scientific machine-learning (ML) techniques, and deep neural networks in particular, have been harnessed to tackle CWI problems. We present a structured framework that consolidates existing research spanning multiple domains, including computational imaging, wave physics, and data science. This study concludes with important lessons learned from existing ML-based methods and identifies technical hurdles and emerging trends through a systematic analysis of the extensive literature on this topic.
Authors: Yurong Liu, R. Teal Witter, Flip Korn, Tarfah Alrashed, Dimitris Paparas, Juliana Freire
Abstract: Banzhaf values offer a simple and interpretable alternative to the widely-used Shapley values. We introduce Kernel Banzhaf, a novel algorithm inspired by KernelSHAP, that leverages an elegant connection between Banzhaf values and linear regression. Through extensive experiments on feature attribution tasks, we demonstrate that Kernel Banzhaf substantially outperforms other algorithms for estimating Banzhaf values in both sample efficiency and robustness to noise. Furthermore, we prove theoretical guarantees on the algorithm's performance, establishing Kernel Banzhaf as a valuable tool for interpretable machine learning.
Authors: Zitong Huang, Mansooreh Montazerin, Ajitesh Srivastava
Abstract: Neural networks are trained by choosing an architecture and training the parameters. The choice of architecture is often by trial and error or with Neural Architecture Search (NAS) methods. While NAS provides some automation, it often relies on discrete steps that optimize the architecture and then train the parameters. We introduce a novel neural network training framework that fundamentally transforms the process by learning architecture and parameters simultaneously with gradient descent. With the appropriate setting of the loss function, it can discover sparse and compact neural networks for given datasets. Central to our approach is a multi-scale encoder-decoder, in which the encoder embeds pairs of neural networks with similar functionalities close to each other (irrespective of their architectures and weights). To train a neural network with a given dataset, we randomly sample a neural network embedding in the embedding space and then perform gradient descent using our custom loss function, which incorporates a sparsity penalty to encourage compactness. The decoder generates a neural network corresponding to the embedding. Experiments demonstrate that our framework can discover sparse and compact neural networks maintaining a high performance.
Authors: Jacob Beck, Shikha Surana, Manus McAuliffe, Oliver Bent, Thomas D. Barrett, Juan Jose Garau Luis, Paul Duckworth
Abstract: Predicting the biophysical and functional properties of proteins is essential for in silico protein design. Machine learning has emerged as a promising technique for such prediction tasks. However, the relative scarcity of in vitro annotations means that these models often have little, or no, specific data on the desired fitness prediction task. As a result of limited data, protein language models (PLMs) are typically trained on general protein sequence modeling tasks, and then fine-tuned, or applied zero-shot, to protein fitness prediction. When no task data is available, the models make strong assumptions about the correlation between the protein sequence likelihood and fitness scores. In contrast, we propose meta-learning over a distribution of standard fitness prediction tasks, and demonstrate positive transfer to unseen fitness prediction tasks. Our method, called Metalic (Meta-Learning In-Context), uses in-context learning and fine-tuning, when data is available, to adapt to new tasks. Crucially, fine-tuning enables considerable generalization, even though it is not accounted for during meta-training. Our fine-tuned models achieve strong results with 18 times fewer parameters than state-of-the-art models. Moreover, our method sets a new state-of-the-art in low-data settings on ProteinGym, an established fitness-prediction benchmark. Due to data scarcity, we believe meta-learning will play a pivotal role in advancing protein engineering.
Authors: Anuran Makur, Japneet Singh
Abstract: The Bradley-Terry-Luce (BTL) model is one of the most widely used models for ranking a collection of items or agents based on pairwise comparisons among them. Given $n$ agents, the BTL model endows each agent $i$ with a latent skill score $\alpha_i > 0$ and posits that the probability that agent $i$ is preferred over agent $j$ is $\alpha_i/(\alpha_i + \alpha_j)$. In this work, our objective is to formulate a hypothesis test that determines whether a given pairwise comparison dataset, with $k$ comparisons per pair of agents, originates from an underlying BTL model. We formalize this testing problem in the minimax sense and define the critical threshold of the problem. We then establish upper bounds on the critical threshold for general induced observation graphs (satisfying mild assumptions) and develop lower bounds for complete induced graphs. Our bounds demonstrate that for complete induced graphs, the critical threshold scales as $\Theta((nk)^{-1/2})$ in a minimax sense. In particular, our test statistic for the upper bounds is based on a new approximation we derive for the separation distance between general pairwise comparison models and the class of BTL models. To further assess the performance of our statistical test, we prove upper bounds on the type I and type II probabilities of error. Much of our analysis is conducted within the context of a fixed observation graph structure, where the graph possesses certain ``nice'' properties, such as expansion and bounded principal ratio. Additionally, we derive several auxiliary results, such as bounds on principal ratios of graphs, $\ell^2$-bounds on BTL parameter estimation under model mismatch, stability of rankings under the BTL model, etc. We validate our theoretical results through experiments on synthetic and real-world datasets and propose a data-driven permutation testing approach to determine test thresholds.
Authors: Raphael C. Kim, Falco J. Bargagli-Stoffi, Kevin L. Chen, Rachel C. Nethery
Abstract: The substantial effect of air pollution on cardiovascular disease and mortality burdens is well-established. Emissions-reducing interventions on coal-fired power plants -- a major source of hazardous air pollution -- have proven to be an effective, but costly, strategy for reducing pollution-related health burdens. Targeting the power plants that achieve maximum health benefits while satisfying realistic cost constraints is challenging. The primary difficulty lies in quantifying the health benefits of intervening at particular plants. This is further complicated because interventions are applied on power plants, while health impacts occur in potentially distant communities, a setting known as bipartite network interference (BNI). In this paper, we introduce novel policy learning methods based on Q- and A-Learning to determine the optimal policy under arbitrary BNI. We derive asymptotic properties and demonstrate finite sample efficacy in simulations. We apply our novel methods to a comprehensive dataset of Medicare claims, power plant data, and pollution transport networks. Our goal is to determine the optimal strategy for installing power plant scrubbers to minimize ischemic heart disease (IHD) hospitalizations under various cost constraints. We find that annual IHD hospitalization rates could be reduced in a range from 20.66-44.51 per 10,000 person-years through optimal policies under different cost constraints.
Authors: Wilson Yan, Matei Zaharia, Volodymyr Mnih, Pieter Abbeel, Aleksandra Faust, Hao Liu
Abstract: Efficient video tokenization remains a key bottleneck in learning general purpose vision models that are capable of processing long video sequences. Prevailing approaches are restricted to encoding videos to a fixed number of tokens, where too few tokens will result in overly lossy encodings, and too many tokens will result in prohibitively long sequence lengths. In this work, we introduce ElasticTok, a method that conditions on prior frames to adaptively encode a frame into a variable number of tokens. To enable this in a computationally scalable way, we propose a masking technique that drops a random number of tokens at the end of each frames's token encoding. During inference, ElasticTok can dynamically allocate tokens when needed -- more complex data can leverage more tokens, while simpler data only needs a few tokens. Our empirical evaluations on images and video demonstrate the effectiveness of our approach in efficient token usage, paving the way for future development of more powerful multimodal models, world models, and agents.
Authors: Andy K Zhang, Kevin Klyman, Yifan Mai, Yoav Levine, Yian Zhang, Rishi Bommasani, Percy Liang
Abstract: Language models are extensively evaluated, but correctly interpreting evaluation results requires knowledge of train-test overlap which refers to the extent to which the language model is trained on the very data it is being tested on. The public currently lacks adequate information about train-test overlap: most models have no public train-test overlap statistics, and third parties cannot directly measure train-test overlap since they do not have access to the training data. To make this clear, we document the practices of 30 model developers, finding that just 9 developers report train-test overlap: 4 developers release training data under open-source licenses, enabling the community to directly measure train-test overlap, and 5 developers publish their train-test overlap methodology and statistics. By engaging with language model developers, we provide novel information about train-test overlap for three additional developers. Overall, we take the position that language model developers should publish train-test overlap statistics and/or training data whenever they report evaluation results on public test sets. We hope our work increases transparency into train-test overlap to increase the community-wide trust in model evaluations.
Authors: Brooke Feinberg, Aiwen Li
Abstract: We build upon a recently introduced class of quasi-graph random features (q-GRFs), which have demonstrated the ability to yield lower variance estimators of the 2-regularized Laplacian kernel (Choromanski 2023). Our research investigates whether similar results can be achieved with alternative kernel functions, specifically the Diffusion (or Heat), Mat\'ern, and Inverse Cosine kernels. We find that the Diffusion kernel performs most similarly to the 2-regularized Laplacian, and we further explore graph types that benefit from the previously established antithetic termination procedure. Specifically, we explore Erd\H{o}s-R\'enyi and Barab\'asi-Albert random graph models, Binary Trees, and Ladder graphs, with the goal of identifying combinations of specific kernel and graph type that benefit from antithetic termination. We assert that q-GRFs achieve lower variance estimators of the Diffusion (or Heat) kernel on Ladder graphs. However, the number of rungs on the Ladder graphs impacts the algorithm's performance; further theoretical results supporting our experimentation are forthcoming. This work builds upon some of the earliest Quasi-Monte Carlo methods for kernels defined on combinatorial objects, paving the way for kernel-based learning algorithms and future real-world applications in various domains.
Authors: Kiwhan Song, Mohamed Ali Dhraief, Muhua Xu, Locke Cai, Xuhao Chen, Arvind, Jie Chen
Abstract: Anti-Money Laundering (AML) involves the identification of money laundering crimes in financial activities, such as cryptocurrency transactions. Recent studies advanced AML through the lens of graph-based machine learning, modeling the web of financial transactions as a graph and developing graph methods to identify suspicious activities. For instance, a recent effort on opensourcing datasets and benchmarks, Elliptic2, treats a set of Bitcoin addresses, considered to be controlled by the same entity, as a graph node and transactions among entities as graph edges. This modeling reveals the "shape" of a money laundering scheme - a subgraph on the blockchain. Despite the attractive subgraph classification results benchmarked by the paper, competitive methods remain expensive to apply due to the massive size of the graph; moreover, existing methods require candidate subgraphs as inputs which may not be available in practice. In this work, we introduce RevTrack, a graph-based framework that enables large-scale AML analysis with a lower cost and a higher accuracy. The key idea is to track the initial senders and the final receivers of funds; these entities offer a strong indication of the nature (licit vs. suspicious) of their respective subgraph. Based on this framework, we propose RevClassify, which is a neural network model for subgraph classification. Additionally, we address the practical problem where subgraph candidates are not given, by proposing RevFilter. This method identifies new suspicious subgraphs by iteratively filtering licit transactions, using RevClassify. Benchmarking these methods on Elliptic2, a new standard for AML, we show that RevClassify outperforms state-of-the-art subgraph classification techniques in both cost and accuracy. Furthermore, we demonstrate the effectiveness of RevFilter in discovering new suspicious subgraphs, confirming its utility for practical AML.
Authors: Aida Mohammadshahi, Yani Ioannou
Abstract: Knowledge Distillation is a commonly used Deep Neural Network compression method, which often maintains overall generalization performance. However, we show that even for balanced image classification datasets, such as CIFAR-100, Tiny ImageNet and ImageNet, as many as 41% of the classes are statistically significantly affected by distillation when comparing class-wise accuracy (i.e. class bias) between a teacher/distilled student or distilled student/non-distilled student model. Changes in class bias are not necessarily an undesirable outcome when considered outside of the context of a model's usage. Using two common fairness metrics, Demographic Parity Difference (DPD) and Equalized Odds Difference (EOD) on models trained with the CelebA, Trifeature, and HateXplain datasets, our results suggest that increasing the distillation temperature improves the distilled student model's fairness -- for DPD, the distilled student even surpasses the fairness of the teacher model at high temperatures. This study highlights the uneven effects of Knowledge Distillation on certain classes and its potentially significant role in fairness, emphasizing that caution is warranted when using distilled models for sensitive application domains.
Authors: Michael T. Pearce, Thomas Dooms, Alice Rigg, Jose M. Oramas, Lee Sharkey
Abstract: A mechanistic understanding of how MLPs do computation in deep neural networks remains elusive. Current interpretability work can extract features from hidden activations over an input dataset but generally cannot explain how MLP weights construct features. One challenge is that element-wise nonlinearities introduce higher-order interactions and make it difficult to trace computations through the MLP layer. In this paper, we analyze bilinear MLPs, a type of Gated Linear Unit (GLU) without any element-wise nonlinearity that nevertheless achieves competitive performance. Bilinear MLPs can be fully expressed in terms of linear operations using a third-order tensor, allowing flexible analysis of the weights. Analyzing the spectra of bilinear MLP weights using eigendecomposition reveals interpretable low-rank structure across toy tasks, image classification, and language modeling. We use this understanding to craft adversarial examples, uncover overfitting, and identify small language model circuits directly from the weights alone. Our results demonstrate that bilinear layers serve as an interpretable drop-in replacement for current activation functions and that weight-based interpretability is viable for understanding deep-learning models.
Authors: Ran Liu, Wenrui Ma, Ellen Zippi, Hadi Pouransari, Jingyun Xiao, Chris Sandino, Behrooz Mahasseni, Juri Minxha, Erdrin Azemi, Eva L. Dyer, Ali Moin
Abstract: Time series data are inherently functions of time, yet current transformers often learn time series by modeling them as mere concatenations of time periods, overlooking their functional properties. In this work, we propose a novel objective for transformers that learn time series by re-interpreting them as temporal functions. We build an alternative sequence of time series by constructing degradation operators of different intensity in the functional space, creating augmented variants of the original sample that are abstracted or simplified to different degrees. Based on the new set of generated sequence, we train an autoregressive transformer that progressively recovers the original sample from the most simplified variant. Analogous to the next word prediction task in languages that learns narratives by connecting different words, our autoregressive transformer aims to learn the Narratives of Time Series (NoTS) by connecting different functions in time. Theoretically, we justify the construction of the alternative sequence through its advantages in approximating functions. When learning time series data with transformers, constructing sequences of temporal functions allows for a broader class of approximable functions (e.g., differentiation) compared to sequences of time periods, leading to a 26\% performance improvement in synthetic feature regression experiments. Experimentally, we validate NoTS in 3 different tasks across 22 real-world datasets, where we show that NoTS significantly outperforms other pre-training methods by up to 6\%. Additionally, combining NoTS on top of existing transformer architectures can consistently boost the performance. Our results demonstrate the potential of NoTS as a general-purpose dynamic learner, offering a viable alternative for developing foundation models for time series analysis.
Authors: Youngwoo Kwon, Qian Qin, Guanyang Wang, Yuchen Wei
Abstract: Restricted Boltzmann Machines are a class of undirected graphical models that play a key role in deep learning and unsupervised learning. In this study, we prove a phase transition phenomenon in the mixing time of the Gibbs sampler for a one-parameter Restricted Boltzmann Machine. Specifically, the mixing time varies logarithmically, polynomially, and exponentially with the number of vertices depending on whether the parameter $c$ is above, equal to, or below a critical value $c_\star\approx-5.87$. A key insight from our analysis is the link between the Gibbs sampler and a dynamical system, which we utilize to quantify the former based on the behavior of the latter. To study the critical case $c= c_\star$, we develop a new isoperimetric inequality for the sampler's stationary distribution by showing that the distribution is nearly log-concave.
Authors: Zain Sarwar, Van Tran, Arjun Nitin Bhagoji, Nick Feamster, Ben Y. Zhao, Supriyo Chakraborty
Abstract: Machine learning (ML) models often require large amounts of data to perform well. When the available data is limited, model trainers may need to acquire more data from external sources. Often, useful data is held by private entities who are hesitant to share their data due to propriety and privacy concerns. This makes it challenging and expensive for model trainers to acquire the data they need to improve model performance. To address this challenge, we propose Mycroft, a data-efficient method that enables model trainers to evaluate the relative utility of different data sources while working with a constrained data-sharing budget. By leveraging feature space distances and gradient matching, Mycroft identifies small but informative data subsets from each owner, allowing model trainers to maximize performance with minimal data exposure. Experimental results across four tasks in two domains show that Mycroft converges rapidly to the performance of the full-information baseline, where all data is shared. Moreover, Mycroft is robust to noise and can effectively rank data owners by utility. Mycroft can pave the way for democratized training of high performance ML models.
Authors: Josiah C. Kratz, Jacob Adamczyk
Abstract: Many organisms and cell types, from bacteria to cancer cells, exhibit a remarkable ability to adapt to fluctuating environments. Additionally, cells can leverage memory of past environments to better survive previously-encountered stressors. From a control perspective, this adaptability poses significant challenges in driving cell populations toward extinction, and is thus an open question with great clinical significance. In this work, we focus on drug dosing in cell populations exhibiting phenotypic plasticity. For specific dynamical models switching between resistant and susceptible states, exact solutions are known. However, when the underlying system parameters are unknown, and for complex memory-based systems, obtaining the optimal solution is currently intractable. To address this challenge, we apply reinforcement learning (RL) to identify informed dosing strategies to control cell populations evolving under novel non-Markovian dynamics. We find that model-free deep RL is able to recover exact solutions and control cell populations even in the presence of long-range temporal dynamics.
Authors: Dom Nasrabadi
Abstract: We introduce JurEE, an ensemble of efficient, encoder-only transformer models designed to strengthen safeguards in AI-User interactions within LLM-based systems. Unlike existing LLM-as-Judge methods, which often struggle with generalization across risk taxonomies and only provide textual outputs, JurEE offers probabilistic risk estimates across a wide range of prevalent risks. Our approach leverages diverse data sources and employs progressive synthetic data generation techniques, including LLM-assisted augmentation, to enhance model robustness and performance. We create an in-house benchmark comprising of other reputable benchmarks such as the OpenAI Moderation Dataset and ToxicChat, where we find JurEE significantly outperforms baseline models, demonstrating superior accuracy, speed, and cost-efficiency. This makes it particularly suitable for applications requiring stringent content moderation, such as customer-facing chatbots. The encoder-ensemble's modular design allows users to set tailored risk thresholds, enhancing its versatility across various safety-related applications. JurEE's collective decision-making process, where each specialized encoder model contributes to the final output, not only improves predictive accuracy but also enhances interpretability. This approach provides a more efficient, performant, and economical alternative to traditional LLMs for large-scale implementations requiring robust content moderation.
Authors: Vikram Krishnamurthy, Cristian Rojas
Abstract: We consider word-of-mouth social learning involving $m$ Kalman filter agents that operate sequentially. The first Kalman filter receives the raw observations, while each subsequent Kalman filter receives a noisy measurement of the conditional mean of the previous Kalman filter. The prior is updated by the $m$-th Kalman filter. When $m=2$, and the observations are noisy measurements of a Gaussian random variable, the covariance goes to zero as $k^{-1/3}$ for $k$ observations, instead of $O(k^{-1})$ in the standard Kalman filter. In this paper we prove that for $m$ agents, the covariance decreases to zero as $k^{-(2^m-1)}$, i.e, the learning slows down exponentially with the number of agents. We also show that by artificially weighing the prior at each time, the learning rate can be made optimal as $k^{-1}$. The implication is that in word-of-mouth social learning, artificially re-weighing the prior can yield the optimal learning rate.
Authors: George Yin, Vikram Krishnamurthy
Abstract: We analyze the finite sample regret of a decreasing step size stochastic gradient algorithm. We assume correlated noise and use a perturbed Lyapunov function as a systematic approach for the analysis. Finally we analyze the escape time of the iterates using large deviations theory.
Authors: Yuting Xie, Xianda Guo, Cong Wang, Kunhua Liu, Long Chen
Abstract: Safety-critical scenarios are infrequent in natural driving environments but hold significant importance for the training and testing of autonomous driving systems. The prevailing approach involves generating safety-critical scenarios automatically in simulation by introducing adversarial adjustments to natural environments. These adjustments are often tailored to specific tested systems, thereby disregarding their transferability across different systems. In this paper, we propose AdvDiffuser, an adversarial framework for generating safety-critical driving scenarios through guided diffusion. By incorporating a diffusion model to capture plausible collective behaviors of background vehicles and a lightweight guide model to effectively handle adversarial scenarios, AdvDiffuser facilitates transferability. Experimental results on the nuScenes dataset demonstrate that AdvDiffuser, trained on offline driving logs, can be applied to various tested systems with minimal warm-up episode data and outperform other existing methods in terms of realism, diversity, and adversarial performance.
Authors: Xin Jiang, Xu Cheng, Zechao Li
Abstract: Pre-training has exhibited notable benefits to downstream tasks by boosting accuracy and speeding up convergence, but the exact reasons for these benefits still remain unclear. To this end, we propose to quantitatively and explicitly explain effects of pre-training on the downstream task from a novel game-theoretic view, which also sheds new light into the learning behavior of deep neural networks (DNNs). Specifically, we extract and quantify the knowledge encoded by the pre-trained model, and further track the changes of such knowledge during the fine-tuning process. Interestingly, we discover that only a small amount of pre-trained model's knowledge is preserved for the inference of downstream tasks. However, such preserved knowledge is very challenging for a model training from scratch to learn. Thus, with the help of this exclusively learned and useful knowledge, the model fine-tuned from pre-training usually achieves better performance than the model training from scratch. Besides, we discover that pre-training can guide the fine-tuned model to learn target knowledge for the downstream task more directly and quickly, which accounts for the faster convergence of the fine-tuned model.
Authors: Abhijnan Nath, Changsoo Jung, Ethan Seefried, Nikhil Krishnaswamy
Abstract: Reward modeling of human preferences is one of the cornerstones of building usable generative large language models (LLMs). While traditional RLHF-based alignment methods explicitly maximize the expected rewards from a separate reward model, more recent supervised alignment methods like Direct Preference Optimization (DPO) circumvent this phase to avoid problems including model drift and reward overfitting. Although popular due to its simplicity, DPO and similar direct alignment methods can still lead to degenerate policies, and rely heavily on the Bradley-Terry-based preference formulation to model reward differences between pairs of candidate outputs. This formulation is challenged by non-deterministic or noisy preference labels, for example human scoring of two candidate outputs is of low confidence. In this paper, we introduce DRDO (Direct Reward Distillation and policy-Optimization), a supervised knowledge distillation-based preference alignment method that simultaneously models rewards and preferences to avoid such degeneracy. DRDO directly mimics rewards assigned by an oracle while learning human preferences from a novel preference likelihood formulation. Our experimental results on the Ultrafeedback and TL;DR datasets demonstrate that policies trained using DRDO surpass previous methods such as DPO and e-DPO in terms of expected rewards and are more robust, on average, to noisy preference signals as well as out-of-distribution (OOD) settings.
Authors: Eunji Kim, Kyuhong Shim, Simyung Chang, Sungroh Yoon
Abstract: A text encoder within Vision-Language Models (VLMs) like CLIP plays a crucial role in translating textual input into an embedding space shared with images, thereby facilitating the interpretative analysis of vision tasks through natural language. Despite the varying significance of different textual elements within a sentence depending on the context, efforts to account for variation of importance in constructing text embeddings have been lacking. We propose a framework of Semantic Token Reweighting to build Interpretable text embeddings (SToRI), which incorporates controllability as well. SToRI refines the text encoding process in CLIP by differentially weighting semantic elements based on contextual importance, enabling finer control over emphasis responsive to data-driven insights and user preferences. The efficacy of SToRI is demonstrated through comprehensive experiments on few-shot image classification and image retrieval tailored to user preferences.
Authors: Guangrui Yang, Ming Li, Han Feng, Xiaosheng Zhuang
Abstract: Graph convolutional networks (GCNs) have emerged as powerful models for graph learning tasks, exhibiting promising performance in various domains. While their empirical success is evident, there is a growing need to understand their essential ability from a theoretical perspective. Existing theoretical research has primarily focused on the analysis of single-layer GCNs, while a comprehensive theoretical exploration of the stability and generalization of deep GCNs remains limited. In this paper, we bridge this gap by delving into the stability and generalization properties of deep GCNs, aiming to provide valuable insights by characterizing rigorously the associated upper bounds. Our theoretical results reveal that the stability and generalization of deep GCNs are influenced by certain key factors, such as the maximum absolute eigenvalue of the graph filter operators and the depth of the network. Our theoretical studies contribute to a deeper understanding of the stability and generalization properties of deep GCNs, potentially paving the way for developing more reliable and well-performing models.
Authors: Bowei Zhu, Shaojie Li, Yong Liu
Abstract: Minimax problems have achieved success in machine learning such as adversarial training, robust optimization, reinforcement learning. For theoretical analysis, current optimal excess risk bounds, which are composed by generalization error and optimization error, present 1/n-rates in strongly-convex-strongly-concave (SC-SC) settings. Existing studies mainly focus on minimax problems with specific algorithms for optimization error, with only a few studies on generalization performance, which limit better excess risk bounds. In this paper, we study the generalization bounds measured by the gradients of primal functions using uniform localized convergence. We obtain a sharper high probability generalization error bound for nonconvex-strongly-concave (NC-SC) stochastic minimax problems. Furthermore, we provide dimension-independent results under Polyak-Lojasiewicz condition for the outer layer. Based on our generalization error bound, we analyze some popular algorithms such as empirical saddle point (ESP), gradient descent ascent (GDA) and stochastic gradient descent ascent (SGDA). We derive better excess primal risk bounds with further reasonable assumptions, which, to the best of our knowledge, are n times faster than exist results in minimax problems.
Authors: Yinan Feng, Yinpeng Chen, Yueh Lee, Youzuo Lin
Abstract: Computational imaging plays a vital role in various scientific and medical applications, such as Full Waveform Inversion (FWI), Computed Tomography (CT), and Electromagnetic (EM) inversion. These methods address inverse problems by reconstructing physical properties (e.g., the acoustic velocity map in FWI) from measurement data (e.g., seismic waveform data in FWI), where both modalities are governed by complex mathematical equations. In this paper, we empirically demonstrate that despite their differing governing equations, three inverse problems (FWI, CT, and EM inversion) share a hidden property within their latent spaces. Specifically, using FWI as an example, we show that both modalities (the velocity map and seismic waveform data) follow the same set of one-way wave equations in the latent space, yet have distinct initial conditions that are linearly correlated. This suggests that after projection into the latent embedding space, the two modalities correspond to different solutions of the same equation, connected through their initial conditions. Our experiments confirm that this hidden property is consistent across all three imaging problems, providing a novel perspective for understanding these computational imaging tasks.
Authors: Binghui Li, Yuanzhi Li
Abstract: Adversarial training is a widely-applied approach to training deep neural networks to be robust against adversarial perturbation. However, although adversarial training has achieved empirical success in practice, it still remains unclear why adversarial examples exist and how adversarial training methods improve model robustness. In this paper, we provide a theoretical understanding of adversarial examples and adversarial training algorithms from the perspective of feature learning theory. Specifically, we focus on a multiple classification setting, where the structured data can be composed of two types of features: the robust features, which are resistant to perturbation but sparse, and the non-robust features, which are susceptible to perturbation but dense. We train a two-layer smoothed ReLU convolutional neural network to learn our structured data. First, we prove that by using standard training (gradient descent over the empirical risk), the network learner primarily learns the non-robust feature rather than the robust feature, which thereby leads to the adversarial examples that are generated by perturbations aligned with negative non-robust feature directions. Then, we consider the gradient-based adversarial training algorithm, which runs gradient ascent to find adversarial examples and runs gradient descent over the empirical risk at adversarial examples to update models. We show that the adversarial training method can provably strengthen the robust feature learning and suppress the non-robust feature learning to improve the network robustness. Finally, we also empirically validate our theoretical findings with experiments on real-image datasets, including MNIST, CIFAR10 and SVHN.
Authors: Shantanu Ghosh, Tiankang Xie, Mikhail Kuznetsov
Abstract: Machine learning (ML) models trained using Empirical Risk Minimization (ERM) often exhibit systematic errors on specific subpopulations of tabular data, known as error slices. Learning robust representation in presence of error slices is challenging, especially in self-supervised settings during the feature reconstruction phase, due to high cardinality features and the complexity of constructing error sets. Traditional robust representation learning methods are largely focused on improving worst group performance in supervised setting in computer vision, leaving a gap in approaches tailored for tabular data. We address this gap by developing a framework to learn robust representation in tabular data during self-supervised pre-training. Our approach utilizes an encoder-decoder model trained with Masked Language Modeling (MLM) loss to learn robust latent representations. This paper applies the Just Train Twice (JTT) and Deep Feature Reweighting (DFR) methods during the pre-training phase for tabular data. These methods fine-tune the ERM pre-trained model by up-weighting error-prone samples or creating balanced datasets for specific categorical features. This results in specialized models for each feature, which are then used in an ensemble approach to enhance downstream classification performance. This methodology improves robustness across slices, thus enhancing overall generalization performance. Extensive experiments across various datasets demonstrate the efficacy of our approach.
Authors: Mohit Gupta, Debjit Bhowmick, Meead Saberi, Shirui Pan, Ben Beck
Abstract: Accurate bicycling volume estimation is crucial for making informed decisions about future investments in bicycling infrastructure. Traditional link-level volume estimation models are effective for motorised traffic but face significant challenges when applied to the bicycling context because of sparse data and the intricate nature of bicycling mobility patterns. To the best of our knowledge, we present the first study to utilize a Graph Convolutional Network (GCN) architecture to model link-level bicycling volumes. We estimate the Annual Average Daily Bicycle (AADB) counts across the City of Melbourne, Australia using Strava Metro bicycling count data. To evaluate the effectiveness of the GCN model, we benchmark it against traditional machine learning models, such as linear regression, support vector machines, and random forest. Our results show that the GCN model performs better than these traditional models in predicting AADB counts, demonstrating its ability to capture the spatial dependencies inherent in bicycle traffic data. We further investigate how varying levels of data sparsity affect performance of the GCN architecture. The GCN architecture performs well and better up to 80% sparsity level, but its limitations become apparent as the data sparsity increases further, emphasizing the need for further research on handling extreme data sparsity in bicycling volume estimation. Our findings offer valuable insights for city planners aiming to improve bicycling infrastructure and promote sustainable transportation.
Authors: Junchao Lin, Zenan Ling, Zhanbo Feng, Feng Zhou, Jingwen Xu, Robert C Qiu
Abstract: Implicit graph neural networks (IGNNs), which exhibit strong expressive power with a single layer, have recently demonstrated remarkable performance in capturing long-range dependencies (LRD) in underlying graphs while effectively mitigating the over-smoothing problem. However, IGNNs rely on computationally expensive fixed-point iterations, which lead to significant speed and scalability limitations, hindering their application to large-scale graphs. To achieve fast fixed-point solving for IGNNs, we propose a novel graph neural solver, IGNN-Solver, which leverages the generalized Anderson Acceleration method, parameterized by a small GNN, and learns iterative updates as a graph-dependent temporal process. Extensive experiments demonstrate that the IGNN-Solver significantly accelerates inference, achieving a $1.5\times$ to $8\times$ speedup without sacrificing accuracy. Moreover, this advantage becomes increasingly pronounced as the graph scale grows, facilitating its large-scale deployment in real-world applications.
Authors: Aldo Gael Carranza, Susan Athey
Abstract: We consider the problem of using observational bandit feedback data from multiple heterogeneous data sources to learn a personalized decision policy that robustly generalizes across diverse target settings. To achieve this, we propose a minimax regret optimization objective to ensure uniformly low regret under general mixtures of the source distributions. We develop a policy learning algorithm tailored to this objective, combining doubly robust offline policy evaluation techniques and no-regret learning algorithms for minimax optimization. Our regret analysis shows that this approach achieves the minimal worst-case mixture regret up to a moderated vanishing rate of the total data across all sources. Our analysis, extensions, and experimental results demonstrate the benefits of this approach for learning robust decision policies from multiple data sources.
Authors: Xinran Li, Ling Pan, Jun Zhang
Abstract: In multi-agent reinforcement learning (MARL), parameter sharing is commonly employed to enhance sample efficiency. However, the popular approach of full parameter sharing often leads to homogeneous policies among agents, potentially limiting the performance benefits that could be derived from policy diversity. To address this critical limitation, we introduce \emph{Kaleidoscope}, a novel adaptive partial parameter sharing scheme that fosters policy heterogeneity while still maintaining high sample efficiency. Specifically, Kaleidoscope maintains one set of common parameters alongside multiple sets of distinct, learnable masks for different agents, dictating the sharing of parameters. It promotes diversity among policy networks by encouraging discrepancy among these masks, without sacrificing the efficiencies of parameter sharing. This design allows Kaleidoscope to dynamically balance high sample efficiency with a broad policy representational capacity, effectively bridging the gap between full parameter sharing and non-parameter sharing across various environments. We further extend Kaleidoscope to critic ensembles in the context of actor-critic algorithms, which could help improve value estimations.Our empirical evaluations across extensive environments, including multi-agent particle environment, multi-agent MuJoCo and StarCraft multi-agent challenge v2, demonstrate the superior performance of Kaleidoscope compared with existing parameter sharing approaches, showcasing its potential for performance enhancement in MARL. The code is publicly available at \url{https://github.com/LXXXXR/Kaleidoscope}.
Authors: Xinyu Liao, Aoyang Qin, Jacob Seidman, Junqi Wang, Wei Wang, Paris Perdikaris
Abstract: Most existing generative models are limited to learning a single probability distribution from the training data and cannot generalize to novel distributions for unseen data. An architecture that can generate samples from both trained datasets and unseen probability distributions would mark a significant breakthrough. Recently, score-based generative models have gained considerable attention for their comprehensive mode coverage and high-quality image synthesis, as they effectively learn an operator that maps a probability distribution to its corresponding score function. In this work, we introduce the $\emph{Score Neural Operator}$, which learns the mapping from multiple probability distributions to their score functions within a unified framework. We employ latent space techniques to facilitate the training of score matching, which tends to over-fit in the original image pixel space, thereby enhancing sample generation quality. Our trained Score Neural Operator demonstrates the ability to predict score functions of probability measures beyond the training space and exhibits strong generalization performance in both 2-dimensional Gaussian Mixture Models and 1024-dimensional MNIST double-digit datasets. Importantly, our approach offers significant potential for few-shot learning applications, where a single image from a new distribution can be leveraged to generate multiple distinct images from that distribution.
Authors: Ruikai Yang, Mingzhen He, Zhengbao He, Youmei Qiu, Xiaolin Huang
Abstract: Machine unlearning (MU) is to make a well-trained model behave as if it had never been trained on specific data. In today's over-parameterized models, dominated by neural networks, a common approach is to manually relabel data and fine-tune the well-trained model. It can approximate the MU model in the output space, but the question remains whether it can achieve exact MU, i.e., in the parameter space. We answer this question by employing random feature techniques to construct an analytical framework. Under the premise of model optimization via stochastic gradient descent, we theoretically demonstrated that over-parameterized linear models can achieve exact MU through relabeling specific data. We also extend this work to real-world nonlinear networks and propose an alternating optimization algorithm that unifies the tasks of unlearning and relabeling. The algorithm's effectiveness, confirmed through numerical experiments, highlights its superior performance in unlearning across various scenarios compared to current state-of-the-art methods, particularly excelling over similar relabeling-based MU approaches.
Authors: Sehun Kim
Abstract: We propose a self-supervised learning method for 12-lead Electrocardiogram (ECG) analysis, named ECG Joint Embedding Predictive Architecture (ECG-JEPA). ECG-JEPA employs a masking strategy to learn semantic representations of ECG data. Unlike existing methods, ECG-JEPA predicts at the hidden representation level rather than reconstructing raw data. This approach offers several advantages in the ECG domain: (1) it avoids producing unnecessary details, such as noise, which is common in standard ECG; and (2) it addresses the limitations of na\"ive L2 loss between raw signals. Another key contribution is the introduction of a special masked attention tailored for 12-lead ECG data, Cross-Pattern Attention (CroPA). CroPA enables the model to effectively capture inter-patch relationships. Additionally, ECG-JEPA is highly scalable, allowing efficient training on large datasets. Our code is openly available https://github.com/sehunfromdaegu/ECG_JEPA.
Authors: Julien Zhou (Thoth, STATIFY), Pierre Gaillard (Thoth), Thibaud Rahier (STATIFY), Julyan Arbel (STATIFY)
Abstract: We address the online unconstrained submodular maximization problem (Online USM), in a setting with stochastic bandit feedback. In this framework, a decision-maker receives noisy rewards from a nonmonotone submodular function, taking values in a known bounded interval. This paper proposes Double-Greedy - Explore-then-Commit (DG-ETC), adapting the Double-Greedy approach from the offline and online full-information settings. DG-ETC satisfies a O(d log(dT)) problemdependent upper bound for the 1/2-approximate pseudo-regret, as well as a O(dT^{2/3}log(dT)^{1/3}) problem-free one at the same time, outperforming existing approaches. To that end, we introduce a notion of hardness for submodular functions, characterizing how difficult it is to maximize them with this type of strategy.
Authors: I-Chun Chen, Hsu-Shen Liu, Wei-Fang Sun, Chen-Hao Chao, Yen-Chang Hsu, Chun-Yi Lee
Abstract: Sparse Mixture-of-Experts (SMoE) models represent a significant breakthrough in large language model development. These models enable performance improvements without a proportional increase in inference costs. By selectively activating a small set of parameters during task execution, SMoEs enhance model capacity. However, their deployment remains challenging due to the substantial memory footprint required to accommodate the growing number of experts. This constraint renders them less feasible in environments with limited hardware resources. To address this challenge, we propose Hierarchical Clustering for Sparsely activated Mixture of Experts (HC-SMoE), a task-agnostic expert merging framework that reduces SMoE model parameters without retraining. Unlike previous methods, HC-SMoE employs hierarchical clustering based on expert outputs. This approach ensures that the merging process remains unaffected by routing decisions. The output-based clustering strategy captures functional similarities between experts, offering an adaptable solution for models with numerous experts. We validate our approach through extensive experiments on eight zero-shot language tasks and demonstrate its effectiveness in large-scale SMoE models such as Qwen and Mixtral. Our comprehensive results demonstrate that HC-SMoE consistently achieves strong performance, which highlights its potential for real-world deployment.
Authors: Jiazhen Chen, Sichao Fu, Zhibin Zhang, Zheng Ma, Mingbin Feng, Tony S. Wirjanto, Qinmu Peng
Abstract: Few-shot graph anomaly detection (GAD) has recently garnered increasing attention, which aims to discern anomalous patterns among abundant unlabeled test nodes under the guidance of a limited number of labeled training nodes. Existing few-shot GAD approaches typically adopt meta-training methods trained on richly labeled auxiliary networks to facilitate rapid adaptation to target networks that possess sparse labels. However, these proposed methods often assume that the auxiliary and target networks exist in the same data distributions-an assumption rarely holds in practical settings. This paper explores a more prevalent and complex scenario of cross-domain few-shot GAD, where the goal is to identify anomalies within sparsely labeled target graphs using auxiliary graphs from a related, yet distinct domain. The challenge here is nontrivial owing to inherent data distribution discrepancies between the source and target domains, compounded by the uncertainties of sparse labeling in the target domain. In this paper, we propose a simple and effective framework, termed CDFS-GAD, specifically designed to tackle the aforementioned challenges. CDFS-GAD first introduces a domain-adaptive graph contrastive learning module, which is aimed at enhancing cross-domain feature alignment. Then, a prompt tuning module is further designed to extract domain-specific features tailored to each domain. Moreover, a domain-adaptive hypersphere classification loss is proposed to enhance the discrimination between normal and anomalous instances under minimal supervision, utilizing domain-sensitive norms. Lastly, a self-training strategy is introduced to further refine the predicted scores, enhancing its reliability in few-shot settings. Extensive experiments on twelve real-world cross-domain data pairs demonstrate the effectiveness of the proposed CDFS-GAD framework in comparison to various existing GAD methods.
Authors: Juno Kim, Taiji Suzuki
Abstract: This work provides the first theoretical analysis of training transformers to solve complex problems by recursively generating intermediate states, analogous to fine-tuning for chain-of-thought (CoT) reasoning. We consider training a one-layer transformer to solve the fundamental $k$-parity problem, extending the work on RNNs by Wies et al. (2023). We establish three key results: (1) any finite-precision gradient-based algorithm, without intermediate supervision, requires substantial iterations to solve parity with finite samples. (2) In contrast, when intermediate parities are incorporated into the loss function, our model can learn parity in one gradient update when aided by \emph{teacher forcing}, where ground-truth labels of the reasoning chain are provided at each generation step. (3) Even without teacher forcing, where the model must generate CoT chains end-to-end, parity can be learned efficiently if augmented data is employed to internally verify the soundness of intermediate steps. These results rigorously show that task decomposition and stepwise reasoning naturally arise from optimizing transformers with CoT; moreover, self-consistency checking can improve reasoning ability, aligning with empirical studies of CoT.
Authors: Yubo Peng, Feibo Jiang, Li Dong, Kezhi Wang, Kun Yang
Abstract: Federated learning (FL) is a commonly distributed algorithm for mobile users (MUs) training artificial intelligence (AI) models, however, several challenges arise when applying FL to real-world scenarios, such as label scarcity, non-IID data, and unexplainability. As a result, we propose an explainable personalized FL framework, called XPFL. First, we introduce a generative AI (GAI) assisted personalized federated semi-supervised learning, called GFed. Particularly, in local training, we utilize a GAI model to learn from large unlabeled data and apply knowledge distillation-based semi-supervised learning to train the local FL model using the knowledge acquired from the GAI model. In global aggregation, we obtain the new local FL model by fusing the local and global FL models in specific proportions, allowing each local model to incorporate knowledge from others while preserving its personalized characteristics. Second, we propose an explainable AI mechanism for FL, named XFed. Specifically, in local training, we apply a decision tree to match the input and output of the local FL model. In global aggregation, we utilize t-distributed stochastic neighbor embedding (t-SNE) to visualize the local models before and after aggregation. Finally, simulation results validate the effectiveness of the proposed XPFL framework.
Authors: Jadon Fowler, Toby Dylan Hocking
Abstract: Receiver Operating Characteristic (ROC) curves are useful for evaluation in binary classification and changepoint detection, but difficult to use for learning since the Area Under the Curve (AUC) is piecewise constant (gradient zero almost everywhere). Recently the Area Under Min (AUM) of false positive and false negative rates has been proposed as a differentiable surrogate for AUC. In this paper we study the piecewise linear/constant nature of the AUM/AUC, and propose new efficient path-following algorithms for choosing the learning rate which is optimal for each step of gradient descent (line search), when optimizing a linear model. Remarkably, our proposed line search algorithm has the same log-linear asymptotic time complexity as gradient descent with constant step size, but it computes a complete representation of the AUM/AUC as a function of step size. In our empirical study of binary classification problems, we verify that our proposed algorithm is fast and exact; in changepoint detection problems we show that the proposed algorithm is just as accurate as grid search, but faster.
Authors: Rafael Pablos Sarabia, Joachim Nyborg, Morten Birk, Jeppe Liborius Sj{\o}rup, Anders Lillevang Vesterholt, Ira Assent
Abstract: Precipitation nowcasting is crucial across various industries and plays a significant role in mitigating and adapting to climate change. We introduce an efficient deep learning model for precipitation nowcasting, capable of predicting rainfall up to 8 hours in advance with greater accuracy than existing operational physics-based and extrapolation-based models. Our model leverages multi-source meteorological data and physics-based forecasts to deliver high-resolution predictions in both time and space. It captures complex spatio-temporal dynamics through temporal attention networks and is optimized using data quality maps and dynamic thresholds. Experiments demonstrate that our model outperforms state-of-the-art, and highlight its potential for fast reliable responses to evolving weather conditions.
Authors: Gleb Radchenko, Victoria Andrea Fill
Abstract: Recent advancements in edge computing have significantly enhanced the AI capabilities of Internet of Things (IoT) devices. However, these advancements introduce new challenges in knowledge exchange and resource management, particularly addressing the spatiotemporal data locality in edge computing environments. This study examines algorithms and methods for deploying distributed machine learning within autonomous, network-capable, AI-enabled edge devices. We focus on determining confidence levels in learning outcomes considering the spatial variability of data encountered by independent agents. Using collaborative mapping as a case study, we explore the application of the Distributed Neural Network Optimization (DiNNO) algorithm extended with Bayesian neural networks (BNNs) for uncertainty estimation. We implement a 3D environment simulation using the Webots platform to simulate collaborative mapping tasks, decouple the DiNNO algorithm into independent processes for asynchronous network communication in distributed learning, and integrate distributed uncertainty estimation using BNNs. Our experiments demonstrate that BNNs can effectively support uncertainty estimation in a distributed learning context, with precise tuning of learning hyperparameters crucial for effective uncertainty assessment. Notably, applying Kullback-Leibler divergence for parameter regularization resulted in a 12-30% reduction in validation loss during distributed BNN training compared to other regularization strategies.
Authors: Toby Dylan Hocking
Abstract: Binary segmentation is the classic greedy algorithm which recursively splits a sequential data set by optimizing some loss or likelihood function. Binary segmentation is widely used for changepoint detection in data sets measured over space or time, and as a sub-routine for decision tree learning. In theory it should be extremely fast for $N$ data and $K$ splits, $O(N K)$ in the worst case, and $O(N \log K)$ in the best case. In this paper we describe new methods for analyzing the time and space complexity of binary segmentation for a given finite $N$, $K$, and minimum segment length parameter. First, we describe algorithms that can be used to compute the best and worst case number of splits the algorithm must consider. Second, we describe synthetic data that achieve the best and worst case and which can be used to test for correct implementation of the algorithm. Finally, we provide an empirical analysis of real data which suggests that binary segmentation is often close to optimal speed in practice.
Authors: Bryce Ferenczi, Rhys Newbury, Michael Burke, Tom Drummond
Abstract: Creation and storage of datasets are often overlooked input costs in machine learning, as many datasets are simple image label pairs or plain text. However, datasets with more complex structures, such as those from the real time strategy game StarCraft II, require more deliberate thought and strategy to reduce cost of ownership. We introduce a serialization framework for StarCraft II that reduces the cost of dataset creation and storage, as well as improving usage ergonomics. We benchmark against the most comparable existing dataset from \textit{AlphaStar-Unplugged} and highlight the benefit of our framework in terms of both the cost of creation and storage. We use our dataset to train deep learning models that exceed the performance of comparable models trained on other datasets. The dataset conversion and usage framework introduced is open source and can be used as a framework for datasets with similar characteristics such as digital twin simulations. Pre-converted StarCraft II tournament data is also available online.
Authors: Peiran Wang, Haohan Wang
Abstract: In this paper, we introduce DistDD, a novel approach within the federated learning framework that reduces the need for repetitive communication by distilling data directly on clients' devices. Unlike traditional federated learning that requires iterative model updates across nodes, DistDD facilitates a one-time distillation process that extracts a global distilled dataset, maintaining the privacy standards of federated learning while significantly cutting down communication costs. By leveraging the DistDD's distilled dataset, the developers of the FL can achieve just-in-time parameter tuning and neural architecture search over FL without repeating the whole FL process multiple times. We provide a detailed convergence proof of the DistDD algorithm, reinforcing its mathematical stability and reliability for practical applications. Our experiments demonstrate the effectiveness and robustness of DistDD, particularly in non-i.i.d. and mislabeled data scenarios, showcasing its potential to handle complex real-world data challenges distinctively from conventional federated learning methods. We also evaluate DistDD's application in the use case and prove its effectiveness and communication-savings in the NAS use case.
Authors: Yanfeng Jiang, Zelan Yang, Bohua Chen, Shen Li, Yong Li, Tao Li
Abstract: Large language models achieve exceptional performance on various downstream tasks through supervised fine-tuning. However, the diversity of downstream tasks and practical requirements makes deploying multiple full-parameter fine-tuned models challenging. Current methods that compress the delta weight struggle to achieve ultra-high compression, failing to minimize the deployment overhead. To address the above issue, we propose a novel distribution-driven delta compression framework DeltaDQ, which utilizes Group-wise Dropout and Separate Quantization to achieve ultra-high compression for the delta weight. We have observed that the matrix-computed intermediate results for the delta weight exhibit extremely small variance and min-max range characteristics, referred to as Balanced Intermediate Results. Exploiting this phenomenon, we introduce Group-wise Dropout to perform dropout on the delta weight using an optimal group size. Furthermore, using Separate Quantization, sparse weights are quantized and decomposed to achieve a lower bit. Experimental results show that DeltaDQ achieves 16x compression with improved accuracy compared to baselines for WizardMath and WizardCoder models across different parameter scales. Moreover, DeltaDQ demonstrates the ability for ultra-high compression ratio, achieving 128x compression for the WizardMath-7B model and 512x compression for the WizardMath-70B model.
Authors: Bryce Ferenczi, Michael Burke, Tom Drummond
Abstract: Various works have aimed at combining the inference efficiency of recurrent models and training parallelism of multi-head attention for sequence modeling. However, most of these works focus on tasks with fixed-dimension observation spaces, such as individual tokens in language modeling or pixels in image completion. To handle an observation space of varying size, we propose a novel algorithm that alternates between cross-attention between a 2D latent state and observation, and a discounted cumulative sum over the sequence dimension to efficiently accumulate historical information. We find this resampling cycle is critical for performance. To evaluate efficient sequence modeling in this domain, we introduce two multi-agent intention tasks: simulated agents chasing bouncing particles and micromanagement analysis in professional StarCraft II games. Our algorithm achieves comparable accuracy with a lower parameter count, faster training and inference compared to existing methods.
Authors: Hanieh Shojaei, Qianqian Zou, Max Mehltretter
Abstract: Safe navigation in new environments requires autonomous vehicles and robots to accurately interpret their surroundings, relying on LiDAR scene segmentation, out-of-distribution (OOD) obstacle detection, and uncertainty computation. We propose a method to distinguish in-distribution (ID) from OOD samples and quantify both epistemic and aleatoric uncertainties using the feature space of a single deterministic model. After training a semantic segmentation network, a Gaussian Mixture Model (GMM) is fitted to its feature space. OOD samples are detected by checking if their squared Mahalanobis distances to each Gaussian component conform to a chi-squared distribution, eliminating the need for an additional OOD training set. Given that the estimated mean and covariance matrix of a multivariate Gaussian distribution follow Gaussian and Inverse-Wishart distributions, multiple GMMs are generated by sampling from these distributions to assess epistemic uncertainty through classification variability. Aleatoric uncertainty is derived from the entropy of responsibility values within Gaussian components. Comparing our method with deep ensembles and logit-sampling for uncertainty computation demonstrates its superior performance in real-world applications for quantifying epistemic and aleatoric uncertainty, as well as detecting OOD samples. While deep ensembles miss some highly uncertain samples, our method successfully detects them and assigns high epistemic uncertainty.
Authors: Satoshi Hayakawa, Yuhta Takida, Masaaki Imaizumi, Hiromi Wakaki, Yuki Mitsufuji
Abstract: Diffusion models have demonstrated exceptional performances in various fields of generative modeling. While they often outperform competitors including VAEs and GANs in sample quality and diversity, they suffer from slow sampling speed due to their iterative nature. Recently, distillation techniques and consistency models are mitigating this issue in continuous domains, but discrete diffusion models have some specific challenges towards faster generation. Most notably, in the current literature, correlations between different dimensions (pixels, locations) are ignored, both by its modeling and loss functions, due to computational limitations. In this paper, we propose "mixture" models in discrete diffusion that are capable of treating dimensional correlations while remaining scalable, and we provide a set of loss functions for distilling the iterations of existing models. Two primary theoretical insights underpin our approach: first, that dimensionally independent models can well approximate the data distribution if they are allowed to conduct many sampling steps, and second, that our loss functions enables mixture models to distill such many-step conventional models into just a few steps by learning the dimensional correlations. We empirically demonstrate that our proposed method for discrete diffusions work in practice, by distilling a continuous-time discrete diffusion model pretrained on the CIFAR-10 dataset.
Authors: Petrus Mikkola, Luigi Acerbi, Arto Klami
Abstract: Eliciting a high-dimensional probability distribution from an expert via noisy judgments is notoriously challenging, yet useful for many applications, such as prior elicitation and reward modeling. We introduce a method for eliciting the expert's belief density as a normalizing flow based solely on preferential questions such as comparing or ranking alternatives. This allows eliciting in principle arbitrarily flexible densities, but flow estimation is susceptible to the challenge of collapsing or diverging probability mass that makes it difficult in practice. We tackle this problem by introducing a novel functional prior for the flow, motivated by a decision-theoretic argument, and show empirically that the belief density can be inferred as the function-space maximum a posteriori estimate. We demonstrate our method by eliciting multivariate belief densities of simulated experts, including the prior belief of a general-purpose large language model over a real-world dataset.
Authors: H. Yi, H. Ren, C. Hu, Y. Li, J. Deng, X. Xie
Abstract: Federated Learning (FL) has become a cornerstone of privacy protection, shifting the paradigm towards localizing sensitive data while only sending model gradients to a central server. This strategy is designed to reinforce privacy protections and minimize the vulnerabilities inherent in centralized data storage systems. Despite its innovative approach, recent empirical studies have highlighted potential weaknesses in FL, notably regarding the exchange of gradients. In response, this study introduces a novel, efficacious method aimed at safeguarding against gradient leakage, namely, ``AdaDefense". Following the idea that model convergence can be achieved by using different types of optimization methods, we suggest using a local stand-in rather than the actual local gradient for global gradient aggregation on the central server. This proposed approach not only effectively prevents gradient leakage, but also ensures that the overall performance of the model remains largely unaffected. Delving into the theoretical dimensions, we explore how gradients may inadvertently leak private information and present a theoretical framework supporting the efficacy of our proposed method. Extensive empirical tests, supported by popular benchmark experiments, validate that our approach maintains model integrity and is robust against gradient leakage, marking an important step in our pursuit of safe and efficient FL.
Authors: Thomas Rupf, Marco Bagatella, Nico G\"urtler, Jonas Frey, Georg Martius
Abstract: Zero-shot imitation learning algorithms hold the promise of reproducing unseen behavior from as little as a single demonstration at test time. Existing practical approaches view the expert demonstration as a sequence of goals, enabling imitation with a high-level goal selector, and a low-level goal-conditioned policy. However, this framework can suffer from myopic behavior: the agent's immediate actions towards achieving individual goals may undermine long-term objectives. We introduce a novel method that mitigates this issue by directly optimizing the occupancy matching objective that is intrinsic to imitation learning. We propose to lift a goal-conditioned value function to a distance between occupancies, which are in turn approximated via a learned world model. The resulting method can learn from offline, suboptimal data, and is capable of non-myopic, zero-shot imitation, as we demonstrate in complex, continuous benchmarks.
Authors: Zhifei Li, Gerrit Gro{\ss}mann, Verena Wolf
Abstract: In recent years, a wide variety of graph neural network (GNN) architectures have emerged, each with its own strengths, weaknesses, and complexities. Various techniques, including rewiring, lifting, and node annotation with centrality values, have been employed as pre-processing steps to enhance GNN performance. However, there are no universally accepted best practices, and the impact of architecture and pre-processing on performance often remains opaque. This study systematically explores the impact of various graph transformations as pre-processing steps on the performance of common GNN architectures across standard datasets. The models are evaluated based on their ability to distinguish non-isomorphic graphs, referred to as expressivity. Our findings reveal that certain transformations, particularly those augmenting node features with centrality measures, consistently improve expressivity. However, these gains come with trade-offs, as methods like graph encoding, while enhancing expressivity, introduce numerical inaccuracies widely-used python packages. Additionally, we observe that these pre-processing techniques are limited when addressing complex tasks involving 3-WL and 4-WL indistinguishable graphs.
Authors: Konstantin Burlachenko, Peter Richt\'arik
Abstract: Federated Learning (FL) is an emerging paradigm that enables intelligent agents to collaboratively train Machine Learning (ML) models in a distributed manner, eliminating the need for sharing their local data. The recent work (arXiv:2106.02969) introduces a family of Federated Newton Learn (FedNL) algorithms, marking a significant step towards applying second-order methods to FL and large-scale optimization. However, the reference FedNL prototype exhibits three serious practical drawbacks: (i) It requires 4.8 hours to launch a single experiment in a sever-grade workstation; (ii) The prototype only simulates multi-node setting; (iii) Prototype integration into resource-constrained applications is challenging. To bridge the gap between theory and practice, we present a self-contained implementation of FedNL, FedNL-LS, FedNL-PP for single-node and multi-node settings. Our work resolves the aforementioned issues and reduces the wall clock time by x1000. With this FedNL outperforms alternatives for training logistic regression in a single-node -- CVXPY (arXiv:1603.00943), and in a multi-node -- Apache Spark (arXiv:1505.06807), Ray/Scikit-Learn (arXiv:1712.05889). Finally, we propose two practical-orientated compressors for FedNL - adaptive TopLEK and cache-aware RandSeqK, which fulfill the theory of FedNL.
Authors: Stefan Feuerriegel, Dennis Frauen, Valentyn Melnychuk, Jonas Schweisthal, Konstantin Hess, Alicia Curth, Stefan Bauer, Niki Kilbertus, Isaac S. Kohane, Mihaela van der Schaar
Abstract: Causal machine learning (ML) offers flexible, data-driven methods for predicting treatment outcomes including efficacy and toxicity, thereby supporting the assessment and safety of drugs. A key benefit of causal ML is that it allows for estimating individualized treatment effects, so that clinical decision-making can be personalized to individual patient profiles. Causal ML can be used in combination with both clinical trial data and real-world data, such as clinical registries and electronic health records, but caution is needed to avoid biased or incorrect predictions. In this Perspective, we discuss the benefits of causal ML (relative to traditional statistical or ML approaches) and outline the key components and steps. Finally, we provide recommendations for the reliable use of causal ML and effective translation into the clinic.
Authors: Rohan Alur, Loren Laine, Darrick K. Li, Dennis Shung, Manish Raghavan, Devavrat Shah
Abstract: We introduce a novel framework for human-AI collaboration in prediction and decision tasks. Our approach leverages human judgment to distinguish inputs which are algorithmically indistinguishable, or "look the same" to any feasible predictive algorithm. We argue that this framing clarifies the problem of human-AI collaboration in prediction and decision tasks, as experts often form judgments by drawing on information which is not encoded in an algorithm's training data. Algorithmic indistinguishability yields a natural test for assessing whether experts incorporate this kind of "side information", and further provides a simple but principled method for selectively incorporating human feedback into algorithmic predictions. We show that this method provably improves the performance of any feasible algorithmic predictor and precisely quantify this improvement. We demonstrate the utility of our framework in a case study of emergency room triage decisions, where we find that although algorithmic risk scores are highly competitive with physicians, there is strong evidence that physician judgments provide signal which could not be replicated by any predictive algorithm. This insight yields a range of natural decision rules which leverage the complementary strengths of human experts and predictive algorithms.
Authors: Mathieu Chevalley, Arash Mehrjou, Patrick Schwab
Abstract: In the algorithm Intersort, Chevalley et al. (2024) proposed a score-based method to discover the causal order of variables in a Directed Acyclic Graph (DAG) model, leveraging interventional data to outperform existing methods. However, as a score-based method over the permutahedron, Intersort is computationally expensive and non-differentiable, limiting its ability to be utilised in problems involving large-scale datasets, such as those in genomics and climate models, or to be integrated into end-to-end gradient-based learning frameworks. We address this limitation by reformulating Intersort using differentiable sorting and ranking techniques. Our approach enables scalable and differentiable optimization of causal orderings, allowing the continuous score function to be incorporated as a regularizer in downstream tasks. Empirical results demonstrate that causal discovery algorithms benefit significantly from regularizing on the causal order, underscoring the effectiveness of our method. Our work opens the door to efficiently incorporating regularization for causal order into the training of differentiable models and thereby addresses a long-standing limitation of purely associational supervised learning.
Authors: Reza Abbasi, Sernam Lim
Abstract: The rapid growth in machine learning models, especially in natural language processing and computer vision, has led to challenges when running these models on hardware with limited resources. This paper introduces Superpipeline, a new framework designed to optimize the execution of large AI models on constrained hardware during both training and inference. Our approach involves dynamically managing model execution by dividing models into individual layers and efficiently transferring these layers between GPU and CPU memory. Superpipeline reduces GPU memory usage by up to 60% in our experiments while maintaining model accuracy and acceptable processing speeds. This allows models that would otherwise exceed available GPU memory to run effectively. Unlike existing solutions that focus mainly on inference or specific model types, Superpipeline can be applied to large language models (LLMs), vision-language models (VLMs), and vision-based models. We tested Superpipeline's performance across various models and hardware setups. The method includes two key parameters that allow fine-tuning the balance between GPU memory use and processing speed. Importantly, Superpipeline does not require retraining or changing model parameters, ensuring that the original model's output remains unchanged. Superpipeline's simplicity and flexibility make it useful for researchers and professionals working with advanced AI models on limited hardware. It enables the use of larger models or bigger batch sizes on existing hardware, potentially speeding up innovation across many machine learning applications. This work marks an important step toward making advanced AI models more accessible and optimizing their deployment in resource-limited environments. The code for Superpipeline is available at https://github.com/abbasiReza/super-pipeline.
Authors: Zhongyi Yu, Zhenghao Wu, Shuhan Zhong, Weifeng Su, S. -H. Gary Chan, Chul-Ho Lee, Weipeng Zhuo
Abstract: Missing values are a common problem that poses significant challenges to data analysis and machine learning. This problem necessitates the development of an effective imputation method to fill in the missing values accurately, thereby enhancing the overall quality and utility of the datasets. Existing imputation methods, however, fall short of explicitly considering the `missingness' information in the data during the embedding initialization stage and modeling the entangled feature and sample correlations during the learning process, thus leading to inferior performance. We propose M$^3$-Impute, which aims to explicitly leverage the missingness information and such correlations with novel masking schemes. M$^3$-Impute first models the data as a bipartite graph and uses a graph neural network to learn node embeddings, where the refined embedding initialization process directly incorporates the missingness information. They are then optimized through M$^3$-Impute's novel feature correlation unit (FRU) and sample correlation unit (SRU) that effectively captures feature and sample correlations for imputation. Experiment results on 25 benchmark datasets under three different missingness settings show the effectiveness of M$^3$-Impute by achieving 20 best and 4 second-best MAE scores on average.
Authors: Felix Teufel, Carsten Stahlhut, Jesper Ferkinghoff-Borg
Abstract: Bayesian optimization (BO) is an attractive machine learning framework for performing sample-efficient global optimization of black-box functions. The optimization process is guided by an acquisition function that selects points to acquire in each round of BO. In batched BO, when multiple points are acquired in parallel, commonly used acquisition functions are often high-dimensional and intractable, leading to the use of sampling-based alternatives. We propose a statistical physics inspired acquisition function for BO with Gaussian processes that can natively handle batches. Batched Energy-Entropy acquisition for BO (BEEBO) enables tight control of the explore-exploit trade-off of the optimization process and generalizes to heteroskedastic black-box problems. We demonstrate the applicability of BEEBO on a range of problems, showing competitive performance to existing methods.
Authors: Chris Cummins, Volker Seeker, Jordi Armengol-Estap\'e, Aram H. Markosyan, Gabriel Synnaeve, Hugh Leather
Abstract: Tools for rewriting, refactoring and optimizing code should be fast and correct. Large language models (LLMs), by their nature, possess neither of these qualities. Yet, there remains tremendous opportunity in using LLMs to improve code. We explore the use of LLMs not to transform code, but to code transforms. We propose a chain-of-thought approach to synthesizing code transformations from a small number of input/output code examples that incorporates execution and feedback. Unlike the direct rewrite approach, LLM-generated transformations are easy to inspect, debug, and validate. The logic of the rewrite is explicitly coded and easy to adapt. The compute required to run code transformations is minute compared to that of LLM rewriting. We test our approach on 16 Python code transformations and find that LLM- generated transforms are perfectly precise for 7 of them and less imprecise than direct LLM rewriting on the others. We hope to encourage further research to improving the precision of LLM code rewriting.
Authors: Thomas Schwarz, Cecilia Casolo, Niki Kilbertus
Abstract: In personalized medicine, the ability to predict and optimize treatment outcomes across various time frames is essential. Additionally, the ability to select cost-effective treatments within specific budget constraints is critical. Despite recent advancements in estimating counterfactual trajectories, a direct link to optimal treatment selection based on these estimates is missing. This paper introduces a novel method integrating counterfactual estimation techniques and uncertainty quantification to recommend personalized treatment plans adhering to predefined cost constraints. Our approach is distinctive in its handling of continuous treatment variables and its incorporation of uncertainty quantification to improve prediction reliability. We validate our method using two simulated datasets, one focused on the cardiovascular system and the other on COVID-19. Our findings indicate that our method has robust performance across different counterfactual estimation baselines, showing that introducing uncertainty quantification in these settings helps the current baselines in finding more reliable and accurate treatment selection. The robustness of our method across various settings highlights its potential for broad applicability in personalized healthcare solutions.
Authors: Malte Mosbach, Jan Niklas Ewertz, Angel Villar-Corrales, Sven Behnke
Abstract: Learning a latent dynamics model provides a task-agnostic representation of an agent's understanding of its environment. Leveraging this knowledge for model-based reinforcement learning holds the potential to improve sample efficiency over model-free methods by learning inside imagined rollouts. Furthermore, because the latent space serves as input to behavior models, the informative representations learned by the world model facilitate efficient learning of desired skills. Most existing methods rely on holistic representations of the environment's state. In contrast, humans reason about objects and their interactions, forecasting how actions will affect specific parts of their surroundings. Inspired by this, we propose Slot-Attention for Object-centric Latent Dynamics (SOLD), a novel algorithm that learns object-centric dynamics models in an unsupervised manner from pixel inputs. We demonstrate that the structured latent space not only improves model interpretability but also provides a valuable input space for behavior models to reason over. Our results show that SOLD outperforms DreamerV3, a state-of-the-art model-based RL algorithm, across a range of benchmark robotic environments that evaluate for both relational reasoning and low-level manipulation capabilities. Videos are available at https://slot-latent-dynamics.github.io/.
Authors: Aghyad Deeb, Fabien Roger
Abstract: Large Language Models' knowledge of how to perform cyber-security attacks, create bioweapons, and manipulate humans poses risks of misuse. Previous work has proposed methods to unlearn this knowledge. Historically, it has been unclear whether unlearning techniques are removing information from the model weights or just making it harder to access. To disentangle these two objectives, we propose an adversarial evaluation method to test for the removal of information from model weights: we give an attacker access to some facts that were supposed to be removed, and using those, the attacker tries to recover other facts from the same distribution that cannot be guessed from the accessible facts. We show that using fine-tuning on the accessible facts can recover 88% of the pre-unlearning accuracy when applied to current unlearning methods, revealing the limitations of these methods in removing information from the model weights.
Authors: Zhuoran Li, Xu Sun, Wanyu Lin, Jiannong Cao
Abstract: Explainable molecular property prediction is essential for various scientific fields, such as drug discovery and material science. Despite delivering intrinsic explainability, linear models struggle with capturing complex, non-linear patterns. Large language models (LLMs), on the other hand, yield accurate predictions through powerful inference capabilities yet fail to provide chemically meaningful explanations for their predictions. This work proposes a novel framework, called MoleX, which leverages LLM knowledge to build a simple yet powerful linear model for accurate molecular property prediction with faithful explanations. The core of MoleX is to model complicated molecular structure-property relationships using a simple linear model, augmented by LLM knowledge and a crafted calibration strategy. Specifically, to extract the maximum amount of task-relevant knowledge from LLM embeddings, we employ information bottleneck-inspired fine-tuning and sparsity-inducing dimensionality reduction. These informative embeddings are then used to fit a linear model for explainable inference. Moreover, we introduce residual calibration to address prediction errors stemming from linear models' insufficient expressiveness of complex LLM embeddings, thus recovering the LLM's predictive power and boosting overall accuracy. Theoretically, we provide a mathematical foundation to justify MoleX's explainability. Extensive experiments demonstrate that MoleX outperforms existing methods in molecular property prediction, establishing a new milestone in predictive performance, explainability, and efficiency. In particular, MoleX enables CPU inference and accelerates large-scale dataset processing, achieving comparable performance 300x faster with 100,000 fewer parameters than LLMs. Additionally, the calibration improves model performance by up to 12.7% without compromising explainability.
Authors: Monika Gierszewska, Tomasz Berezowski
Abstract: The flooding extent area in a river valley is related to river gauge observations. The higher the water elevation, the larger the flooding area. Due to synthetic aperture radar\textquoteright s (SAR) capabilities to penetrate through clouds, radar images have been commonly used to estimate flooding extent area with various methods, from simple thresholding to deep learning models. In this study, we propose a physics-guided neural network for flooding area detection. Our approach takes as input data the Sentinel 1 time-series images and the water elevations in the river assigned to each image. We apply the Pearson correlation coefficient between the predicted sum of water extent areas and the local water level observations of river water elevations as the loss function. The effectiveness of our method is evaluated in five different study areas by comparing the predicted water maps with reference water maps obtained from digital terrain models and optical satellite images. The highest Intersection over Union (IoU) score achieved by our models was 0.89 for the water class and 0.96 for the non-water class. Additionally, we compared the results with other unsupervised methods. The proposed neural network provided a higher IoU than the other methods, especially for SAR images registered during low water elevation in the river.
Authors: Noam Razin, Sadhika Malladi, Adithya Bhaskar, Danqi Chen, Sanjeev Arora, Boris Hanin
Abstract: Direct Preference Optimization (DPO) and its variants are increasingly used for aligning language models with human preferences. Although these methods are designed to teach a model to generate preferred responses more frequently relative to dispreferred responses, prior work has observed that the likelihood of preferred responses often decreases during training. The current work sheds light on the causes and implications of this counter-intuitive phenomenon, which we term likelihood displacement. We demonstrate that likelihood displacement can be catastrophic, shifting probability mass from preferred responses to responses with an opposite meaning. As a simple example, training a model to prefer $\texttt{No}$ over $\texttt{Never}$ can sharply increase the probability of $\texttt{Yes}$. Moreover, when aligning the model to refuse unsafe prompts, we show that such displacement can unintentionally lead to unalignment, by shifting probability mass from preferred refusal responses to harmful responses (e.g., reducing the refusal rate of Llama-3-8B-Instruct from 74.4% to 33.4%). We theoretically characterize that likelihood displacement is driven by preferences that induce similar embeddings, as measured by a centered hidden embedding similarity (CHES) score. Empirically, the CHES score enables identifying which training samples contribute most to likelihood displacement in a given dataset. Filtering out these samples effectively mitigated unintentional unalignment in our experiments. More broadly, our results highlight the importance of curating data with sufficiently distinct preferences, for which we believe the CHES score may prove valuable.
Authors: Zijiang Yan, Hao Zhou, Hina Tabassum, Xue Liu
Abstract: Large language models (LLMs) have received considerable interest recently due to their outstanding reasoning and comprehension capabilities. This work explores applying LLMs to vehicular networks, aiming to jointly optimize vehicle-to-infrastructure (V2I) communications and autonomous driving (AD) policies. We deploy LLMs for AD decision-making to maximize traffic flow and avoid collisions for road safety, and a double deep Q-learning algorithm (DDQN) is used for V2I optimization to maximize the received data rate and reduce frequent handovers. In particular, for LLM-enabled AD, we employ the Euclidean distance to identify previously explored AD experiences, and then LLMs can learn from past good and bad decisions for further improvement. Then, LLM-based AD decisions will become part of states in V2I problems, and DDQN will optimize the V2I decisions accordingly. After that, the AD and V2I decisions are iteratively optimized until convergence. Such an iterative optimization approach can better explore the interactions between LLMs and conventional reinforcement learning techniques, revealing the potential of using LLMs for network optimization and management. Finally, the simulations demonstrate that our proposed hybrid LLM-DDQN approach outperforms the conventional DDQN algorithm, showing faster convergence and higher average rewards.
Authors: Grzegorz G{\l}uch, Berkant Turan, Sai Ganesh Nagarajan, Sebastian Pokutta
Abstract: We formalize and extend existing definitions of backdoor-based watermarks and adversarial defenses as interactive protocols between two players. The existence of these schemes is inherently tied to the learning tasks for which they are designed. Our main result shows that for almost every discriminative learning task, at least one of the two -- a watermark or an adversarial defense -- exists. The term "almost every" indicates that we also identify a third, counterintuitive but necessary option, i.e., a scheme we call a transferable attack. By transferable attack, we refer to an efficient algorithm computing queries that look indistinguishable from the data distribution and fool all efficient defenders. To this end, we prove the necessity of a transferable attack via a construction that uses a cryptographic tool called homomorphic encryption. Furthermore, we show that any task that satisfies our notion of a transferable attack implies a cryptographic primitive, thus requiring the underlying task to be computationally complex. These two facts imply an "equivalence" between the existence of transferable attacks and cryptography. Finally, we show that the class of tasks of bounded VC-dimension has an adversarial defense, and a subclass of them has a watermark.
Authors: Lilin Fan, Xuqing Chai, Zhixiong Tian, Yihang Qiao, Zhen Wang, Yifan Zhang
Abstract: Analysis of the genome sequence of Perccottus glenii, the only fish known to possess freeze tolerance, holds significant importance for understanding how organisms adapt to extreme environments, Traditional biological analysis methods are time-consuming and have limited accuracy, To address these issues, we will employ machine learning techniques to analyze the gene sequences of Perccottus glenii, with Neodontobutis hainanens as a comparative group, Firstly, we have proposed five gene sequence vectorization methods and a method for handling ultra-long gene sequences, We conducted a comparative study on the three vectorization methods: ordinal encoding, One-Hot encoding, and K-mer encoding, to identify the optimal encoding method, Secondly, we constructed four classification models: Random Forest, LightGBM, XGBoost, and Decision Tree, The dataset used by these classification models was extracted from the National Center for Biotechnology Information database, and we vectorized the sequence matrices using the optimal encoding method, K-mer, The Random Forest model, which is the optimal model, achieved a classification accuracy of up to 99, 98 , Lastly, we utilized SHAP values to conduct an interpretable analysis of the optimal classification model, Through ten-fold cross-validation and the AUC metric, we identified the top 10 features that contribute the most to the model's classification accuracy, This demonstrates that machine learning methods can effectively replace traditional manual analysis in identifying genes associated with the freeze tolerance phenotype in Perccottus glenii.
Authors: Navdeep Kumar, Priyank Agrawal, Giorgia Ramponi, Kfir Yehuda Levy, Shie Mannor
Abstract: In this paper, we establish the global convergence of the actor-critic algorithm with a significantly improved sample complexity of $O(\epsilon^{-3})$, advancing beyond the existing local convergence results. Previous works provide local convergence guarantees with a sample complexity of $O(\epsilon^{-2})$ for bounding the squared gradient of the return, which translates to a global sample complexity of $O(\epsilon^{-4})$ using the gradient domination lemma. In contrast to traditional methods that employ decreasing step sizes for both the actor and critic, we demonstrate that a constant step size for the critic is sufficient to ensure convergence in expectation. This key insight reveals that using a decreasing step size for the actor alone is sufficient to handle the noise for both the actor and critic. Our findings provide theoretical support for the practical success of many algorithms that rely on constant step sizes.
Authors: Daniel Balcells, Benjamin Lerner, Michael Oesterle, Ediz Ucar, Stefan Heimersheim
Abstract: Sparse Autoencoders for transformer-based language models are typically defined independently per layer. In this work we analyze statistical relationships between features in adjacent layers to understand how features evolve through a forward pass. We provide a graph visualization interface for features and their most similar next-layer neighbors, and build communities of related features across layers. We find that a considerable amount of features are passed through from a previous layer, some features can be expressed as quasi-boolean combinations of previous features, and some features become more specialized in later layers.
Authors: Claas A Voelcker, Marcel Hussing, Marcel Hussing
Abstract: Empirical, benchmark-driven testing is a fundamental paradigm in the current RL community. While using off-the-shelf benchmarks in reinforcement learning (RL) research is a common practice, this choice is rarely discussed. Benchmark choices are often done based on intuitive ideas like "legged robots" or "visual observations". In this paper, we argue that benchmarking in RL needs to be treated as a scientific discipline itself. To illustrate our point, we present a case study on different variants of the Hopper environment to show that the selection of standard benchmarking suites can drastically change how we judge performance of algorithms. The field does not have a cohesive notion of what the different Hopper environments are representative - they do not even seem to be representative of each other. Our experimental results suggests a larger issue in the deep RL literature: benchmark choices are neither commonly justified, nor does there exist a language that could be used to justify the selection of certain environments. This paper concludes with a discussion of the requirements for proper discussion and evaluations of benchmarks and recommends steps to start a dialogue towards this goal.
Authors: Isha Gupta, Hidde Lycklama, Emanuel Opel, Evan Rose, Anwar Hithnawi
Abstract: As machine learning models become increasingly complex, concerns about their robustness and trustworthiness have become more pressing. A critical vulnerability of these models is data poisoning attacks, where adversaries deliberately alter training data to degrade model performance. One particularly stealthy form of these attacks is subpopulation poisoning, which targets distinct subgroups within a dataset while leaving overall performance largely intact. The ability of these attacks to generalize within subpopulations poses a significant risk in real-world settings, as they can be exploited to harm marginalized or underrepresented groups within the dataset. In this work, we investigate how model complexity influences susceptibility to subpopulation poisoning attacks. We introduce a theoretical framework that explains how overparameterized models, due to their large capacity, can inadvertently memorize and misclassify targeted subpopulations. To validate our theory, we conduct extensive experiments on large-scale image and text datasets using popular model architectures. Our results show a clear trend: models with more parameters are significantly more vulnerable to subpopulation poisoning. Moreover, we find that attacks on smaller, human-interpretable subgroups often go undetected by these models. These results highlight the need to develop defenses that specifically address subpopulation vulnerabilities.
Authors: Yuanyi Wang, Haifeng Sun, Chengsen Wang, Mengde Zhu, Jingyu Wang, Wei Tang, Qi Qi, Zirui Zhuang, Jianxin Liao
Abstract: Anomaly detection in multivariate time series (MTS) is crucial for various applications in data mining and industry. Current industrial methods typically approach anomaly detection as an unsupervised learning task, aiming to identify deviations by estimating the normal distribution in noisy, label-free datasets. These methods increasingly incorporate interdependencies between channels through graph structures to enhance accuracy. However, the role of interdependencies is more critical than previously understood, as shifts in interdependencies between MTS channels from normal to anomalous data are significant. This observation suggests that \textit{anomalies could be detected by changes in these interdependency graph series}. To capitalize on this insight, we introduce MADGA (MTS Anomaly Detection via Graph Alignment), which redefines anomaly detection as a graph alignment (GA) problem that explicitly utilizes interdependencies for anomaly detection. MADGA dynamically transforms subsequences into graphs to capture the evolving interdependencies, and Graph alignment is performed between these graphs, optimizing an alignment plan that minimizes cost, effectively minimizing the distance for normal data and maximizing it for anomalous data. Uniquely, our GA approach involves explicit alignment of both nodes and edges, employing Wasserstein distance for nodes and Gromov-Wasserstein distance for edges. To our knowledge, this is the first application of GA to MTS anomaly detection that explicitly leverages interdependency for this purpose. Extensive experiments on diverse real-world datasets validate the effectiveness of MADGA, demonstrating its capability to detect anomalies and differentiate interdependencies, consistently achieving state-of-the-art across various scenarios.
Authors: F M Ahosanul Haque, Md. Mahedi Hassan
Abstract: Banks are important for the development of economies in any financial ecosystem through consumer and business loans. Lending, however, presents risks; thus, banks have to determine the applicant's financial position to reduce the probabilities of default. A number of banks have currently, therefore, adopted data analytics and state-of-the-art technology to arrive at better decisions in the process. The probability of payback is prescribed by a predictive modeling technique in which machine learning algorithms are applied. In this research project, we will apply several machine learning methods to further improve the accuracy and efficiency of loan approval processes. Our work focuses on the prediction of bank loan approval; we have worked on a dataset of 148,670 instances and 37 attributes using machine learning methods. The target property segregates the loan applications into "Approved" and "Denied" groups. various machine learning techniques have been used, namely, Decision Tree Categorization, AdaBoosting, Random Forest Classifier, SVM, and GaussianNB. Following that, the models were trained and evaluated. Among these, the best-performing algorithm was AdaBoosting, which achieved an incredible accuracy of 99.99%. The results therefore show how ensemble learning works effectively to improve the prediction skills of loan approval decisions. The presented work points to the possibility of achieving extremely accurate and efficient loan prediction models that provide useful insights for applying machine learning to financial domains.
Authors: Katharine Daly, Hubert Eichner, Peter Kairouz, H. Brendan McMahan, Daniel Ramage, Zheng Xu
Abstract: Federated Learning (FL) is a machine learning technique that enables multiple entities to collaboratively learn a shared model without exchanging their local data. Over the past decade, FL systems have achieved substantial progress, scaling to millions of devices across various learning domains while offering meaningful differential privacy (DP) guarantees. Production systems from organizations like Google, Apple, and Meta demonstrate the real-world applicability of FL. However, key challenges remain, including verifying server-side DP guarantees and coordinating training across heterogeneous devices, limiting broader adoption. Additionally, emerging trends such as large (multi-modal) models and blurred lines between training, inference, and personalization challenge traditional FL frameworks. In response, we propose a redefined FL framework that prioritizes privacy principles rather than rigid definitions. We also chart a path forward by leveraging trusted execution environments and open-source ecosystems to address these challenges and facilitate future advancements in FL.
Authors: Wenlong Wang, Ivana Dusparic, Yucheng Shi, Ke Zhang, Vinny Cahill
Abstract: Model-based reinforcement learning (RL) offers a solution to the data inefficiency that plagues most model-free RL algorithms. However, learning a robust world model often demands complex and deep architectures, which are expensive to compute and train. Within the world model, dynamics models are particularly crucial for accurate predictions, and various dynamics-model architectures have been explored, each with its own set of challenges. Currently, recurrent neural network (RNN) based world models face issues such as vanishing gradients and difficulty in capturing long-term dependencies effectively. In contrast, use of transformers suffers from the well-known issues of self-attention mechanisms, where both memory and computational complexity scale as $O(n^2)$, with $n$ representing the sequence length. To address these challenges we propose a state space model (SSM) based world model, specifically based on Mamba, that achieves $O(n)$ memory and computational complexity while effectively capturing long-term dependencies and facilitating the use of longer training sequences efficiently. We also introduce a novel sampling method to mitigate the suboptimality caused by an incorrect world model in the early stages of training, combining it with the aforementioned technique to achieve a normalised score comparable to other state-of-the-art model-based RL algorithms using only a 7 million trainable parameter world model. This model is accessible and can be trained on an off-the-shelf laptop. Our code is available at https://github.com/realwenlongwang/drama.git.
Authors: Claas A Voelcker, Marcel Hussing, Eric Eaton, Amir-massoud Farahmand, Igor Gilitschenski
Abstract: Building deep reinforcement learning (RL) agents that find a good policy with few samples has proven notoriously challenging. To achieve sample efficiency, recent work has explored updating neural networks with large numbers of gradient steps for every new sample. While such high update-to-data (UTD) ratios have shown strong empirical performance, they also introduce instability to the training process. Previous approaches need to rely on periodic neural network parameter resets to address this instability, but restarting the training process is infeasible in many real-world applications and requires tuning the resetting interval. In this paper, we focus on one of the core difficulties of stable training with limited samples: the inability of learned value functions to generalize to unobserved on-policy actions. We mitigate this issue directly by augmenting the off-policy RL training process with a small amount of data generated from a learned world model. Our method, Model-Augmented Data for Temporal Difference learning (MAD-TD) uses small amounts of generated data to stabilize high UTD training and achieve competitive performance on the most challenging tasks in the DeepMind control suite. Our experiments further highlight the importance of employing a good model to generate data, MAD-TD's ability to combat value overestimation, and its practical stability gains for continued learning.
Authors: Yang Chen, Yitao Liang, Zhouchen Lin
Abstract: Low-Dimension-to-High-Dimension (LDHD) generalization is a special case of Out-of-Distribution (OOD) generalization, where the training data are restricted to a low-dimensional subspace of the high-dimensional testing space. Assuming that each instance is generated from a latent variable and the dimension of the latent variable reflects the problem scale, the inherent scaling challenge in length generalization can be captured by the LDHD generalization in the latent space. We theoretically demonstrate that LDHD generalization is generally unattainable without exploiting prior knowledge to provide appropriate inductive bias. Specifically, we explore LDHD generalization in Boolean functions. We verify that different architectures trained with (S)GD converge to \emph{min-degree interpolators w.r.t. different independent sets}. LDHD generalization is achievable if and only if the target function coincides with this inductive bias. Applying the insights from LDHD generalization to length generalization, we explain the effectiveness of CoT as changing the structure latent space to enable better LDHD generalization. We also propose a principle for position embedding design to handle both the inherent LDHD generalization and the nuisances such as the data format. Following the principle, we propose a novel position embedding called RPE-Square that remedies the RPE for dealing with the data format nuisance.
Authors: Yuwei Geng, Olena Burkovska, Lili Ju, Guannan Zhang, Max Gunzburger
Abstract: We propose an efficient end-to-end deep learning method for solving nonlocal Allen-Cahn (AC) and Cahn-Hilliard (CH) phase-field models. One motivation for this effort emanates from the fact that discretized partial differential equation-based AC or CH phase-field models result in diffuse interfaces between phases, with the only recourse for remediation is to severely refine the spatial grids in the vicinity of the true moving sharp interface whose width is determined by a grid-independent parameter that is substantially larger than the local grid size. In this work, we introduce non-mass conserving nonlocal AC or CH phase-field models with regular, logarithmic, or obstacle double-well potentials. Because of non-locality, some of these models feature totally sharp interfaces separating phases. The discretization of such models can lead to a transition between phases whose width is only a single grid cell wide. Another motivation is to use deep learning approaches to ameliorate the otherwise high cost of solving discretized nonlocal phase-field models. To this end, loss functions of the customized neural networks are defined using the residual of the fully discrete approximations of the AC or CH models, which results from applying a Fourier collocation method and a temporal semi-implicit approximation. To address the long-range interactions in the models, we tailor the architecture of the neural network by incorporating a nonlocal kernel as an input channel to the neural network model. We then provide the results of extensive computational experiments to illustrate the accuracy, structure-preserving properties, predictive capabilities, and cost reductions of the proposed method.
Authors: Ruinan Wang, Ian Nabney, Mohammad Golbabaee
Abstract: Hyperparameter selection is an essential aspect of the machine learning pipeline, profoundly impacting models' robustness, stability, and generalization capabilities. Given the complex hyperparameter spaces associated with Neural Networks and the constraints of computational resources and time, optimizing all hyperparameters becomes impractical. In this context, leveraging hyperparameter importance assessment (HIA) can provide valuable guidance by narrowing down the search space. This enables machine learning practitioners to focus their optimization efforts on the hyperparameters with the most significant impact on model performance while conserving time and resources. This paper aims to quantify the importance weights of some hyperparameters in Convolutional Neural Networks (CNNs) with an algorithm called N-RReliefF, laying the groundwork for applying HIA methodologies in the Deep Learning field. We conduct an extensive study by training over ten thousand CNN models across ten popular image classification datasets, thereby acquiring a comprehensive dataset containing hyperparameter configuration instances and their corresponding performance metrics. It is demonstrated that among the investigated hyperparameters, the top five important hyperparameters of the CNN model are the number of convolutional layers, learning rate, dropout rate, optimizer and epoch.
Authors: Matt L. Sampson, Peter Melchior
Abstract: Latent ODE models provide flexible descriptions of dynamic systems, but they can struggle with extrapolation and predicting complicated non-linear dynamics. The latent ODE approach implicitly relies on encoders to identify unknown system parameters and initial conditions, whereas the evaluation times are known and directly provided to the ODE solver. This dichotomy can be exploited by encouraging time-independent latent representations. By replacing the common variational penalty in latent space with an $\ell_2$ penalty on the path length of each system, the models learn data representations that can easily be distinguished from those of systems with different configurations. This results in faster training, smaller models, more accurate interpolation and long-time extrapolation compared to the baseline ODE models with GRU, RNN, and LSTM encoder/decoders on tests with damped harmonic oscillator, self-gravitating fluid, and predator-prey systems. We also demonstrate superior results for simulation-based inference of the Lotka-Volterra parameters and initial conditions by using the latents as data summaries for a conditional normalizing flow. Our change to the training loss is agnostic to the specific recognition network used by the decoder and can therefore easily be adopted by other latent ODE models.
Authors: Yuchen Ma, Valentyn Melnychuk, Jonas Schweisthal, Stefan Feuerriegel
Abstract: Predicting potential outcomes of interventions from observational data is crucial for decision-making in medicine, but the task is challenging due to the fundamental problem of causal inference. Existing methods are largely limited to point estimates of potential outcomes with no uncertain quantification; thus, the full information about the distributions of potential outcomes is typically ignored. In this paper, we propose a novel causal diffusion model called DiffPO, which is carefully designed for reliable inferences in medicine by learning the distribution of potential outcomes. In our DiffPO, we leverage a tailored conditional denoising diffusion model to learn complex distributions, where we address the selection bias through a novel orthogonal diffusion loss. Another strength of our DiffPO method is that it is highly flexible (e.g., it can also be used to estimate different causal quantities such as CATE). Across a wide range of experiments, we show that our method achieves state-of-the-art performance.
Authors: Maximilian Xiling Li, Korbinian Franz Rudolf, Nils Blank, Rudolf Lioutikov
Abstract: Prototype Learning methods provide an interpretable alternative to black-box deep learning models. Approaches such as ProtoPNet learn, which part of a test image "look like" known prototypical parts from training images, combining predictive power with the inherent interpretability of case-based reasoning. However, existing approaches have two main drawbacks: A) They rely solely on deterministic similarity scores without statistical confidence. B) The prototypes are learned in a black-box manner without human input. This work introduces HyperPg, a new prototype representation leveraging Gaussian distributions on a hypersphere in latent space, with learnable mean and variance. HyperPg prototypes adapt to the spread of clusters in the latent space and output likelihood scores. The new architecture, HyperPgNet, leverages HyperPg to learn prototypes aligned with human concepts from pixel-level annotations. Consequently, each prototype represents a specific concept such as color, image texture, or part of the image subject. A concept extraction pipeline built on foundation models provides pixel-level annotations, significantly reducing human labeling effort. Experiments on CUB-200-2011 and Stanford Cars datasets demonstrate that HyperPgNet outperforms other prototype learning architectures while using fewer parameters and training steps. Additionally, the concept-aligned HyperPg prototypes are learned transparently, enhancing model interpretability.
Authors: Clayton Leite, Yu Xiao
Abstract: Text-to-motion models that generate sequences of human poses from textual descriptions are garnering significant attention. However, due to data scarcity, the range of motions these models can produce is still limited. For instance, current text-to-motion models cannot generate a motion of kicking a football with the instep of the foot, since the training data only includes martial arts kicks. We propose a novel method that uses short video clips or images as conditions to modify existing basic motions. In this approach, the model's understanding of a kick serves as the prior, while the video or image of a football kick acts as the posterior, enabling the generation of the desired motion. By incorporating these additional modalities as conditions, our method can create motions not present in the training set, overcoming the limitations of text-motion datasets. A user study with 26 participants demonstrated that our approach produces unseen motions with realism comparable to commonly represented motions in text-motion datasets (e.g., HumanML3D), such as walking, running, squatting, and kicking.
Authors: Aymane El Firdoussi, Mohamed El Amine Seddik, Soufiane Hayou, Reda Alami, Ahmed Alzubaidi, Hakim Hacid
Abstract: Synthetic data has gained attention for training large language models, but poor-quality data can harm performance (see, e.g., Shumailov et al. (2023); Seddik et al. (2024)). A potential solution is data pruning, which retains only high-quality data based on a score function (human or machine feedback). Previous work Feng et al. (2024) analyzed models trained on synthetic data as sample size increases. We extend this by using random matrix theory to derive the performance of a binary classifier trained on a mix of real and pruned synthetic data in a high dimensional setting. Our findings identify conditions where synthetic data could improve performance, focusing on the quality of the generative model and verification strategy. We also show a smooth phase transition in synthetic label noise, contrasting with prior sharp behavior in infinite sample limits. Experiments with toy models and large language models validate our theoretical results.
Authors: Weijia Zhang, Jindong Han, Hao Liu, Wei Fan, Hao Wang, Hui Xiong
Abstract: Real estate appraisal is important for a variety of endeavors such as real estate deals, investment analysis, and real property taxation. Recently, deep learning has shown great promise for real estate appraisal by harnessing substantial online transaction data from web platforms. Nonetheless, deep learning is data-hungry, and thus it may not be trivially applicable to enormous small cities with limited data. To this end, we propose Meta-Transfer Learning Empowered Temporal Graph Networks (MetaTransfer) to transfer valuable knowledge from multiple data-rich metropolises to the data-scarce city to improve valuation performance. Specifically, by modeling the ever-growing real estate transactions with associated residential communities as a temporal event heterogeneous graph, we first design an Event-Triggered Temporal Graph Network to model the irregular spatiotemporal correlations between evolving real estate transactions. Besides, we formulate the city-wide real estate appraisal as a multi-task dynamic graph link label prediction problem, where the valuation of each community in a city is regarded as an individual task. A Hypernetwork-Based Multi-Task Learning module is proposed to simultaneously facilitate intra-city knowledge sharing between multiple communities and task-specific parameters generation to accommodate the community-wise real estate price distribution. Furthermore, we propose a Tri-Level Optimization Based Meta- Learning framework to adaptively re-weight training transaction instances from multiple source cities to mitigate negative transfer, and thus improve the cross-city knowledge transfer effectiveness. Finally, extensive experiments based on five real-world datasets demonstrate the significant superiority of MetaTransfer compared with eleven baseline algorithms.
Authors: Yisen Wang, Yichuan Mo, Dongxian Wu, Mingjie Li, Xingjun Ma, Zhouchen Lin
Abstract: Skip connection is an essential ingredient for modern deep models to be deeper and more powerful. Despite their huge success in normal scenarios (state-of-the-art classification performance on natural examples), we investigate and identify an interesting property of skip connections under adversarial scenarios, namely, the use of skip connections allows easier generation of highly transferable adversarial examples. Specifically, in ResNet-like models (with skip connections), we find that using more gradients from the skip connections rather than the residual modules according to a decay factor during backpropagation allows one to craft adversarial examples with high transferability. The above method is termed as Skip Gradient Method (SGM). Although starting from ResNet-like models in vision domains, we further extend SGM to more advanced architectures, including Vision Transformers (ViTs) and models with length-varying paths and other domains, i.e. natural language processing. We conduct comprehensive transfer attacks against various models including ResNets, Transformers, Inceptions, Neural Architecture Search, and Large Language Models (LLMs). We show that employing SGM can greatly improve the transferability of crafted attacks in almost all cases. Furthermore, considering the big complexity for practical use, we further demonstrate that SGM can even improve the transferability on ensembles of models or targeted attacks and the stealthiness against current defenses. At last, we provide theoretical explanations and empirical insights on how SGM works. Our findings not only motivate new adversarial research into the architectural characteristics of models but also open up further challenges for secure model architecture design. Our code is available at https://github.com/mo666666/SGM.
Authors: Arthur Mendon\c{c}a Sasse, Claudio Miceli de Farias
Abstract: Federated Kolmogorov-Arnold Networks (F-KANs) have already been proposed, but their assessment is at an initial stage. We present a comparison between KANs (using B-splines and Radial Basis Functions as activation functions) and Multi- Layer Perceptrons (MLPs) with a similar number of parameters for 100 rounds of federated learning in the MNIST classification task using non-IID partitions with 100 clients. After 15 trials for each model, we show that the best accuracies achieved by MLPs can be achieved by Spline-KANs in half of the time (in rounds), with just a moderate increase in computing time.
Authors: Michalis Korakakis, Andreas Vlachos, Adrian Weller
Abstract: Active Learning aims to minimize annotation effort by selecting the most useful instances from a pool of unlabeled data. However, typical active learning methods overlook the presence of distinct example groups within a class, whose prevalence may vary, e.g., in occupation classification datasets certain demographics are disproportionately represented in specific classes. This oversight causes models to rely on shortcuts for predictions, i.e., spurious correlations between input attributes and labels occurring in well-represented groups. To address this issue, we propose Active Learning Via INterpolation (ALVIN), which conducts intra-class interpolations between examples from under-represented and well-represented groups to create anchors, i.e., artificial points situated between the example groups in the representation space. By selecting instances close to the anchors for annotation, ALVIN identifies informative examples exposing the model to regions of the representation space that counteract the influence of shortcuts. Crucially, since the model considers these examples to be of high certainty, they are likely to be ignored by typical active learning methods. Experimental results on six datasets encompassing sentiment analysis, natural language inference, and paraphrase detection demonstrate that ALVIN outperforms state-of-the-art active learning methods in both in-distribution and out-of-distribution generalization.
Authors: Jonas Schweisthal, Dennis Frauen, Maresa Schr\"oder, Konstantin Hess, Niki Kilbertus, Stefan Feuerriegel
Abstract: Reliable estimation of treatment effects from observational data is important in many disciplines such as medicine. However, estimation is challenging when unconfoundedness as a standard assumption in the causal inference literature is violated. In this work, we leverage arbitrary (potentially high-dimensional) instruments to estimate bounds on the conditional average treatment effect (CATE). Our contributions are three-fold: (1) We propose a novel approach for partial identification through a mapping of instruments to a discrete representation space so that we yield valid bounds on the CATE. This is crucial for reliable decision-making in real-world applications. (2) We derive a two-step procedure that learns tight bounds using a tailored neural partitioning of the latent instrument space. As a result, we avoid instability issues due to numerical approximations or adversarial training. Furthermore, our procedure aims to reduce the estimation variance in finite-sample settings to yield more reliable estimates. (3) We show theoretically that our procedure obtains valid bounds while reducing estimation variance. We further perform extensive experiments to demonstrate the effectiveness across various settings. Overall, our procedure offers a novel path for practitioners to make use of potentially high-dimensional instruments (e.g., as in Mendelian randomization).
Authors: Devdhar Patel, Hava Siegelmann
Abstract: Reinforcement learning (RL) is rapidly reaching and surpassing human-level control capabilities. However, state-of-the-art RL algorithms often require timesteps and reaction times significantly faster than human capabilities, which is impractical in real-world settings and typically necessitates specialized hardware. Such speeds are difficult to achieve in the real world and often requires specialized hardware. We introduce Sequence Reinforcement Learning (SRL), an RL algorithm designed to produce a sequence of actions for a given input state, enabling effective control at lower decision frequencies. SRL addresses the challenges of learning action sequences by employing both a model and an actor-critic architecture operating at different temporal scales. We propose a "temporal recall" mechanism, where the critic uses the model to estimate intermediate states between primitive actions, providing a learning signal for each individual action within the sequence. Once training is complete, the actor can generate action sequences independently of the model, achieving model-free control at a slower frequency. We evaluate SRL on a suite of continuous control tasks, demonstrating that it achieves performance comparable to state-of-the-art algorithms while significantly reducing actor sample complexity. To better assess performance across varying decision frequencies, we introduce the Frequency-Averaged Score (FAS) metric. Our results show that SRL significantly outperforms traditional RL algorithms in terms of FAS, making it particularly suitable for applications requiring variable decision frequencies. Additionally, we compare SRL with model-based online planning, showing that SRL achieves superior FAS while leveraging the same model during training that online planners use for planning.
Authors: Ziming Yu, Pan Zhou, Sike Wang, Jia Li, Hua Huang
Abstract: Fine-tuning Large Language Models (LLMs) has proven effective for a variety of downstream tasks. However, as LLMs grow in size, the memory demands for backpropagation become increasingly prohibitive. Zeroth-order (ZO) optimization methods offer a memory-efficient alternative by using forward passes to estimate gradients, but the variance of gradient estimates typically scales linearly with the model's parameter dimension$\unicode{x2013}$a significant issue for LLMs. In this paper, we propose the random Subspace Zeroth-order (SubZero) optimization to address the challenges posed by LLMs' high dimensionality. We introduce a low-rank perturbation tailored for LLMs that significantly reduces memory consumption while improving training performance. Additionally, we prove that our gradient estimation closely approximates the backpropagation gradient, exhibits lower variance than traditional ZO methods, and ensures convergence when combined with SGD. Experimental results show that SubZero enhances fine-tuning performance and achieves faster convergence compared to standard ZO approaches like MeZO across various language modeling tasks.
Authors: Rushiv Arora
Abstract: There have been key advancements to building universal approximators for multi-goal collections of reinforcement learning value functions -- key elements in estimating long-term returns of states in a parameterized manner. We extend this to hierarchical reinforcement learning, using the options framework, by introducing hierarchical universal value function approximators (H-UVFAs). This allows us to leverage the added benefits of scaling, planning, and generalization expected in temporal abstraction settings. We develop supervised and reinforcement learning methods for learning embeddings of the states, goals, options, and actions in the two hierarchical value functions: $Q(s, g, o; \theta)$ and $Q(s, g, o, a; \theta)$. Finally we demonstrate generalization of the HUVFAs and show they outperform corresponding UVFAs.
Authors: Kevin Galim, Wonjun Kang, Yuchen Zeng, Hyung Il Koo, Kangwook Lee
Abstract: Deep State Space Models (SSMs), such as Mamba (Gu & Dao, 2024), have emerged as powerful tools for language modeling, offering high performance with efficient inference and linear scaling in sequence length. However, the application of parameter-efficient fine-tuning (PEFT) methods to SSM-based models remains largely unexplored. This paper aims to systematically study two key questions: (i) How do existing PEFT methods perform on SSM-based models? (ii) Which modules are most effective for fine-tuning? We conduct an empirical benchmark of four basic PEFT methods on SSM-based models. Our findings reveal that prompt-based methods (e.g., prefix-tuning) are no longer effective, an empirical result further supported by theoretical analysis. In contrast, LoRA remains effective for SSM-based models. We further investigate the optimal application of LoRA within these models, demonstrating both theoretically and experimentally that applying LoRA to linear projection matrices without modifying SSM modules yields the best results, as LoRA is not effective at tuning SSM modules. To further improve performance, we introduce LoRA with Selective Dimension tuning (SDLoRA), which selectively updates certain channels and states on SSM modules while applying LoRA to linear projection matrices. Extensive experimental results show that this approach outperforms standard LoRA.
Authors: Maksym Andriushchenko, Alexandra Souly, Mateusz Dziemian, Derek Duenas, Maxwell Lin, Justin Wang, Dan Hendrycks, Andy Zou, Zico Kolter, Matt Fredrikson, Eric Winsor, Jerome Wynne, Yarin Gal, Xander Davies
Abstract: The robustness of LLMs to jailbreak attacks, where users design prompts to circumvent safety measures and misuse model capabilities, has been studied primarily for LLMs acting as simple chatbots. Meanwhile, LLM agents -- which use external tools and can execute multi-stage tasks -- may pose a greater risk if misused, but their robustness remains underexplored. To facilitate research on LLM agent misuse, we propose a new benchmark called AgentHarm. The benchmark includes a diverse set of 110 explicitly malicious agent tasks (440 with augmentations), covering 11 harm categories including fraud, cybercrime, and harassment. In addition to measuring whether models refuse harmful agentic requests, scoring well on AgentHarm requires jailbroken agents to maintain their capabilities following an attack to complete a multi-step task. We evaluate a range of leading LLMs, and find (1) leading LLMs are surprisingly compliant with malicious agent requests without jailbreaking, (2) simple universal jailbreak templates can be adapted to effectively jailbreak agents, and (3) these jailbreaks enable coherent and malicious multi-step agent behavior and retain model capabilities. We publicly release AgentHarm to enable simple and reliable evaluation of attacks and defenses for LLM-based agents. We publicly release the benchmark at https://huggingface.co/ai-safety-institute/AgentHarm.
Authors: Wanyu Bian
Abstract: This paper presents a comprehensive analysis of both conventional and deep learning methods for eliminating electromagnetic interference (EMI) in MRI systems. We explore the underlying principles and implementation of traditional analytical and adaptive EMI elimination techniques, as well as cutting-edge deep learning approaches. Through a detailed comparison, the strengths and limitations of each method are highlighted. Recent advancements in active EMI elimination utilizing multiple external EMI receiver coils and analytical techniques are discussed alongside the superior performance of deep learning methods, which leverage neural networks trained on extensive MRI data. While deep learning methods demonstrate significant improvements in EMI suppression, enhancing diagnostic capabilities and accessibility of MRI technology, they also introduce potential security and safety concerns, especially in production and commercial applications. This study underscores the need to address these challenges to fully realize the benefits of deep learning in EMI elimination. The findings suggest a balanced approach, combining the reliability of conventional methods with the advanced capabilities of deep learning, to develop more robust and effective EMI suppression strategies in MRI systems.
Authors: Xinyuan Wang, Haozhou Li, Dingfang Zheng, Qinke Peng
Abstract: The global COVID-19 pandemic underscored major deficiencies in traditional healthcare systems, hastening the advancement of online medical services, especially in medical triage and consultation. However, existing studies face two main challenges. First, the scarcity of large-scale, publicly available, domain-specific medical datasets due to privacy concerns, with current datasets being small and limited to a few diseases, limiting the effectiveness of triage methods based on Pre-trained Language Models (PLMs). Second, existing methods lack medical knowledge and struggle to accurately understand professional terms and expressions in patient-doctor consultations. To overcome these obstacles, we construct the Large-scale Chinese Medical Dialogue Corpora (LCMDC), comprising a Coarse-grained Triage dataset with 439,630 samples, a Fine-grained Diagnosis dataset with 199,600 samples, and a Medical Consultation dataset with 472,418 items, thereby addressing the data shortage in this field. Moreover, we further propose a novel triage system that combines BERT-based supervised learning with prompt learning, as well as a GPT-based medical consultation model using reinforcement learning. To enhance domain knowledge acquisition, we pre-trained PLMs using our self-constructed background corpus. Experimental results on the LCMDC demonstrate the efficacy of our proposed systems.
Authors: Sam Gunn, Xuandong Zhao, Dawn Song
Abstract: We present the first undetectable watermarking scheme for generative image models. Undetectability ensures that no efficient adversary can distinguish between watermarked and un-watermarked images, even after making many adaptive queries. In particular, an undetectable watermark does not degrade image quality under any efficiently computable metric. Our scheme works by selecting the initial latents of a diffusion model using a pseudorandom error-correcting code (Christ and Gunn, 2024), a strategy which guarantees undetectability and robustness. We experimentally demonstrate that our watermarks are quality-preserving and robust using Stable Diffusion 2.1. Our experiments verify that, in contrast to every prior scheme we tested, our watermark does not degrade image quality. Our experiments also demonstrate robustness: existing watermark removal attacks fail to remove our watermark from images without significantly degrading the quality of the images. Finally, we find that we can robustly encode 512 bits in our watermark, and up to 2500 bits when the images are not subjected to watermark removal attacks. Our code is available at https://github.com/XuandongZhao/PRC-Watermark.
Authors: Marwan Hamze (LISV), Mitsuharu Morisawa (AIST), Eiichi Yoshida (CNRS-AIST JRL)
Abstract: Deep reinforcement learning has seen successful implementations on humanoid robots to achieve dynamic walking. However, these implementations have been so far successful in simple environments void of obstacles. In this paper, we aim to achieve bipedal locomotion in an environment where obstacles are present using a policy-based reinforcement learning. By adding simple distance reward terms to a state of art reward function that can achieve basic bipedal locomotion, the trained policy succeeds in navigating the robot towards the desired destination without colliding with the obstacles along the way.
Authors: Yuqing He, Yousef Heider, Bernd Markert
Abstract: This manuscript presents a practical method for incorporating trained Neural Networks (NNs) into the Finite Element (FE) framework using a user material (UMAT) subroutine. The work exemplifies crystal plasticity, a complex inelastic non-linear path-dependent material response, with a wide range of applications in ABAQUS UMAT. However, this approach can be extended to other material behaviors and FE tools. The use of a UMAT subroutine serves two main purposes: (1) it predicts and updates the stress or other mechanical properties of interest directly from the strain history; (2) it computes the Jacobian matrix either through backpropagation or numerical differentiation, which plays an essential role in the solution convergence. By implementing NNs in a UMAT subroutine, a trained machine learning model can be employed as a data-driven constitutive law within the FEM framework, preserving multiscale information that conventional constitutive laws often neglect or average. The versatility of this method makes it a powerful tool for integrating machine learning into mechanical simulation. While this approach is expected to provide higher accuracy in reproducing realistic material behavior, the reliability of the solution process and the convergence conditions must be paid special attention. While the theory of the model is explained in [Heider et al. 2020], exemplary source code is also made available for interested readers [https://doi.org/10.25835/6n5uu50y]
Authors: Akash Agrawal, Mayesh Mohapatra, Abhinav Raja, Paritosh Tiwari, Vishwajeet Pattanaik, Neeru Jaiswal, Arpit Agarwal, Punit Rathore
Abstract: Estimating the location and intensity of tropical cyclones holds crucial significance for predicting catastrophic weather events. In this study, we approach this task as a detection and regression challenge, specifically over the North Indian Ocean (NIO) region where best tracks location and wind speed information serve as the labels. The current process for cyclone detection and intensity estimation involves physics-based simulation studies which are time-consuming, only using image features will automate the process for significantly faster and more accurate predictions. While conventional methods typically necessitate substantial prior knowledge for training, we are exploring alternative approaches to enhance efficiency. This research aims to focus specifically on cyclone detection, intensity estimation and related aspects using only image input and data-driven approaches and will lead to faster inference time and automate the process as opposed to current NWP models being utilized at SAC. In context to algorithm development, a novel two stage detection and intensity estimation module is proposed. In the first level detection we try to localize the cyclone over an entire image as captured by INSAT3D over the NIO (North Indian Ocean). For the intensity estimation task, we propose a CNN-LSTM network, which works on the cyclone centered images, utilizing a ResNet-18 backbone, by which we are able to capture both temporal and spatial characteristics.
Authors: Yulong Feng, Jing Xu, Liujun Hu, Guanghui Yu
Abstract: Semantic communication technology emerges as a pivotal bridge connecting AI with classical communication. The current semantic communication systems are generally modeled as an Auto-Encoder (AE). AE lacks a deep integration of AI principles with communication strategies due to its inability to effectively capture channel dynamics. This gap makes it difficult to justify the need for joint source-channel coding (JSCC) and to explain why performance improves. This paper begins by exploring lossless and lossy communication, highlighting that the inclusion of data distortion distinguishes semantic communication from classical communication. It breaks the conditions for the separation theorem to hold and explains why the amount of data transferred by semantic communication is less. Therefore, employing JSCC becomes imperative for achieving optimal semantic communication. Moreover, a Variational Source-Channel Coding (VSCC) method is proposed for constructing semantic communication systems based on data distortion theory, integrating variational inference and channel characteristics. Using a deep learning network, we develop a semantic communication system employing the VSCC method and demonstrate its capability for semantic transmission. We also establish semantic communication systems of equivalent complexity employing the AE method and the VAE method. Experimental results reveal that the VSCC model offers superior interpretability compared to AE model, as it clearly captures the semantic features of the transmitted data, represented as the variance of latent variables in our experiments. In addition, VSCC model exhibits superior semantic transmission capabilities compared to VAE model. At the same level of data distortion evaluated by PSNR, VSCC model exhibits stronger human interpretability, which can be partially assessed by SSIM.
Authors: Pengfei Wang, Huanran Zheng, Silong Dai, Yiqiao Wang, Xiaotian Gu, Yuanbin Wu, Xiaoling Wang
Abstract: In recent years, the field of electroencephalography (EEG) analysis has witnessed remarkable advancements, driven by the integration of machine learning and artificial intelligence. This survey aims to encapsulate the latest developments, focusing on emerging methods and technologies that are poised to transform our comprehension and interpretation of brain activity. We delve into self-supervised learning methods that enable the robust representation of brain signals, which are fundamental for a variety of downstream applications. We also explore emerging discriminative methods, including graph neural networks (GNN), foundation models, and large language models (LLMs)-based approaches. Furthermore, we examine generative technologies that harness EEG data to produce images or text, offering novel perspectives on brain activity visualization and interpretation. The survey provides an extensive overview of these cutting-edge techniques, their current applications, and the profound implications they hold for future research and clinical practice. The relevant literature and open-source materials have been compiled and are consistently being refreshed at \url{https://github.com/wpf535236337/LLMs4TS}
Authors: Samuel Stockman, Daniel Lawson, Maximilian Werner
Abstract: Classical point process models, such as the epidemic-type aftershock sequence (ETAS) model, have been widely used for forecasting the event times and locations of earthquakes for decades. Recent advances have led to Neural Point Processes (NPPs), which promise greater flexibility and improvements over classical models. However, the currently-used benchmark dataset for NPPs does not represent an up-to-date challenge in the seismological community since it lacks a key earthquake sequence from the region and improperly splits training and testing data. Furthermore, initial earthquake forecast benchmarking lacks a comparison to state-of-the-art earthquake forecasting models typically used by the seismological community. To address these gaps, we introduce EarthquakeNPP: a collection of benchmark datasets to facilitate testing of NPPs on earthquake data, accompanied by a credible implementation of the ETAS model. The datasets cover a range of small to large target regions within California, dating from 1971 to 2021, and include different methodologies for dataset generation. In a benchmarking experiment, we compare three spatio-temporal NPPs against ETAS and find that none outperform ETAS in either spatial or temporal log-likelihood. These results indicate that current NPP implementations are not yet suitable for practical earthquake forecasting. However, EarthquakeNPP will serve as a platform for collaboration between the seismology and machine learning communities with the goal of improving earthquake predictability.
Authors: Jiaxing Xu, Mengcheng Lan, Xia Dong, Kai He, Wei Zhang, Qingtian Bian, Yiping Ke
Abstract: In the realm of neuroscience, identifying distinctive patterns associated with neurological disorders via brain networks is crucial. Resting-state functional magnetic resonance imaging (fMRI) serves as a primary tool for mapping these networks by correlating blood-oxygen-level-dependent (BOLD) signals across different brain regions, defined as regions of interest (ROIs). Constructing these brain networks involves using atlases to parcellate the brain into ROIs based on various hypotheses of brain division. However, there is no standard atlas for brain network classification, leading to limitations in detecting abnormalities in disorders. Some recent methods have proposed utilizing multiple atlases, but they neglect consistency across atlases and lack ROI-level information exchange. To tackle these limitations, we propose an Atlas-Integrated Distillation and Fusion network (AIDFusion) to improve brain network classification using fMRI data. AIDFusion addresses the challenge of utilizing multiple atlases by employing a disentangle Transformer to filter out inconsistent atlas-specific information and distill distinguishable connections across atlases. It also incorporates subject- and population-level consistency constraints to enhance cross-atlas consistency. Additionally, AIDFusion employs an inter-atlas message-passing mechanism to fuse complementary information across brain regions. Experimental results on four datasets of different diseases demonstrate the effectiveness and efficiency of AIDFusion compared to state-of-the-art methods. A case study illustrates AIDFusion extract patterns that are both interpretable and consistent with established neuroscience findings.
Authors: Shahriar Ahmad Fahim
Abstract: Rapid urbanization in megacities around the world, like Dhaka, has caused numerous transportation challenges that need to be addressed. Emerging technologies of deep learning and artificial intelligence can help us solve these problems to move towards Intelligent Transportation Systems (ITS) in the city. The government of Bangladesh recognizes the integration of ITS to ensure smart mobility as a vital step towards the development plan "Smart Bangladesh Vision 2041", but faces challenges in understanding ITS, its effects, and directions to implement. A vehicle detection system can pave the way to understanding traffic congestion, finding mobility patterns, and ensuring traffic surveillance. So, this paper proposes a fine-tuned object detector, the YOLOv9 model to detect native vehicles trained on a Bangladesh-based dataset. Results show that the fine-tuned YOLOv9 model achieved a mean Average Precision (mAP) of 0.934 at the Intersection over Union (IoU) threshold of 0.5, achieving state-of-the-art performance over past studies on Bangladesh-based datasets, shown through a comparison. Later, by suggesting the model to be deployed on CCTVs (closed circuit television) on the roads, a conceptual technique is proposed to process the vehicle detection model output data in a graph structure creating a vehicle detection system in the city. Finally, applications of such vehicle detection system are discussed showing a framework on how it can solve further ITS research questions, to provide a rationale for policymakers to implement the proposed vehicle detection system in the city.
Authors: Pablo M. Barros, Roosevelt de L. Sardinha, Giovanny A. M. Arboleda, Lessandro de S. S. Valente, Isabelle R. V. de Melo, Albino Aveleda, Andr\'e Bulc\~ao, Sergio L. Netto, Alexandre G. Evsukoff
Abstract: The recent development of deep learning (DL) methods for computer vision has been driven by the creation of open benchmark datasets on which new algorithms can be tested and compared with reproducible results. Although DL methods have many applications in geophysics, few real seismic datasets are available for benchmarking DL models, especially for denoising real data, which is one of the main problems in seismic data processing scenarios in the oil and gas industry. This article presents a benchmark dataset composed of synthetic seismic data corrupted with noise extracted from a filtering process implemented on real data. In this work, a comparison between two well-known DL-based denoising models is conducted on this dataset, which is proposed as a benchmark for accelerating the development of new solutions for seismic data denoising. This work also introduces a new evaluation metric that can capture small variations in model results. The results show that DL models are effective at denoising seismic data, but some issues remain to be solved.
Authors: Kemal Altwlkany, Sead Delalic, Elmedin Selmanovic, Adis Alihodzic, Ivica Lovric
Abstract: In the field of telecommunications and cloud communications, accurately and in real-time detecting whether a human or an answering machine has answered an outbound call is of paramount importance. This problem is of particular significance during campaigns as it enhances service quality, efficiency and cost reduction through precise caller identification. Despite the significance of the field, it remains inadequately explored in the existing literature. This paper presents an innovative approach to answering machine detection that leverages transfer learning through the YAMNet model for feature extraction. The YAMNet architecture facilitates the training of a recurrent-based classifier, enabling real-time processing of audio streams, as opposed to fixed-length recordings. The results demonstrate an accuracy of over 96% on the test set. Furthermore, we conduct an in-depth analysis of misclassified samples and reveal that an accuracy exceeding 98% can be achieved with the integration of a silence detection algorithm, such as the one provided by FFmpeg.
Authors: Claudia Fournier, Raul Fernandez-Fernandez, Samuel Cir\'es, Jos\'e A. L\'opez-Orozco, Eva Besada-Portas, Antonio Quesada
Abstract: Cyanobacteria are the most frequent dominant species of algal blooms in inland waters, threatening ecosystem function and water quality, especially when toxin-producing strains predominate. Enhanced by anthropogenic activities and global warming, cyanobacterial blooms are expected to increase in frequency and global distribution. Early warning systems (EWS) for cyanobacterial blooms development allow timely implementation of management measures, reducing the risks associated to these blooms. In this paper, we propose an effective EWS for cyanobacterial bloom forecasting, which uses 6 years of incomplete high-frequency spatio-temporal data from multiparametric probes, including phycocyanin (PC) fluorescence as a proxy for cyanobacteria. A probe agnostic and replicable method is proposed to pre-process the data and to generate time series specific for cyanobacterial bloom forecasting. Using these pre-processed data, six different non-site/species-specific predictive models were compared including the autoregressive and multivariate versions of Linear Regression, Random Forest, and Long-Term Short-Term (LSTM) neural networks. Results were analyzed for seven forecasting time horizons ranging from 4 to 28 days evaluated with a hybrid system that combined regression metrics (MSE, R2, MAPE) for PC values, classification metrics (Accuracy, F1, Kappa) for a proposed alarm level of 10 ug PC/L, and a forecasting-specific metric to measure prediction improvement over the displaced signal (skill). The multivariate version of LSTM showed the best and most consistent results across all forecasting horizons and metrics, achieving accuracies of up to 90% in predicting the proposed PC alarm level. Additionally, positive skill values indicated its outstanding effectiveness to forecast cyanobacterial blooms from 16 to 28 days in advance.
Authors: F\'elix Marcoccia, C\'edric Adjih, Paul M\"uhlethaler
Abstract: This work introduces NetDiff, an expressive graph denoising diffusion probabilistic architecture that generates wireless ad hoc network link topologies. Such networks, with directional antennas, can achieve unmatched performance when the communication links are designed to provide good geometric properties, notably by reducing interference between these links while respecting diverse physical constraints. How to craft such a link assignment algorithm is yet a real problem. Deep graph generation offers multiple advantages compared to traditional approaches: it allows to relieve the network nodes of the communication burden caused by the search of viable links and to avoid resorting to heavy combinatorial methods to find a good link topology. Denoising diffusion also provides a built-in method to update the network over time. Given that graph neural networks sometimes tend to struggle with global, structural properties, we augment the popular graph transformer with cross-attentive modulation tokens in order to improve global control over the predicted topology. We also incorporate simple node and edge features, as well as additional loss terms, to facilitate the compliance with the network topology physical constraints. A network evolution algorithm based on partial diffusion is also proposed to maintain a stable network topology over time when the nodes move. Our results show that the generated links are realistic, present structural properties similar to the dataset graphs', and require only minor corrections and verification steps to be operational.
Authors: Nuria Rodr\'iguez-Barroso, M. Victoria Luz\'on, Francisco Herrera
Abstract: At the same time that artificial intelligence is becoming popular, concern and the need for regulation is growing, including among other requirements the data privacy. In this context, Federated Learning is proposed as a solution to data privacy concerns derived from different source data scenarios due to its distributed learning. The defense mechanisms proposed in literature are just focused on defending against adversarial attacks and the performance, leaving aside other important qualities such as explainability, fairness to poor quality clients, dynamism in terms of attacks configuration and generality in terms of being resilient against different kinds of attacks. In this work, we propose RAB$^2$-DEF, a $\textbf{r}$esilient $\textbf{a}$gainst $\textbf{b}\text{yzantine}$ and $\textbf{b}$ackdoor attacks which is $\textbf{d}$ynamic, $\textbf{e}$xplainable and $\textbf{f}$air to poor clients using local linear explanations. We test the performance of RAB$^2$-DEF in image datasets and both byzantine and backdoor attacks considering the state-of-the-art defenses and achieve that RAB$^2$-DEF is a proper defense at the same time that it boosts the other qualities towards trustworthy artificial intelligence.
Authors: Junyi Cao, Shanyan Guan, Yanhao Ge, Wei Li, Xiaokang Yang, Chao Ma
Abstract: While humans effortlessly discern intrinsic dynamics and adapt to new scenarios, modern AI systems often struggle. Current methods for visual grounding of dynamics either use pure neural-network-based simulators (black box), which may violate physical laws, or traditional physical simulators (white box), which rely on expert-defined equations that may not fully capture actual dynamics. We propose the Neural Material Adaptor (NeuMA), which integrates existing physical laws with learned corrections, facilitating accurate learning of actual dynamics while maintaining the generalizability and interpretability of physical priors. Additionally, we propose Particle-GS, a particle-driven 3D Gaussian Splatting variant that bridges simulation and observed images, allowing back-propagate image gradients to optimize the simulator. Comprehensive experiments on various dynamics in terms of grounded particle accuracy, dynamic rendering quality, and generalization ability demonstrate that NeuMA can accurately capture intrinsic dynamics.
Authors: Amanda Muyskens, Benjamin W. Priest, Imene R. Goumiri, Michael D. Schneider
Abstract: Kernels representing limiting cases of neural network architectures have recently gained popularity. However, the application and performance of these new kernels compared to existing options, such as the Matern kernel, is not well studied. We take a practical approach to explore the neural network Gaussian process (NNGP) kernel and its application to data in Gaussian process regression. We first demonstrate the necessity of normalization to produce valid NNGP kernels and explore related numerical challenges. We further demonstrate that the predictions from this model are quite inflexible, and therefore do not vary much over the valid hyperparameter sets. We then demonstrate a surprising result that the predictions given from the NNGP kernel correspond closely to those given by the Matern kernel under specific circumstances, which suggests a deep similarity between overparameterized deep neural networks and the Matern kernel. Finally, we demonstrate the performance of the NNGP kernel as compared to the Matern kernel on three benchmark data cases, and we conclude that for its flexibility and practical performance, the Matern kernel is preferred to the novel NNGP in practical applications.
Authors: Roberto Barcel\'o, Crist\'obal Alc\'azar, Felipe Tobar
Abstract: Fine-tuning foundation models via reinforcement learning (RL) has proven promising for aligning to downstream objectives. In the case of diffusion models (DMs), though RL training improves alignment from early timesteps, critical issues such as training instability and mode collapse arise. We address these drawbacks by exploiting the hierarchical nature of DMs: we train them dynamically at each epoch with a tailored RL method, allowing for continual evaluation and step-by-step refinement of the model performance (or alignment). Furthermore, we find that not every denoising step needs to be fine-tuned to align DMs to downstream tasks. Consequently, in addition to clipping, we regularise model parameters at distinct learning phases via a sliding-window approach. Our approach, termed Hierarchical Reward Fine-tuning (HRF), is validated on the Denoising Diffusion Policy Optimisation method, where we show that models trained with HRF achieve better preservation of diversity in downstream tasks, thus enhancing the fine-tuning robustness and at uncompromising mean rewards.
Authors: Zhuohang Li, Jiaxin Zhang, Chao Yan, Kamalika Das, Sricharan Kumar, Murat Kantarcioglu, Bradley A. Malin
Abstract: Language models (LMs) are known to suffer from hallucinations and misinformation. Retrieval augmented generation (RAG) that retrieves verifiable information from an external knowledge corpus to complement the parametric knowledge in LMs provides a tangible solution to these problems. However, the generation quality of RAG is highly dependent on the relevance between a user's query and the retrieved documents. Inaccurate responses may be generated when the query is outside of the scope of knowledge represented in the external knowledge corpus or if the information in the corpus is out-of-date. In this work, we establish a statistical framework that assesses how well a query can be answered by an RAG system by capturing the relevance of knowledge. We introduce an online testing procedure that employs goodness-of-fit (GoF) tests to inspect the relevance of each user query to detect out-of-knowledge queries with low knowledge relevance. Additionally, we develop an offline testing framework that examines a collection of user queries, aiming to detect significant shifts in the query distribution which indicates the knowledge corpus is no longer sufficiently capable of supporting the interests of the users. We demonstrate the capabilities of these strategies through a systematic evaluation on eight question-answering (QA) datasets, the results of which indicate that the new testing framework is an efficient solution to enhance the reliability of existing RAG systems.
Authors: Benno Weck, Frederic Font
Abstract: This study examines textual, user-written search queries within the context of sound search engines, encompassing various applications such as foley, sound effects, and general audio retrieval. Current research inadequately addresses real-world user needs and behaviours in designing text-based audio retrieval systems. To bridge this gap, we analysed search queries from two sources: a custom survey and Freesound website query logs. The survey was designed to collect queries for an unrestricted, hypothetical sound search engine, resulting in a dataset that captures user intentions without the constraints of existing systems. This dataset is also made available for sharing with the research community. In contrast, the Freesound query logs encompass approximately 9 million search requests, providing a comprehensive view of real-world usage patterns. Our findings indicate that survey queries are generally longer than Freesound queries, suggesting users prefer detailed queries when not limited by system constraints. Both datasets predominantly feature keyword-based queries, with few survey participants using full sentences. Key factors influencing survey queries include the primary sound source, intended usage, perceived location, and the number of sound sources. These insights are crucial for developing user-centred, effective text-based audio retrieval systems, enhancing our understanding of user behaviour in sound search contexts.
Authors: Yiwei Zhao, Ziyun Li, Win-San Khwa, Xiaoyu Sun, Sai Qian Zhang, Syed Shakib Sarwar, Kleber Hugo Stangherlin, Yi-Lun Lu, Jorge Tomas Gomez, Jae-Sun Seo, Phillip B. Gibbons, Barbara De Salvo, Chiao Liu
Abstract: Low-Latency and Low-Power Edge AI is essential for Virtual Reality and Augmented Reality applications. Recent advances show that hybrid models, combining convolution layers (CNN) and transformers (ViT), often achieve superior accuracy/performance tradeoff on various computer vision and machine learning (ML) tasks. However, hybrid ML models can pose system challenges for latency and energy-efficiency due to their diverse nature in dataflow and memory access patterns. In this work, we leverage the architecture heterogeneity from Neural Processing Units (NPU) and Compute-In-Memory (CIM) and perform diverse execution schemas to efficiently execute these hybrid models. We also introduce H4H-NAS, a Neural Architecture Search framework to design efficient hybrid CNN/ViT models for heterogeneous edge systems with both NPU and CIM. Our H4H-NAS approach is powered by a performance estimator built with NPU performance results measured on real silicon, and CIM performance based on industry IPs. H4H-NAS searches hybrid CNN/ViT models with fine granularity and achieves significant (up to 1.34%) top-1 accuracy improvement on ImageNet dataset. Moreover, results from our Algo/HW co-design reveal up to 56.08% overall latency and 41.72% energy improvements by introducing such heterogeneous computing over baseline solutions. The framework guides the design of hybrid network architectures and system architectures of NPU+CIM heterogeneous systems.
Authors: Konstantina Christakopoulou, Shibl Mourad, Maja Matari\'c
Abstract: Large language models have enabled agents of all kinds to interact with users through natural conversation. Consequently, agents now have two jobs: conversing and planning/reasoning. Their conversational responses must be informed by all available information, and their actions must help to achieve goals. This dichotomy between conversing with the user and doing multi-step reasoning and planning can be seen as analogous to the human systems of "thinking fast and slow" as introduced by Kahneman. Our approach is comprised of a "Talker" agent (System 1) that is fast and intuitive, and tasked with synthesizing the conversational response; and a "Reasoner" agent (System 2) that is slower, more deliberative, and more logical, and is tasked with multi-step reasoning and planning, calling tools, performing actions in the world, and thereby producing the new agent state. We describe the new Talker-Reasoner architecture and discuss its advantages, including modularity and decreased latency. We ground the discussion in the context of a sleep coaching agent, in order to demonstrate real-world relevance.
Authors: Tirthankar Mittra
Abstract: This paper investigates how children learn numbers using the framework of reinforcement learning (RL), with a focus on the impact of language instructions. The motivation for using reinforcement learning stems from its parallels with psychological learning theories in controlled environments. By using state of the art deep reinforcement learning models, we simulate and analyze the effects of various forms of language instructions on number acquisition. Our findings indicate that certain linguistic structures more effectively improve numerical comprehension in RL agents. Additionally, our model predicts optimal sequences for presenting numbers to RL agents which enhance their speed of learning. This research provides valuable insights into the interplay between language and numerical cognition, with implications for both educational strategies and the development of artificial intelligence systems designed to support early childhood learning.
Authors: Priyanka Roy, Susanne Saminger-Platz
Abstract: One of the key problems in learning theory is to compute a function $f$ that closely approximates the relationship between some input $x$ and corresponding output $y$, such that $y\approx f(x)$. This approximation is based on sample points $(x_t,y_t)_{t=1}^{m}$, where the function $f$ can be approximated within reproducing kernel Hilbert spaces using various learning algorithms. In the context of learning theory, it is usually customary to assume that the sample points are drawn independently and identically distributed (i.i.d.) from an unknown underlying distribution. However, we relax this i.i.d. assumption by considering an input sequence $(x_t)_{t\in {\mathbb N}}$ as a trajectory generated by an iterated function system, which forms a particular Markov chain, with $(y_t)_{t\in {\mathbb N}}$ corresponding to an observation sequence when the model is in the corresponding state $x_t$. For such a process, we approximate the function $f$ using the Markov chain stochastic gradient algorithm and estimate the error by deriving upper bounds within reproducing kernel Hilbert spaces.
Authors: Thomas Gauthier-Caron, Shamane Siriwardhana, Elliot Stein, Malikeh Ehghaghi, Charles Goddard, Mark McQuade, Jacob Solawetz, Maxime Labonne
Abstract: By merging models, AI systems can combine the distinct strengths of separate language models, achieving a balance between multiple capabilities without requiring substantial retraining. However, the integration process can be intricate due to differences in training methods and fine-tuning, typically necessitating specialized knowledge and repeated refinement. This paper explores model merging techniques across a spectrum of complexity, examining where automated methods like evolutionary strategies stand compared to hyperparameter-driven approaches such as DARE, TIES-Merging and simpler methods like Model Soups. In addition, we introduce Differentiable Adaptive Merging (DAM), an efficient, adaptive merging approach as an alternative to evolutionary merging that optimizes model integration through scaling coefficients, minimizing computational demands. Our findings reveal that even simple averaging methods, like Model Soups, perform competitively when model similarity is high, underscoring each technique's unique strengths and limitations. We open-sourced DAM, including the implementation code and experiment pipeline, on GitHub: https://github.com/arcee-ai/DAM.
Authors: Andy Zhou, Xiaojun Xu, Ramesh Raghunathan, Alok Lal, Xinze Guan, Bin Yu, Bo Li
Abstract: Graph-based anomaly detection is pivotal in diverse security applications, such as fraud detection in transaction networks and intrusion detection for network traffic. Standard approaches, including Graph Neural Networks (GNNs), often struggle to generalize across shifting data distributions. Meanwhile, real-world domain knowledge is more stable and a common existing component of real-world detection strategies. To explicitly integrate such knowledge into data-driven models such as GCNs, we propose KnowGraph, which integrates domain knowledge with data-driven learning for enhanced graph-based anomaly detection. KnowGraph comprises two principal components: (1) a statistical learning component that utilizes a main model for the overarching detection task, augmented by multiple specialized knowledge models that predict domain-specific semantic entities; (2) a reasoning component that employs probabilistic graphical models to execute logical inferences based on model outputs, encoding domain knowledge through weighted first-order logic formulas. Extensive experiments on these large-scale real-world datasets show that KnowGraph consistently outperforms state-of-the-art baselines in both transductive and inductive settings, achieving substantial gains in average precision when generalizing to completely unseen test graphs. Further ablation studies demonstrate the effectiveness of the proposed reasoning component in improving detection performance, especially under extreme class imbalance. These results highlight the potential of integrating domain knowledge into data-driven models for high-stakes, graph-based security applications.
Authors: Kanan Gupta, Stephan Wojtowytsch
Abstract: While momentum-based optimization algorithms are commonly used in the notoriously non-convex optimization problems of deep learning, their analysis has historically been restricted to the convex and strongly convex setting. In this article, we partially close this gap between theory and practice and demonstrate that virtually identical guarantees can be obtained in optimization problems with a `benign' non-convexity. We show that these weaker geometric assumptions are well justified in overparametrized deep learning, at least locally. Variations of this result are obtained for a continuous time model of Nesterov's accelerated gradient descent algorithm (NAG), the classical discrete time version of NAG, and versions of NAG with stochastic gradient estimates with purely additive noise and with noise that exhibits both additive and multiplicative scaling.
Authors: Tingyu Zhu, Haoyu Liu, Zhimin Jiang, Zeyu Zheng
Abstract: The problem of symbolic music generation presents unique challenges due to the combination of limited data availability and the need for high precision in note pitch. To overcome these difficulties, we introduce Fine-grained Textural Guidance (FTG) within diffusion models to correct errors in the learned distributions. By incorporating FTG, the diffusion models improve the accuracy of music generation, which makes them well-suited for advanced tasks such as progressive music generation, improvisation and interactive music creation. We derive theoretical characterizations for both the challenges in symbolic music generation and the effect of the FTG approach. We provide numerical experiments and a demo page for interactive music generation with user input to showcase the effectiveness of our approach.
Authors: Michael H. Freedman
Abstract: Kolmogorov and Arnold, in answering Hilbert's 13th problem (in the context of continuous functions), laid the foundations for the modern theory of Neural Networks (NNs). Their proof divides the representation of a multivariate function into two steps: The first (non-linear) inter-layer map gives a universal embedding of the data manifold into a single hidden layer whose image is patterned in such a way that a subsequent dynamic can then be defined to solve for the second inter-layer map. I interpret this pattern as "minor concentration" of the almost everywhere defined Jacobians of the interlayer map. Minor concentration amounts to sparsity for higher exterior powers of the Jacobians. We present a conceptual argument for how such sparsity may set the stage for the emergence of successively higher order concepts in today's deep NNs and suggest two classes of experiments to test this hypothesis.
Authors: Yagnik Bandyopadhyay, Harshil Avlani, Houlong L. Zhuang
Abstract: A wide range of deep learning-based machine learning techniques are extensively applied to the design of high-entropy alloys (HEAs), yielding numerous valuable insights. Kolmogorov-Arnold Networks (KAN) is a recently developed architecture that aims to improve both the accuracy and interpretability of input features. In this work, we explore three different datasets for HEA design and demonstrate the application of KAN for both classification and regression models. In the first example, we use a KAN classification model to predict the probability of single-phase formation in high-entropy carbide ceramics based on various properties such as mixing enthalpy and valence electron concentration. In the second example, we employ a KAN regression model to predict the yield strength and ultimate tensile strength of HEAs based on their chemical composition and process conditions including annealing time, cold rolling percentage, and homogenization temperature. The third example involves a KAN classification model to determine whether a certain composition is an HEA or non-HEA, followed by a KAN regressor model to predict the bulk modulus of the identified HEA, aiming to identify HEAs with high bulk modulus. In all three examples, KAN either outperform or match the performance in terms of accuracy such as F1 score for classification and Mean Square Error (MSE), and coefficient of determination (R2) for regression of the multilayer perceptron (MLP) by demonstrating the efficacy of KAN in handling both classification and regression tasks. We provide a promising direction for future research to explore advanced machine learning techniques, which lead to more accurate predictions and better interpretability of complex materials, ultimately accelerating the discovery and optimization of HEAs with desirable properties.
Authors: Yan Li, Mingyi Li, Xiao Zhang, Guangwei Xu, Feng Chen, Yuan Yuan, Yifei Zou, Mengying Zhao, Jianbo Lu, Dongxiao Yu
Abstract: In this work, we study to release the potential of massive heterogeneous weak computing power to collaboratively train large-scale models on dispersed datasets. In order to improve both efficiency and accuracy in resource-adaptive collaborative learning, we take the first step to consider the \textit{unstructured pruning}, \textit{varying submodel architectures}, \textit{knowledge loss}, and \textit{straggler} challenges simultaneously. We propose a novel semi-asynchronous collaborative training framework, namely ${Co\text{-}S}^2{P}$, with data distribution-aware structured pruning and cross-block knowledge transfer mechanism to address the above concerns. Furthermore, we provide theoretical proof that ${Co\text{-}S}^2{P}$ can achieve asymptotic optimal convergence rate of $O(1/\sqrt{N^*EQ})$. Finally, we conduct extensive experiments on a real-world hardware testbed, in which 16 heterogeneous Jetson devices can be united to train large-scale models with parameters up to 0.11 billion. The experimental results demonstrate that $Co\text{-}S^2P$ improves accuracy by up to 8.8\% and resource utilization by up to 1.2$\times$ compared to state-of-the-art methods, while reducing memory consumption by approximately 22\% and training time by about 24\% on all resource-limited devices.
Authors: Krish Parikh
Abstract: Smart vehicles produce large amounts of data, much of which is sensitive and at risk of privacy breaches. As attackers increasingly exploit anonymised metadata within these datasets to profile drivers, it's important to find solutions that mitigate this information leakage without hindering innovation and ongoing research. Synthetic data has emerged as a promising tool to address these privacy concerns, as it allows for the replication of real-world data relationships while minimising the risk of revealing sensitive information. In this paper, we examine the use of synthetic data to tackle these challenges. We start by proposing a comprehensive taxonomy of 14 in-vehicle sensors, identifying potential attacks and categorising their vulnerability. We then focus on the most vulnerable signals, using the Passive Vehicular Sensor (PVS) dataset to generate synthetic data with a Tabular Variational Autoencoder (TVAE) model, which included over 1 million data points. Finally, we evaluate this against 3 core metrics: fidelity, utility, and privacy. Our results show that we achieved 90.1% statistical similarity and 78% classification accuracy when tested on its original intent while also preventing the profiling of the driver. The code can be found at https://github.com/krish-parikh/Synthetic-Data-Generation
URLs: https://github.com/krish-parikh/Synthetic-Data-Generation
Authors: Zhiwei Li, Guodong Long, Jing Jiang, Chengqi Zhang
Abstract: Federated recommendation systems play a crucial role in protecting user privacy. However, existing methods primarily rely on ID-based item embeddings, overlooking the rich multimodal information of items. To address this limitation, we propose a novel Federated Multimodal Recommendation System called FedMR. FedMR leverages a foundation model on the server side to encode multimodal data, such as images and text, associated with items. To tackle the challenge of data heterogeneity caused by varying user preferences, FedMR introduces a Mixing Feature Fusion Module on the client. This module dynamically adjusts the weights of different fusion strategies based on user interaction history, generating personalized item embeddings that capture fine-grained user preferences. FedMR is compatible with existing ID-based federated recommendation systems, improving their performances without modifying the original framework. Our experiments on four real-world multimodal recommendation datasets demonstrate the effectiveness of FedMR. Our code is available at https://anonymous.4open.science/r/FedMR.
Authors: Duraimurugan Rajamanickam
Abstract: Legal Entity Recognition (LER) is critical in automating legal workflows such as contract analysis, compliance monitoring, and litigation support. Existing approaches, including rule-based systems and classical machine learning models, struggle with the complexity of legal documents and domain specificity, particularly in handling ambiguities and nested entity structures. This paper proposes a novel hybrid model that enhances the accuracy and precision of Legal-BERT, a transformer model fine-tuned for legal text processing, by introducing a semantic similarity-based filtering mechanism. We evaluate the model on a dataset of 15,000 annotated legal documents, achieving an F1 score of 93.4%, demonstrating significant improvements in precision and recall over previous methods.
Authors: Yangyi Chen, Binxuan Huang, Yifan Gao, Zhengyang Wang, Jingfeng Yang, Heng Ji
Abstract: Precise estimation of downstream performance in large language models (LLMs) prior to training is essential for guiding their development process. Scaling laws analysis utilizes the statistics of a series of significantly smaller sampling language models (LMs) to predict the performance of the target LLM. For downstream performance prediction, the critical challenge lies in the emergent abilities in LLMs that occur beyond task-specific computational thresholds. In this work, we focus on the pre-training loss as a more computation-efficient metric for performance estimation. Our two-stage approach consists of first estimating a function that maps computational resources (e.g., FLOPs) to the pre-training Loss using a series of sampling models, followed by mapping the pre-training loss to downstream task Performance after the critical "emergent phase". In preliminary experiments, this FLP solution accurately predicts the performance of LLMs with 7B and 13B parameters using a series of sampling LMs up to 3B, achieving error margins of 5% and 10%, respectively, and significantly outperforming the FLOPs-to-Performance approach. This motivates FLP-M, a fundamental approach for performance prediction that addresses the practical need to integrate datasets from multiple sources during pre-training, specifically blending general corpora with code data to accurately represent the common necessity. FLP-M extends the power law analytical function to predict domain-specific pre-training loss based on FLOPs across data sources, and employs a two-layer neural network to model the non-linear relationship between multiple domain-specific loss and downstream performance. By utilizing a 3B LLM trained on a specific ratio and a series of smaller sampling LMs, FLP-M can effectively forecast the performance of 3B and 7B LLMs across various data mixtures for most benchmarks within 10% error margins.
Authors: Akihiro Fujii, Yoshitaka Ushiku, Koji Shimizu, Anh Khoa Augustin Lu, Satoshi Watanabe
Abstract: In materials science, finding crystal structures that have targeted properties is crucial. While recent methodologies such as Bayesian optimization and deep generative models have made some advances on this issue, these methods often face difficulties in adaptively incorporating various constraints, such as electrical neutrality and targeted properties optimization, while keeping the desired specific crystal structure. To address these challenges, we have developed the Simultaneous Multi-property Optimization using Adaptive Crystal Synthesizer (SMOACS), which utilizes state-of-the-art property prediction models and their gradients to directly optimize input crystal structures for targeted properties simultaneously. SMOACS enables the integration of adaptive constraints into the optimization process without necessitating model retraining. Thanks to this feature, SMOACS has succeeded in simultaneously optimizing targeted properties while maintaining perovskite structures, even with models trained on diverse crystal types. We have demonstrated the band gap optimization while meeting a challenging constraint, that is, maintaining electrical neutrality in large atomic configurations up to 135 atom sites, where the verification of the electrical neutrality is challenging. The properties of the most promising materials have been confirmed by density functional theory calculations.
Authors: Ho-Chien Huang, Chao-Lin Liu
Abstract: In the Taiwanese judicial system, Cause of Actions (COAs) are essential for identifying relevant legal judgments. However, the lack of standardized COA labeling creates challenges in filtering cases using basic methods. This research addresses this issue by leveraging embedding and clustering techniques to analyze the similarity between COAs based on cited legal articles. The study implements various similarity measures, including Dice coefficient and Pearson's correlation coefficient. An ensemble model combines rankings, and social network analysis identifies clusters of related COAs. This approach enhances legal analysis by revealing inconspicuous connections between COAs, offering potential applications in legal research beyond civil law.
Authors: Peng Jiang, Kun Wang, Jiaxing Wang, Zeliang Feng, Shengjie Qiao, Runhuai Deng, Fengkai Zhang
Abstract: GPR full-waveform inversion optimizes the subsurface property model iteratively to match the entire waveform information. However, the model gradients derived from wavefield continuation often contain errors, such as ghost values and excessively large values at transmitter and receiver points. Furthermore, models updated based on these gradients frequently exhibit unclear characterization of anomalous bodies or false anomalies, making it challenging to obtain accurate inversion results. To address these issues, we introduced a novel full-waveform inversion (FWI) framework that incorporates an embedded convolutional neural network (CNN) to adaptively filter model parameters and gradients. Specifically, we embedded the CNN module before the forward modeling process and ensured the entire FWI process remains differentiable. This design leverages the auto-grad tool of the deep learning library, allowing model values to pass through the CNN module during forward computation and model gradients to pass through the CNN module during backpropagation. Experiments have shown that filtering the model parameters during forward computation and the model gradients during backpropagation can ultimately yield high-quality inversion results.
Authors: Joris Guerin, Shray Bansal, Amirreza Shaban, Paulo Mann, Harshvardhan Gazula
Abstract: This work tackles the challenge of efficiently selecting high-performance pre-trained vision backbones for specific target tasks. Although exhaustive search within a finite set of backbones can solve this problem, it becomes impractical for large datasets and backbone pools. To address this, we introduce Vision Backbone Efficient Selection (VIBES), which aims to quickly find well-suited backbones, potentially trading off optimality for efficiency. We propose several simple yet effective heuristics to address VIBES and evaluate them across four diverse computer vision datasets. Our results show that these approaches can identify backbones that outperform those selected from generic benchmarks, even within a limited search budget of one hour on a single GPU. We reckon VIBES marks a paradigm shift from benchmarks to task-specific optimization.
Authors: Shojiro Yamabe, Tsubasa Takahashi, Futa Waseda, Koki Wataoka
Abstract: As the cost of training large language models (LLMs) rises, protecting their intellectual property has become increasingly critical. Model merging, which integrates multiple expert models into a single model capable of performing multiple tasks, presents a growing risk of unauthorized and malicious usage. While fingerprinting techniques have been studied for asserting model ownership, existing methods have primarily focused on fine-tuning, leaving model merging underexplored. To address this gap, we propose a novel fingerprinting method MergePrint that embeds robust fingerprints designed to preserve ownership claims even after model merging. By optimizing against a pseudo-merged model, which simulates post-merged model weights, MergePrint generates fingerprints that remain detectable after merging. Additionally, we optimize the fingerprint inputs to minimize performance degradation, enabling verification through specific outputs from targeted inputs. This approach provides a practical fingerprinting strategy for asserting ownership in cases of misappropriation through model merging.
Authors: Mehrshad Momen-Tayefeh
Abstract: Generating realistic images from human texts is one of the most challenging problems in the field of computer vision (CV). The meaning of descriptions given can be roughly reflected by existing text-to-image approaches. In this paper, our main purpose is to propose a brief comparison between five different methods base on the Generative Adversarial Networks (GAN) to make image from the text. In addition, each model architectures synthesis images with different resolution. Furthermore, the best and worst obtained resolutions is 64*64, 256*256 respectively. However, we checked and compared some metrics that introduce the accuracy of each model. Also, by doing this study, we found out the best model for this problem by comparing these different approaches essential metrics.
Authors: Purushothaman Natarajan, Kamal Basha, Athira Nambiar
Abstract: Sonar image synthesis is crucial for advancing applications in underwater exploration, marine biology, and defence. Traditional methods often rely on extensive and costly data collection using sonar sensors, jeopardizing data quality and diversity. To overcome these limitations, this study proposes a new sonar image synthesis framework, Synth-SONAR leveraging diffusion models and GPT prompting. The key novelties of Synth-SONAR are threefold: First, by integrating Generative AI-based style injection techniques along with publicly available real/simulated data, thereby producing one of the largest sonar data corpus for sonar research. Second, a dual text-conditioning sonar diffusion model hierarchy synthesizes coarse and fine-grained sonar images with enhanced quality and diversity. Third, high-level (coarse) and low-level (detailed) text-based sonar generation methods leverage advanced semantic information available in visual language models (VLMs) and GPT-prompting. During inference, the method generates diverse and realistic sonar images from textual prompts, bridging the gap between textual descriptions and sonar image generation. This marks the application of GPT-prompting in sonar imagery for the first time, to the best of our knowledge. Synth-SONAR achieves state-of-the-art results in producing high-quality synthetic sonar datasets, significantly enhancing their diversity and realism.
Authors: Yi Zhou, Yilai Li, Jing Yuan, Quanquan Gu
Abstract: Cryo-electron microscopy (cryo-EM) is a powerful technique in structural biology and drug discovery, enabling the study of biomolecules at high resolution. Significant advancements by structural biologists using cryo-EM have led to the production of over 38,626 protein density maps at various resolutions1. However, cryo-EM data processing algorithms have yet to fully benefit from our knowledge of biomolecular density maps, with only a few recent models being data-driven but limited to specific tasks. In this study, we present CryoFM, a foundation model designed as a generative model, learning the distribution of high-quality density maps and generalizing effectively to downstream tasks. Built on flow matching, CryoFM is trained to accurately capture the prior distribution of biomolecular density maps. Furthermore, we introduce a flow posterior sampling method that leverages CRYOFM as a flexible prior for several downstream tasks in cryo-EM and cryo-electron tomography (cryo-ET) without the need for fine-tuning, achieving state-of-the-art performance on most tasks and demonstrating its potential as a foundational model for broader applications in these fields.
Authors: Unai Ruiz-Gonzalez, Alain Andres, Pedro G. Bascoy, Javier Del Ser
Abstract: Sparse reward environments in reinforcement learning (RL) pose significant challenges for exploration, often leading to inefficient or incomplete learning processes. To tackle this issue, this work proposes a teacher-student RL framework that leverages Large Language Models (LLMs) as "teachers" to guide the agent's learning process by decomposing complex tasks into subgoals. Due to their inherent capability to understand RL environments based on a textual description of structure and purpose, LLMs can provide subgoals to accomplish the task defined for the environment in a similar fashion to how a human would do. In doing so, three types of subgoals are proposed: positional targets relative to the agent, object representations, and language-based instructions generated directly by the LLM. More importantly, we show that it is possible to query the LLM only during the training phase, enabling agents to operate within the environment without any LLM intervention. We assess the performance of this proposed framework by evaluating three state-of-the-art open-source LLMs (Llama, DeepSeek, Qwen) eliciting subgoals across various procedurally generated environment of the MiniGrid benchmark. Experimental results demonstrate that this curriculum-based approach accelerates learning and enhances exploration in complex tasks, achieving up to 30 to 200 times faster convergence in training steps compared to recent baselines designed for sparse reward environments.
Authors: Toby Dylan Hocking, Gabrielle Thibault, Cameron Scott Bodine, Paul Nelson Arellano, Alexander F Shenkin, Olivia Jasmine Lindly
Abstract: In many real-world applications of machine learning, we are interested to know if it is possible to train on the data that we have gathered so far, and obtain accurate predictions on a new test data subset that is qualitatively different in some respect (time period, geographic region, etc). Another question is whether data subsets are similar enough so that it is beneficial to combine subsets during model training. We propose SOAK, Same/Other/All K-fold cross-validation, a new method which can be used to answer both questions. SOAK systematically compares models which are trained on different subsets of data, and then used for prediction on a fixed test subset, to estimate the similarity of learnable/predictable patterns in data subsets. We show results of using SOAK on six new real data sets (with geographic/temporal subsets, to check if predictions are accurate on new subsets), 3 image pair data sets (subsets are different image types, to check that we get smaller prediction error on similar images), and 11 benchmark data sets with predefined train/test splits (to check similarity of predefined splits).
Authors: Changhun Lee, Jun-gyu Jin, Younghyun Cho, Eunhyeok Park
Abstract: With the rapid growth in the use of fine-tuning for large language models (LLMs), optimizing fine-tuning while keeping inference efficient has become highly important. However, this is a challenging task as it requires improvements in all aspects, including inference speed, fine-tuning speed, memory consumption, and, most importantly, model quality. Previous studies have attempted to achieve this by combining quantization with fine-tuning, but they have failed to enhance all four aspects simultaneously. In this study, we propose a new lightweight technique called Quantization for Efficient Fine-Tuning (QEFT). QEFT accelerates both inference and fine-tuning, is supported by robust theoretical foundations, offers high flexibility, and maintains good hardware compatibility. Our extensive experiments demonstrate that QEFT matches the quality and versatility of full-precision parameter-efficient fine-tuning, while using fewer resources. Our code is available at https://github.com/xvyaward/qeft.
Authors: Beatrice Achilli, Enrico Ventura, Gianluigi Silvestri, Bao Pham, Gabriel Raya, Dmitry Krotov, Carlo Lucibello, Luca Ambrogioni
Abstract: Generative diffusion processes are state-of-the-art machine learning models deeply connected with fundamental concepts in statistical physics. Depending on the dataset size and the capacity of the network, their behavior is known to transition from an associative memory regime to a generalization phase in a phenomenon that has been described as a glassy phase transition. Here, using statistical physics techniques, we extend the theory of memorization in generative diffusion to manifold-supported data. Our theoretical and experimental findings indicate that different tangent subspaces are lost due to memorization effects at different critical times and dataset sizes, which depend on the local variance of the data along their directions. Perhaps counterintuitively, we find that, under some conditions, subspaces of higher variance are lost first due to memorization effects. This leads to a selective loss of dimensionality where some prominent features of the data are memorized without a full collapse on any individual training point. We validate our theory with a comprehensive set of experiments on networks trained both in image datasets and on linear manifolds, which result in a remarkable qualitative agreement with the theoretical predictions.
Authors: Beichen Wang, Juexiao Zhang, Shuwen Dong, Irving Fang, Chen Feng
Abstract: Vision Language Models (VLMs) have recently been adopted in robotics for their capability in common sense reasoning and generalizability. Existing work has applied VLMs to generate task and motion planning from natural language instructions and simulate training data for robot learning. In this work, we explore using VLM to interpret human demonstration videos and generate robot task planning. Our method integrates keyframe selection, visual perception, and VLM reasoning into a pipeline. We named it SeeDo because it enables the VLM to ''see'' human demonstrations and explain the corresponding plans to the robot for it to ''do''. To validate our approach, we collected a set of long-horizon human videos demonstrating pick-and-place tasks in three diverse categories and designed a set of metrics to comprehensively benchmark SeeDo against several baselines, including state-of-the-art video-input VLMs. The experiments demonstrate SeeDo's superior performance. We further deployed the generated task plans in both a simulation environment and on a real robot arm.
Authors: Disha Hegde, Mohamed Adil, Jon Cockayne
Abstract: Gaussian processes are notorious for scaling cubically with the size of the training set, preventing application to very large regression problems. Computation-aware Gaussian processes (CAGPs) tackle this scaling issue by exploiting probabilistic linear solvers to reduce complexity, widening the posterior with additional computational uncertainty due to reduced computation. However, the most commonly used CAGP framework results in (sometimes dramatically) conservative uncertainty quantification, making the posterior unrealistic in practice. In this work, we prove that if the utilised probabilistic linear solver is calibrated, in a rigorous statistical sense, then so too is the induced CAGP. We thus propose a new CAGP framework, CAGP-GS, based on using Gauss-Seidel iterations for the underlying probabilistic linear solver. CAGP-GS performs favourably compared to existing approaches when the test set is low-dimensional and few iterations are performed. We test the calibratedness on a synthetic problem, and compare the performance to existing approaches on a large-scale global temperature regression problem.
Authors: Alessandro Bombini, Fernando Garc\'ia-Avello Bof\'ias, Francesca Giambi, Chiara Ruberto
Abstract: In this contribution, we define (and test) a pipeline to perform virtual painting recolouring using raw data of X-Ray Fluorescence (XRF) analysis on pictorial artworks. To circumvent the small dataset size, we generate a synthetic dataset, starting from a database of XRF spectra; furthermore, to ensure a better generalisation capacity (and to tackle the issue of in-memory size and inference time), we define a Deep Variational Embedding network to embed the XRF spectra into a lower dimensional, K-Means friendly, metric space. We thus train a set of models to assign coloured images to embedded XRF images. We report here the devised pipeline performances in terms of visual quality metrics, and we close on a discussion on the results.
Authors: Lorenzo Magnino, Yuchen Zhu, Mathieu Lauri\`ere
Abstract: Optimal stopping is a fundamental problem in optimization that has found applications in risk management, finance, economics, and recently in the fields of computer science. We extend the standard framework to a multi-agent setting, named multi-agent optimal stopping (MAOS), where a group of agents cooperatively solves finite-space, discrete-time optimal stopping problems. Solving the finite-agent case is computationally prohibitive when the number of agents is very large, so this work studies the mean field optimal stopping (MFOS) problem, obtained as the number of agents approaches infinity. We prove that MFOS provides a good approximate solution to MAOS. We also prove a dynamic programming principle (DPP), based on the theory of mean field control. We then propose two deep learning methods: one simulates full trajectories to learn optimal decisions, whereas the other leverages DPP with backward induction; both methods train neural networks for the optimal stopping decisions. We demonstrate the effectiveness of these approaches through numerical experiments on 6 different problems in spatial dimension up to 300. To the best of our knowledge, this is the first work to study MFOS in finite space and discrete time, and to propose efficient and scalable computational methods for this type of problem.
Authors: Michelle Zhao, Reid Simmons, Henny Admoni, Aaditya Ramdas, Andrea Bajcsy
Abstract: In interactive imitation learning (IL), uncertainty quantification offers a way for the learner (i.e. robot) to contend with distribution shifts encountered during deployment by actively seeking additional feedback from an expert (i.e. human) online. Prior works use mechanisms like ensemble disagreement or Monte Carlo dropout to quantify when black-box IL policies are uncertain; however, these approaches can lead to overconfident estimates when faced with deployment-time distribution shifts. Instead, we contend that we need uncertainty quantification algorithms that can leverage the expert human feedback received during deployment time to adapt the robot's uncertainty online. To tackle this, we draw upon online conformal prediction, a distribution-free method for constructing prediction intervals online given a stream of ground-truth labels. Human labels, however, are intermittent in the interactive IL setting. Thus, from the conformal prediction side, we introduce a novel uncertainty quantification algorithm called intermittent quantile tracking (IQT) that leverages a probabilistic model of intermittent labels, maintains asymptotic coverage guarantees, and empirically achieves desired coverage levels. From the interactive IL side, we develop ConformalDAgger, a new approach wherein the robot uses prediction intervals calibrated by IQT as a reliable measure of deployment-time uncertainty to actively query for more expert feedback. We compare ConformalDAgger to prior uncertainty-aware DAgger methods in scenarios where the distribution shift is (and isn't) present because of changes in the expert's policy. We find that in simulated and hardware deployments on a 7DOF robotic manipulator, ConformalDAgger detects high uncertainty when the expert shifts and increases the number of interventions compared to baselines, allowing the robot to more quickly learn the new behavior.
Authors: Klaudia Thellmann, Bernhard Stadler, Michael Fromm, Jasper Schulze Buschhoff, Alex Jude, Fabio Barth, Johannes Leveling, Nicolas Flores-Herr, Joachim K\"ohler, Ren\'e J\"akel, Mehdi Ali
Abstract: The rise of Large Language Models (LLMs) has revolutionized natural language processing across numerous languages and tasks. However, evaluating LLM performance in a consistent and meaningful way across multiple European languages remains challenging, especially due to the scarcity of multilingual benchmarks. We introduce a cross-lingual evaluation approach tailored for European languages. We employ translated versions of five widely-used benchmarks to assess the capabilities of 40 LLMs across 21 European languages. Our contributions include examining the effectiveness of translated benchmarks, assessing the impact of different translation services, and offering a multilingual evaluation framework for LLMs that includes newly created datasets: EU20-MMLU, EU20-HellaSwag, EU20-ARC, EU20-TruthfulQA, and EU20-GSM8K. The benchmarks and results are made publicly available to encourage further research in multilingual LLM evaluation.
Authors: Xin Yu, Zelin He, Ying Sun, Lingzhou Xue, Runze Li
Abstract: FedProx is a simple yet effective federated learning method that enables model personalization via regularization. Despite remarkable success in practice, a rigorous analysis of how such a regularization provably improves the statistical accuracy of each client's local model hasn't been fully established. Setting the regularization strength heuristically presents a risk, as an inappropriate choice may even degrade accuracy. This work fills in the gap by analyzing the effect of regularization on statistical accuracy, thereby providing a theoretical guideline for setting the regularization strength for achieving personalization. We prove that by adaptively choosing the regularization strength under different statistical heterogeneity, FedProx can consistently outperform pure local training and achieve a nearly minimax-optimal statistical rate. In addition, to shed light on resource allocation, we design an algorithm, provably showing that stronger personalization reduces communication complexity without increasing the computation cost overhead. Finally, our theory is validated on both synthetic and real-world datasets and its generalizability is verified in a non-convex setting.
Authors: Benson Chen, Tomasz Danel, Patrick J. McEnaney, Nikhil Jain, Kirill Novikov, Spurti Umesh Akki, Joshua L. Turnbull, Virja Atul Pandya, Boris P. Belotserkovskii, Jared Bryce Weaver, Ankita Biswas, Dat Nguyen, Gabriel H. S. Dreiman, Mohammad Sultan, Nathaniel Stanley, Daniel M Whalen, Divya Kanichar, Christoph Klein, Emily Fox, R. Edward Watts
Abstract: DNA-Encoded Libraries (DEL) are combinatorial small molecule libraries that offer an efficient way to characterize diverse chemical spaces. Selection experiments using DELs are pivotal to drug discovery efforts, enabling high-throughput screens for hit finding. However, limited availability of public DEL datasets hinders the advancement of computational techniques designed to process such data. To bridge this gap, we present KinDEL, one of the first large, publicly available DEL datasets on two kinases: Mitogen-Activated Protein Kinase 14 (MAPK14) and Discoidin Domain Receptor Tyrosine Kinase 1 (DDR1). Interest in this data modality is growing due to its ability to generate extensive supervised chemical data that densely samples around select molecular structures. Demonstrating one such application of the data, we benchmark different machine learning techniques to develop predictive models for hit identification; in particular, we highlight recent structure-based probabilistic approaches. Finally, we provide biophysical assay data, both on- and off-DNA, to validate our models on a smaller subset of molecules. Data and code for our benchmarks can be found at: https://github.com/insitro/kindel.
Authors: Brighton Ancelin, Alex Saad-Falcon, Kason Ancelin, Justin Romberg
Abstract: We propose new algorithms to efficiently average a collection of points on a Grassmannian manifold in both the centralized and decentralized settings. Grassmannian points are used ubiquitously in machine learning, computer vision, and signal processing to represent data through (often low-dimensional) subspaces. While averaging these points is crucial to many tasks (especially in the decentralized setting), existing methods unfortunately remain computationally expensive due to the non-Euclidean geometry of the manifold. Our proposed algorithms, Rapid Grassmannian Averaging (RGrAv) and Decentralized Rapid Grassmannian Averaging (DRGrAv), overcome this challenge by leveraging the spectral structure of the problem to rapidly compute an average using only small matrix multiplications and QR factorizations. We provide a theoretical guarantee of optimality and present numerical experiments which demonstrate that our algorithms outperform state-of-the-art methods in providing high accuracy solutions in minimal time. Additional experiments showcase the versatility of our algorithms to tasks such as K-means clustering on video motion data, establishing RGrAv and DRGrAv as powerful tools for generic Grassmannian averaging.
Authors: Daniel Salnikov, Kevin Michalewicz, Dan Leonte
Abstract: We propose the $\textit{lifted linear model}$, and derive model-free prediction intervals that become tighter as the correlation between predictions and observations increases. These intervals motivate the $\textit{Lifted Coefficient of Determination}$, a model comparison criterion for arbitrary loss functions in prediction-based settings, e.g., regression, classification or counts. We extend the prediction intervals to more general error distributions, and propose a fast model-free outlier detection algorithm for regression. Finally, we illustrate the framework via numerical experiments.
Authors: Baptiste Ab\'el\`es, Eugenio Clerico, Gergely Neu
Abstract: Traditional generalization results in statistical learning require a training data set made of independently drawn examples. Most of the recent efforts to relax this independence assumption have considered either purely temporal (mixing) dependencies, or graph-dependencies, where non-adjacent vertices correspond to independent random variables. Both approaches have their own limitations, the former requiring a temporal ordered structure, and the latter lacking a way to quantify the strength of inter-dependencies. In this work, we bridge these two lines of work by proposing a framework where dependencies decay with graph distance. We derive generalization bounds leveraging the online-to-PAC framework, by deriving a concentration result and introducing an online learning framework incorporating the graph structure. The resulting high-probability generalization guarantees depend on both the mixing rate and the graph's chromatic number.
Authors: Jiaxu Wang, Jingkai Sun, Junhao He, Ziyi Zhang, Qiang Zhang, Mingyuan Sun, Renjing Xu
Abstract: Learning-based simulators show great potential for simulating particle dynamics when 3D groundtruth is available, but per-particle correspondences are not always accessible. The development of neural rendering presents a new solution to this field to learn 3D dynamics from 2D images by inverse rendering. However, existing approaches still suffer from ill-posed natures resulting from the 2D to 3D uncertainty, for example, specific 2D images can correspond with various 3D particle distributions. To mitigate such uncertainty, we consider a conventional, mechanically interpretable framework as the physical priors and extend it to a learning-based version. In brief, we incorporate the learnable graph kernels into the classic Discrete Element Analysis (DEA) framework to implement a novel mechanics-integrated learning system. In this case, the graph network kernels are only used for approximating some specific mechanical operators in the DEA framework rather than the whole dynamics mapping. By integrating the strong physics priors, our methods can effectively learn the dynamics of various materials from the partial 2D observations in a unified manner. Experiments show that our approach outperforms other learned simulators by a large margin in this context and is robust to different renderers, fewer training samples, and fewer camera views.
Authors: Rebecca M. M. Hicke, Ross Deans Kristensen-McLachlan
Abstract: Metaphors are everywhere. They appear extensively across all domains of natural language, from the most sophisticated poetry to seemingly dry academic prose. A significant body of research in the cognitive science of language argues for the existence of conceptual metaphors, the systematic structuring of one domain of experience in the language of another. Conceptual metaphors are not simply rhetorical flourishes but are crucial evidence of the role of analogical reasoning in human cognition. In this paper, we ask whether Large Language Models (LLMs) can accurately identify and explain the presence of such conceptual metaphors in natural language data. Using a novel prompting technique based on metaphor annotation guidelines, we demonstrate that LLMs are a promising tool for large-scale computational research on conceptual metaphors. Further, we show that LLMs are able to apply procedural guidelines designed for human annotators, displaying a surprising depth of linguistic knowledge.
Authors: Yan Chen, Jose Blanchet, Krzysztof Dembczynski, Laura Fee Nern, Aaron Flores
Abstract: Downsampling or under-sampling is a technique that is utilized in the context of large and highly imbalanced classification models. We study optimal downsampling for imbalanced classification using generalized linear models (GLMs). We propose a pseudo maximum likelihood estimator and study its asymptotic normality in the context of increasingly imbalanced populations relative to an increasingly large sample size. We provide theoretical guarantees for the introduced estimator. Additionally, we compute the optimal downsampling rate using a criterion that balances statistical accuracy and computational efficiency. Our numerical experiments, conducted on both synthetic and empirical data, further validate our theoretical results, and demonstrate that the introduced estimator outperforms commonly available alternatives.
Authors: Roman Worschech, Bernd Rosenow
Abstract: Neural scaling laws describe how the performance of deep neural networks scales with key factors such as training data size, model complexity, and training time, often following power-law behaviors over multiple orders of magnitude. Despite their empirical observation, the theoretical understanding of these scaling laws remains limited. In this work, we employ techniques from statistical mechanics to analyze one-pass stochastic gradient descent within a student-teacher framework, where both the student and teacher are two-layer neural networks. Our study primarily focuses on the generalization error and its behavior in response to data covariance matrices that exhibit power-law spectra. For linear activation functions, we derive analytical expressions for the generalization error, exploring different learning regimes and identifying conditions under which power-law scaling emerges. Additionally, we extend our analysis to non-linear activation functions in the feature learning regime, investigating how power-law spectra in the data covariance matrix impact learning dynamics. Importantly, we find that the length of the symmetric plateau depends on the number of distinct eigenvalues of the data covariance matrix and the number of hidden units, demonstrating how these plateaus behave under various configurations. In addition, our results reveal a transition from exponential to power-law convergence in the specialized phase when the data covariance matrix possesses a power-law spectrum. This work contributes to the theoretical understanding of neural scaling laws and provides insights into optimizing learning performance in practical scenarios involving complex data structures.
Authors: Zhexiao Lin, Pablo Crespo
Abstract: Online controlled experiments (A/B testing) are essential in data-driven decision-making for many companies. Increasing the sensitivity of these experiments, particularly with a fixed sample size, relies on reducing the variance of the estimator for the average treatment effect (ATE). Existing methods like CUPED and CUPAC use pre-experiment data to reduce variance, but their effectiveness depends on the correlation between the pre-experiment data and the outcome. In contrast, in-experiment data is often more strongly correlated with the outcome and thus more informative. In this paper, we introduce a novel method that combines both pre-experiment and in-experiment data to achieve greater variance reduction than CUPED and CUPAC, without introducing bias or additional computation complexity. We also establish asymptotic theory and provide consistent variance estimators for our method. Applying this method to multiple online experiments at Etsy, we reach substantial variance reduction over CUPAC with the inclusion of only a few in-experiment covariates. These results highlight the potential of our approach to significantly improve experiment sensitivity and accelerate decision-making.
Authors: Pratinav Seth, Michelle Lin, Brefo Dwamena Yaw, Jade Boutot, Mary Kang, David Rolnick
Abstract: Millions of abandoned oil and gas wells are scattered across the world, leaching methane into the atmosphere and toxic compounds into the groundwater. Many of these locations are unknown, preventing the wells from being plugged and their polluting effects averted. Remote sensing is a relatively unexplored tool for pinpointing abandoned wells at scale. We introduce the first large-scale benchmark dataset for this problem, leveraging medium-resolution multi-spectral satellite imagery from Planet Labs. Our curated dataset comprises over 213,000 wells (abandoned, suspended, and active) from Alberta, a region with especially high well density, sourced from the Alberta Energy Regulator and verified by domain experts. We evaluate baseline algorithms for well detection and segmentation, showing the promise of computer vision approaches but also significant room for improvement.
Authors: Peter Potaptchik, Iskander Azangulov, George Deligiannidis
Abstract: Score-matching generative models have proven successful at sampling from complex high-dimensional data distributions. In many applications, this distribution is believed to concentrate on a much lower $d$-dimensional manifold embedded into $D$-dimensional space; this is known as the manifold hypothesis. The current best-known convergence guarantees are either linear in $D$ or polynomial (superlinear) in $d$. The latter exploits a novel integration scheme for the backward SDE. We take the best of both worlds and show that the number of steps diffusion models require in order to converge in Kullback-Leibler~(KL) divergence is linear (up to logarithmic terms) in the intrinsic dimension $d$. Moreover, we show that this linear dependency is sharp.
Authors: Qin Liu, Chao Shang, Ling Liu, Nikolaos Pappas, Jie Ma, Neha Anna John, Srikanth Doss, Lluis Marquez, Miguel Ballesteros, Yassine Benajiba
Abstract: The safety alignment ability of Vision-Language Models (VLMs) is prone to be degraded by the integration of the vision module compared to its LLM backbone. We investigate this phenomenon, dubbed as ''safety alignment degradation'' in this paper, and show that the challenge arises from the representation gap that emerges when introducing vision modality to VLMs. In particular, we show that the representations of multi-modal inputs shift away from that of text-only inputs which represent the distribution that the LLM backbone is optimized for. At the same time, the safety alignment capabilities, initially developed within the textual embedding space, do not successfully transfer to this new multi-modal representation space. To reduce safety alignment degradation, we introduce Cross-Modality Representation Manipulation (CMRM), an inference time representation intervention method for recovering the safety alignment ability that is inherent in the LLM backbone of VLMs, while simultaneously preserving the functional capabilities of VLMs. The empirical results show that our framework significantly recovers the alignment ability that is inherited from the LLM backbone with minimal impact on the fluency and linguistic capabilities of pre-trained VLMs even without additional training. Specifically, the unsafe rate of LLaVA-7B on multi-modal input can be reduced from 61.53% to as low as 3.15% with only inference-time intervention. WARNING: This paper contains examples of toxic or harmful language.
Authors: Jongseong Jang, Daeun Kyung, Seung Hwan Kim, Honglak Lee, Kyunghoon Bae, Edward Choi
Abstract: Deep neural networks are increasingly used in medical imaging for tasks such as pathological classification, but they face challenges due to the scarcity of high-quality, expert-labeled training data. Recent efforts have utilized pre-trained contrastive image-text models like CLIP, adapting them for medical use by fine-tuning the model with chest X-ray images and corresponding reports for zero-shot pathology classification, thus eliminating the need for pathology-specific annotations. However, most studies continue to use the same contrastive learning objectives as in the general domain, overlooking the multi-labeled nature of medical image-report pairs. In this paper, we propose a new fine-tuning strategy that includes positive-pair loss relaxation and random sentence sampling. We aim to improve the performance of zero-shot pathology classification without relying on external knowledge. Our method can be applied to any pre-trained contrastive image-text encoder and easily transferred to out-of-domain datasets without further training, as it does not use external data. Our approach consistently improves overall zero-shot pathology classification across four chest X-ray datasets and three pre-trained models, with an average macro AUROC increase of 4.3%. Additionally, our method outperforms the state-of-the-art and marginally surpasses board-certified radiologists in zero-shot classification for the five competition pathologies in the CheXpert dataset.
Authors: Ahmad Faraz Khan, Xinran Wang, Qi Le, Zain ul Abdeen, Azal Ahmad Khan, Haider Ali, Ming Jin, Jie Ding, Ali R. Butt, Ali Anwar
Abstract: Existing incentive solutions for traditional Federated Learning (FL) focus on individual contributions to a single global objective, neglecting the nuances of clustered personalization with multiple cluster-level models and the non-monetary incentives such as personalized model appeal for clients. In this paper, we first propose to treat incentivization and personalization as interrelated challenges and solve them with an incentive mechanism that fosters personalized learning. Additionally, current methods depend on an aggregator for client clustering, which is limited by a lack of access to clients' confidential information due to privacy constraints, leading to inaccurate clustering. To overcome this, we propose direct client involvement, allowing clients to indicate their cluster membership preferences based on data distribution and incentive-driven feedback. Our approach enhances the personalized model appeal for self-aware clients with high-quality data leading to their active and consistent participation. Our evaluation demonstrates significant improvements in test accuracy (8-45%), personalized model appeal (3-38%), and participation rates (31-100%) over existing FL models, including those addressing data heterogeneity and personalization.
Authors: Juyoung Yun, Sol Choi, Francois Rameau, Byungkon Kang, Zhoulai Fu
Abstract: With the increasing complexity of machine learning models, managing computational resources like memory and processing power has become a critical concern. Mixed precision techniques, which leverage different numerical precisions during model training and inference to optimize resource usage, have been widely adopted. However, access to hardware that supports lower precision formats (e.g., FP8 or FP4) remains limited, especially for practitioners with hardware constraints. For many with limited resources, the available options are restricted to using 32-bit, 16-bit, or a combination of the two. While it is commonly believed that 16-bit precision can achieve results comparable to full (32-bit) precision, this study is the first to systematically validate this assumption through both rigorous theoretical analysis and extensive empirical evaluation. Our theoretical formalization of floating-point errors and classification tolerance provides new insights into the conditions under which 16-bit precision can approximate 32-bit results. This study fills a critical gap, proving for the first time that standalone 16-bit precision neural networks match 32-bit and mixed-precision in accuracy while boosting computational speed. Given the widespread availability of 16-bit across GPUs, these findings are especially valuable for machine learning practitioners with limited hardware resources to make informed decisions.
Authors: Aldo Gael Carranza, Susan Athey
Abstract: We consider the problem of learning personalized decision policies from observational bandit feedback data across multiple heterogeneous data sources. In our approach, we introduce a novel regret analysis that establishes finite-sample upper bounds on distinguishing notions of global regret for all data sources on aggregate and of local regret for any given data source. We characterize these regret bounds by expressions of source heterogeneity and distribution shift. Moreover, we examine the practical considerations of this problem in the federated setting where a central server aims to train a policy on data distributed across the heterogeneous sources without collecting any of their raw data. We present a policy learning algorithm amenable to federation based on the aggregation of local policies trained with doubly robust offline policy evaluation strategies. Our analysis and supporting experimental results provide insights into tradeoffs in the participation of heterogeneous data sources in offline policy learning.
Authors: Binghui Li, Yuanzhi Li
Abstract: Similar to surprising performance in the standard deep learning, deep nets trained by adversarial training also generalize well for $\textit{unseen clean data (natural data)}$. However, despite adversarial training can achieve low robust training error, there exists a significant $\textit{robust generalization gap}$. We call this phenomenon the $\textit{Clean Generalization and Robust Overfitting (CGRO)}$. In this work, we study the CGRO phenomenon in adversarial training from two views: $\textit{representation complexity}$ and $\textit{training dynamics}$. Specifically, we consider a binary classification setting with $N$ separated training data points. $\textit{First}$, we prove that, based on the assumption that we assume there is $\operatorname{poly}(D)$-size clean classifier (where $D$ is the data dimension), ReLU net with only $O(N D)$ extra parameters is able to leverages robust memorization to achieve the CGRO, while robust classifier still requires exponential representation complexity in worst case. $\textit{Next}$, we focus on a structured-data case to analyze training dynamics, where we train a two-layer convolutional network with $O(N D)$ width against adversarial perturbation. We then show that a three-stage phase transition occurs during learning process and the network provably converges to robust memorization regime, which thereby results in the CGRO. $\textit{Besides}$, we also empirically verify our theoretical analysis by experiments in real-image recognition datasets.
Authors: Jesse S. Ghashti, John R. J. Thompson
Abstract: Distance-based clustering and classification are widely used in various fields to group mixed numeric and categorical data. In many algorithms, a predefined distance measurement is used to cluster data points based on their dissimilarity. While there exist numerous distance-based measures for data with pure numerical attributes and several ordered and unordered categorical metrics, an efficient and accurate distance for mixed-type data that utilizes the continuous and discrete properties simulatenously is an open problem. Many metrics convert numerical attributes to categorical ones or vice versa. They handle the data points as a single attribute type or calculate a distance between each attribute separately and add them up. We propose a metric called KDSUM that uses mixed kernels to measure dissimilarity, with cross-validated optimal bandwidth selection. We demonstrate that KDSUM is a shrinkage method from existing mixed-type metrics to a uniform dissimilarity metric, and improves clustering accuracy when utilized in existing distance-based clustering algorithms on simulated and real-world datasets containing continuous-only, categorical-only, and mixed-type data.
Authors: Karol Dziedziul, Sergey Kryzhevich, Pawe{\l} Wieczy\'nski
Abstract: To present Mercer large-scale kernel machines from a ridge function perspective, we recall the results by Lin and Pinkus from {\it Fundamentality of ridge functions}. We consider the main result of the recent paper by Rachimi and Recht, 2008, {\it Random features for large-scale kernel machines} from the Approximation Theory point of view. We study which kernels could be approximated by a sum of products of cosine functions with arguments depending on $x$ and $y$ and present the obstacles of such an approach. The results of this article are applied to Image Processing by procedure "one-vs-rest".
Authors: \'Etienne de Montbrun (TSE-R), S\'ebastien Gerchinovitz (IMT)
Abstract: We consider the problem of multi-fidelity zeroth-order optimization, where one can evaluate a function $f$ at various approximation levels (of varying costs), and the goal is to optimize $f$ with the cheapest evaluations possible. In this paper, we study certified algorithms, which are additionally required to output a data-driven upper bound on the optimization error. We first formalize the problem in terms of a min-max game between an algorithm and an evaluation environment. We then propose a certified variant of the MFDOO algorithm and derive a bound on its cost complexity for any Lipschitz function $f$. We also prove an $f$-dependent lower bound showing that this algorithm has a near-optimal cost complexity. As a direct example, we close the paper by addressing the special case of noisy (stochastic) evaluations, which corresponds to $\eps$-best arm identification in Lipschitz bandits with continuously many arms.
Authors: Junette Hsin, Shubhankar Agarwal, Adam Thorpe, Luis Sentis, David Fridovich-Keil
Abstract: In this paper, we address the challenge of deriving dynamical models from sparse and noisy data. High-quality data is crucial for symbolic regression algorithms; limited and noisy data can present modeling challenges. To overcome this, we combine Gaussian process regression with a sparse identification of nonlinear dynamics (SINDy) method to denoise the data and identify nonlinear dynamical equations. Our approach GPSINDy offers improved robustness with sparse, noisy data compared to SINDy alone. We demonstrate its effectiveness on simulation data from Lotka-Volterra and unicycle models and hardware data from an NVIDIA JetRacer system. We show superior performance over baselines including more than 50% improvement over SINDy and other baselines in predicting future trajectories from noise-corrupted and sparse 5 Hz data.
Authors: Yujin Tang, Jiaming Zhou, Xiang Pan, Zeying Gong, Junwei Liang
Abstract: Accurate precipitation forecasting is a vital challenge of societal importance. Though data-driven approaches have emerged as a widely used solution, solely relying on data-driven approaches has limitations in modeling the underlying physics, making accurate predictions difficult. We focus on the Numerical Weather Prediction (NWP) post-processing based precipitation forecasting task to couple Machine Learning techniques with traditional NWP. This task remains challenging due to the imbalanced precipitation data and complex relationships between multiple meteorological variables. To address these limitations, we introduce the \textbf{PostRainBench}, a comprehensive multi-variable NWP post-processing benchmark, and \textbf{CAMT}, a simple yet effective Channel Attention Enhanced Multi-task Learning framework with a specially designed weighted loss function. Extensive experimental results on the proposed benchmark show that our method outperforms state-of-the-art methods by 6.3\%, 4.7\%, and 26.8\% in rain CSI and improvements of 15.6\%, 17.4\%, and 31.8\% over NWP predictions in heavy rain CSI on respective datasets. Most notably, our model is the first deep learning-based method to outperform NWP approaches in heavy rain conditions. These results highlight the potential impact of our model in reducing the severe consequences of extreme rainfall events. Our datasets and code are available at https://github.com/yyyujintang/PostRainBench.
Authors: Tianjun Ke, Haoqun Cao, Zenan Ling, Feng Zhou
Abstract: Meta-learning has demonstrated promising results in few-shot classification (FSC) by learning to solve new problems using prior knowledge. Bayesian methods are effective at characterizing uncertainty in FSC, which is crucial in high-risk fields. In this context, the logistic-softmax likelihood is often employed as an alternative to the softmax likelihood in multi-class Gaussian process classification due to its conditional conjugacy property. However, the theoretical property of logistic-softmax is not clear and previous research indicated that the inherent uncertainty of logistic-softmax leads to suboptimal performance. To mitigate these issues, we revisit and redesign the logistic-softmax likelihood, which enables control of the \textit{a priori} confidence level through a temperature parameter. Furthermore, we theoretically and empirically show that softmax can be viewed as a special case of logistic-softmax and logistic-softmax induces a larger family of data distribution than softmax. Utilizing modified logistic-softmax, we integrate the data augmentation technique into the deep kernel based Gaussian process meta-learning framework, and derive an analytical mean-field approximation for task-specific updates. Our approach yields well-calibrated uncertainty estimates and achieves comparable or superior results on standard benchmark datasets. Code is publicly available at \url{https://github.com/keanson/revisit-logistic-softmax}.
Authors: Vedant Shah, Frederik Tr\"auble, Ashish Malik, Hugo Larochelle, Michael Mozer, Sanjeev Arora, Yoshua Bengio, Anirudh Goyal
Abstract: Machine \emph{unlearning}, which involves erasing knowledge about a \emph{forget set} from a trained model, can prove to be costly and infeasible by existing techniques. We propose a nearly compute-free zero-shot unlearning technique based on a discrete representational bottleneck. We show that the proposed technique efficiently unlearns the forget set and incurs negligible damage to the model's performance on the rest of the data set. We evaluate the proposed technique on the problem of \textit{class unlearning} using three datasets: CIFAR-10, CIFAR-100, and LACUNA-100. We compare the proposed technique to SCRUB, a state-of-the-art approach which uses knowledge distillation for unlearning. Across all three datasets, the proposed technique performs as well as, if not better than SCRUB while incurring almost no computational cost.
Authors: Wei Wang, Takashi Ishida, Yu-Jie Zhang, Gang Niu, Masashi Sugiyama
Abstract: Complementary-label learning is a weakly supervised learning problem in which each training example is associated with one or multiple complementary labels indicating the classes to which it does not belong. Existing consistent approaches have relied on the uniform distribution assumption to model the generation of complementary labels, or on an ordinary-label training set to estimate the transition matrix in non-uniform cases. However, either condition may not be satisfied in real-world scenarios. In this paper, we propose a novel consistent approach that does not rely on these conditions. Inspired by the positive-unlabeled (PU) learning literature, we propose an unbiased risk estimator based on the Selected-Completely-at-Random assumption for complementary-label learning. We then introduce a risk-correction approach to address overfitting problems. Furthermore, we find that complementary-label learning can be expressed as a set of negative-unlabeled binary classification problems when using the one-versus-rest strategy. Extensive experimental results on both synthetic and real-world benchmark datasets validate the superiority of our proposed approach over state-of-the-art methods.
Authors: Kevin Slote, Elaine Lee
Abstract: In the industrial practice of machine learning and statistical modeling, practitioners often work under the assumption of accessible, static, labeled data for evaluation and training. However, this assumption often deviates from reality, where data may be private, encrypted, difficult-to-measure, or unlabeled. In this paper, we bridge this gap by adapting the Hui-Walter paradigm, a method traditionally applied in epidemiology and medicine, to the field of machine learning. This approach enables us to estimate key performance metrics such as false positive rate, false negative rate, and priors in scenarios where no ground truth is available. We further extend this paradigm for handling online data, opening up new possibilities for dynamic data environments. Our methodology involves partitioning data into latent classes to simulate multiple data populations (if natural populations are unavailable) and independently training models to replicate multiple tests. By cross-tabulating binary outcomes across multiple categorizers and multiple populations, we are able to estimate unknown parameters through Gibbs sampling, eliminating the need for ground-truth or labeled data. This paper showcases the potential of our methodology to transform machine learning practices by allowing for accurate model assessment under dynamic and uncertain data conditions.
Authors: Niccol\`o Turcato, Alberto Sinigaglia, Alberto Dalla Libera, Ruggero Carli, Gian Antonio Susto
Abstract: Continuous control Deep Reinforcement Learning (RL) approaches are known to suffer from estimation biases, leading to suboptimal policies. This paper introduces innovative methods in RL, focusing on addressing and exploiting estimation biases in Actor-Critic methods for continuous control tasks, using Deep Double Q-Learning. We design a Bias Exploiting (BE) mechanism to dynamically select the most advantageous estimation bias during training of the RL agent. Most State-of-the-art Deep RL algorithms can be equipped with the BE mechanism, without hindering performance or computational complexity. Our extensive experiments across various continuous control tasks demonstrate the effectiveness of our approaches. We show that RL algorithms equipped with this method can match or surpass their counterparts, particularly in environments where estimation biases significantly impact learning. The results underline the importance of bias exploitation in improving policy learning in RL.
Authors: Reza Ghane, Danil Akhtiamov, Babak Hassibi
Abstract: We study the use of linear regression for multiclass classification in the over-parametrized regime where some of the training data is mislabeled. In such scenarios it is necessary to add an explicit regularization term, $\lambda f(w)$, for some convex function $f(\cdot)$, to avoid overfitting the mislabeled data. In our analysis, we assume that the data is sampled from a Gaussian Mixture Model with equal class sizes, and that a proportion $c$ of the training labels is corrupted for each class. Under these assumptions, we prove that the best classification performance is achieved when $f(\cdot) = \|\cdot\|^2_2$ and $\lambda \to \infty$. We then proceed to analyze the classification errors for $f(\cdot) = \|\cdot\|_1$ and $f(\cdot) = \|\cdot\|_\infty$ in the large $\lambda$ regime and notice that it is often possible to find sparse and one-bit solutions, respectively, that perform almost as well as the one corresponding to $f(\cdot) = \|\cdot\|_2^2$.
Authors: Henning Schwarz, Micha \"Uberr\"uck, Jens-Peter M. Zemke, Thomas Rung
Abstract: We present approaches to predict dynamic ditching loads on aircraft fuselages using machine learning. The employed learning procedure is structured into two parts, the reconstruction of the spatial loads using a convolutional autoencoder (CAE) and the transient evolution of these loads in a subsequent part. Different CAE strategies are assessed and combined with either long short-term memory (LSTM) networks or Koopman-operator based methods to predict the transient behaviour. The training data is compiled by an extension of the momentum method of von-Karman and Wagner and the rationale of the training approach is briefly summarised. The application included refers to a full-scale fuselage of a DLR-D150 aircraft for a range of horizontal and vertical approach velocities at 6{\deg} incidence. Results indicate a satisfactory level of predictive agreement for all four investigated surrogate models examined, with the combination of an LSTM and a deep decoder CAE showing the best performance.
Authors: Huy Hoang, Tien Mai, Pradeep Varakantham
Abstract: We focus on offline imitation learning (IL), which aims to mimic an expert's behavior using demonstrations without any interaction with the environment. One of the main challenges in offline IL is the limited support of expert demonstrations, which typically cover only a small fraction of the state-action space. While it may not be feasible to obtain numerous expert demonstrations, it is often possible to gather a larger set of sub-optimal demonstrations. For example, in treatment optimization problems, there are varying levels of doctor treatments available for different chronic conditions. These range from treatment specialists and experienced general practitioners to less experienced general practitioners. Similarly, when robots are trained to imitate humans in routine tasks, they might learn from individuals with different levels of expertise and efficiency. In this paper, we propose an offline IL approach that leverages the larger set of sub-optimal demonstrations while effectively mimicking expert trajectories. Existing offline IL methods based on behavior cloning or distribution matching often face issues such as overfitting to the limited set of expert demonstrations or inadvertently imitating sub-optimal trajectories from the larger dataset. Our approach, which is based on inverse soft-Q learning, learns from both expert and sub-optimal demonstrations. It assigns higher importance (through learned weights) to aligning with expert demonstrations and lower importance to aligning with sub-optimal ones. A key contribution of our approach, called SPRINQL, is transforming the offline IL problem into a convex optimization over the space of Q functions. Through comprehensive experimental evaluations, we demonstrate that the SPRINQL algorithm achieves state-of-the-art (SOTA) performance on offline IL benchmarks. Code is available at https://github.com/hmhuy0/SPRINQL.
Authors: Aymen Rayane Khouas, Mohamed Reda Bouadjenek, Hakim Hacid, Sunil Aryal
Abstract: Edge computing has gained significant traction in recent years, promising enhanced efficiency by integrating artificial intelligence capabilities at the edge. While the focus has primarily been on the deployment and inference of Machine Learning (ML) models at the edge, the training aspect remains less explored. This survey, explores the concept of edge learning, specifically the optimization of ML model training at the edge. The objective is to comprehensively explore diverse approaches and methodologies in edge learning, synthesize existing knowledge, identify challenges, and highlight future trends. Utilizing Scopus and Web of science advanced search, relevant literature on edge learning was identified, revealing a concentration of research efforts in distributed learning methods, particularly federated learning. This survey further provides a guideline for comparing techniques used to optimize ML for edge learning, along with an exploration of the different frameworks, libraries, and simulation tools available. In doing so, the paper contributes to a holistic understanding of the current landscape and future directions in the intersection of edge computing and machine learning, paving the way for informed comparisons between optimization methods and techniques designed for training on the edge.
Authors: Minsu Kim, Sanghyeok Choi, Hyeonah Kim, Jiwoo Son, Jinkyoo Park, Yoshua Bengio
Abstract: We present the Generative Flow Ant Colony Sampler (GFACS), a novel meta-heuristic method that hierarchically combines amortized inference and parallel stochastic search. Our method first leverages Generative Flow Networks (GFlowNets) to amortize a multi-modal prior distribution over combinatorial solution space that encompasses both high-reward and diversified solutions. This prior is iteratively updated via parallel stochastic search in the spirit of Ant Colony Optimization (ACO), leading to the posterior distribution that generates near-optimal solutions. Extensive experiments across seven combinatorial optimization problems demonstrate GFACS's promising performances.
Authors: Daniel Mayfrank, Na Young Ahn, Alexander Mitsos, Manuel Dahmen
Abstract: We present a method for end-to-end learning of Koopman surrogate models for optimal performance in a specific control task. In contrast to previous contributions that employ standard reinforcement learning (RL) algorithms, we use a training algorithm that exploits the potential differentiability of environments based on mechanistic simulation models to aid the policy optimization. We evaluate the performance of our method by comparing it to that of other controller type and training algorithm combinations on an existing economic nonlinear model predictive control (eNMPC) case study of a continuous stirred-tank reactor (CSTR) model. Compared to the benchmark methods, our method produces similar economic performance but causes considerably fewer and less severe constraint violations. Thus, for this case study, our method outperforms the others and offers a promising path toward more performant controllers that employ dynamic surrogate models.
Authors: Guoxuan Xia, Olivier Laurent, Gianni Franchi, Christos-Savvas Bouganis
Abstract: Label smoothing (LS) is a popular regularisation method for training neural networks as it is effective in improving test accuracy and is simple to implement. Hard one-hot labels are smoothed by uniformly distributing probability mass to other classes, reducing overfitting. Prior work has suggested that in some cases LS can degrade selective classification (SC) -- where the aim is to reject misclassifications using a model's uncertainty. In this work, we first demonstrate empirically across an extended range of large-scale tasks and architectures that LS consistently degrades SC. We then address a gap in existing knowledge, providing an explanation for this behaviour by analysing logit-level gradients: LS degrades the uncertainty rank ordering of correct vs incorrect predictions by regularising the max logit more when a prediction is likely to be correct, and less when it is likely to be wrong. This elucidates previously reported experimental results where strong classifiers underperform in SC. We then demonstrate the empirical effectiveness of post-hoc logit normalisation for recovering lost SC performance caused by LS. Furthermore, linking back to our gradient analysis, we again provide an explanation for why such normalisation is effective.
Authors: Cameron Gordon, Lachlan Ewen MacDonald, Hemanth Saratchandran, Simon Lucey
Abstract: Deep implicit functions have been found to be an effective tool for efficiently encoding all manner of natural signals. Their attractiveness stems from their ability to compactly represent signals with little to no offline training data. Instead, they leverage the implicit bias of deep networks to decouple hidden redundancies within the signal. In this paper, we explore the hypothesis that additional compression can be achieved by leveraging redundancies that exist between layers. We propose to use a novel runtime decoder-only hypernetwork - that uses no offline training data - to better exploit cross-layer parameter redundancy. Previous applications of hypernetworks with deep implicit functions have employed feed-forward encoder/decoder frameworks that rely on large offline datasets that do not generalize beyond the signals they were trained on. We instead present a strategy for the optimization of runtime deep implicit functions for single-instance signals through a Decoder-Only randomly projected Hypernetwork (D'OH). By directly changing the latent code dimension, we provide a natural way to vary the memory footprint of neural representations without the costly need for neural architecture search on a space of alternative low-rate structures.
Authors: Ruiqi Zhang, Licong Lin, Yu Bai, Song Mei
Abstract: Large Language Models (LLMs) often memorize sensitive, private, or copyrighted data during pre-training. LLM unlearning aims to eliminate the influence of undesirable data from the pre-trained model while preserving the model's utilities on other tasks. Several practical methods have recently been proposed for LLM unlearning, mostly based on gradient ascent (GA) on the loss of undesirable data. However, on certain unlearning tasks, these methods either fail to effectively unlearn the target data or suffer from catastrophic collapse -- a drastic degradation of the model's utilities. In this paper, we propose Negative Preference Optimization (NPO), a simple alignment-inspired method that could efficiently and effectively unlearn a target dataset. We theoretically show that the progression toward catastrophic collapse by minimizing the NPO loss is exponentially slower than GA. Through experiments on synthetic data and the benchmark TOFU dataset, we demonstrate that NPO-based methods achieve a better balance between unlearning the undesirable data and maintaining the model's utilities. We also observe that NPO-based methods generate more sensible outputs than GA-based methods, whose outputs are often gibberish. Remarkably, on TOFU, NPO-based methods are the first to achieve reasonable unlearning results in forgetting 50% (or more) of the training data, whereas existing methods already struggle with forgetting 10% of training data.
Authors: Alexey Gorbatovski, Boris Shaposhnikov, Alexey Malakhov, Nikita Surnachev, Yaroslav Aksenov, Ian Maksimov, Nikita Balagansky, Daniil Gavrilov
Abstract: Despite the fact that offline methods for Large Language Models (LLMs) alignment do not require a direct reward model, they remain susceptible to overoptimization. This issue arises when the trained model deviates excessively from the reference policy, leading to a decrease in sample quality. We propose a new paradigm of offline alignment methods, called Trust Region (including variants TR-DPO, TR-IPO, TR-KTO), which dynamically updates the reference policy throughout the training process. Our results show that TR alignment methods effectively mitigate overoptimization, enabling models to maintain strong performance even when substantially deviating from the initial reference policy. We demonstrate the efficacy of these approaches not only through toy examples that exhibit reduced overoptimization, but also through direct, side-by-side comparisons in specific tasks such as helpful and harmless dialogue, as well as summarization, where they surpass conventional methods. Additionally, we report significant improvements in general-purpose assistant setups with the Llama3 model on the AlpacaEval 2 and Arena-Hard benchmarks, highlighting the advantages of Trust Region methods over classical approaches.
Authors: Lunjia Hu, Yifan Wu
Abstract: Calibration allows predictions to be reliably interpreted as probabilities by decision makers. We propose a decision-theoretic calibration error, the Calibration Decision Loss (CDL), defined as the maximum improvement in decision payoff obtained by calibrating the predictions, where the maximum is over all payoff-bounded decision tasks. Vanishing CDL guarantees the payoff loss from miscalibration vanishes simultaneously for all downstream decision tasks. We show separations between CDL and existing calibration error metrics, including the most well-studied metric Expected Calibration Error (ECE). Our main technical contribution is a new efficient algorithm for online calibration that achieves near-optimal $O(\frac{\log T}{\sqrt{T}})$ expected CDL, bypassing the $\Omega(T^{-0.472})$ lower bound for ECE by Qiao and Valiant (2021).
Authors: Mikkel Odgaard, Kiril Vadimovic Klein, Sanne M{\o}ller Thysen, Espen Jimenez-Solem, Martin Sillesen, Mads Nielsen
Abstract: The widespread adoption of Electronic Health Records (EHR) has significantly increased the amount of available healthcare data. This has allowed models inspired by Natural Language Processing (NLP) and Computer Vision, which scale exceptionally well, to be used in EHR research. Particularly, BERT-based models have surged in popularity following the release of BEHRT and Med-BERT. Subsequent models have largely built on these foundations despite the fundamental design choices of these pioneering models remaining underexplored. Through incremental optimization, we study BERT-based EHR modeling and isolate the sources of improvement for key design choices, giving us insights into the effect of data representation, individual technical components, and training procedure. Evaluating this across a set of generic tasks (death, pain treatment, and general infection), we showed that improving data representation can increase the average downstream performance from 0.785 to 0.797 AUROC ($p<10^{-7}$), primarily when including medication and timestamps. Improving the architecture and training protocol on top of this increased average downstream performance to 0.801 AUROC ($p<10^{-7}$). We then demonstrated the consistency of our optimization through a rigorous evaluation across 25 diverse clinical prediction tasks. We observed significant performance increases in 17 out of 25 tasks and improvements in 24 tasks, highlighting the generalizability of our results. Our findings provide a strong foundation for future work and aim to increase the trustworthiness of BERT-based EHR models.
Authors: Seong-Joon Park, Hee-Youl Kwak, Sang-Hyo Kim, Yongjune Kim, Jong-Seon No
Abstract: Error correcting codes (ECCs) are indispensable for reliable transmission in communication systems. The recent advancements in deep learning have catalyzed the exploration of ECC decoders based on neural networks. Among these, transformer-based neural decoders have achieved state-of-the-art decoding performance. In this paper, we propose a novel Cross-attention Message-Passing Transformer (CrossMPT), which shares key operational principles with conventional message-passing decoders. While conventional transformer-based decoders employ self-attention mechanism without distinguishing between the types of input vectors (i.e., magnitude and syndrome vectors), CrossMPT updates the two types of input vectors separately and iteratively using two masked cross-attention blocks. The mask matrices are determined by the code's parity-check matrix, which explicitly captures the irrelevant relationship between two input vectors. Our experimental results show that CrossMPT significantly outperforms existing neural network-based decoders for various code classes. Notably, CrossMPT achieves this decoding performance improvement, while significantly reducing the memory usage, complexity, inference time, and training time.
Authors: Tian Xie, Xueru Zhang
Abstract: As machine learning (ML) models are increasingly used in social domains to make consequential decisions about humans, they often have the power to reshape data distributions. Humans, as strategic agents, continuously adapt their behaviors in response to the learning system. As populations change dynamically, ML systems may need frequent updates to ensure high performance. However, acquiring high-quality human-annotated samples can be highly challenging and even infeasible in social domains. A common practice to address this issue is using the model itself to annotate unlabeled data samples. This paper investigates the long-term impacts when ML models are retrained with model-annotated samples when they incorporate human strategic responses. We first formalize the interactions between strategic agents and the model and then analyze how they evolve under such dynamic interactions. We find that agents are increasingly likely to receive positive decisions as the model gets retrained, whereas the proportion of agents with positive labels may decrease over time. We thus propose a refined retraining process to stabilize the dynamics. Last, we examine how algorithmic fairness can be affected by these retraining processes and find that enforcing common fairness constraints at every round may not benefit the disadvantaged group in the long run. Experiments on (semi-)synthetic and real data validate the theoretical findings.
Authors: Xingzhou Lou, Junge Zhang, Jian Xie, Lifeng Liu, Dong Yan, Kaiqi Huang
Abstract: Human preference alignment is critical in building powerful and reliable large language models (LLMs). However, current methods either ignore the multi-dimensionality of human preferences (e.g. helpfulness and harmlessness) or struggle with the complexity of managing multiple reward models. To address these issues, we propose Sequential Preference Optimization (SPO), a method that sequentially fine-tunes LLMs to align with multiple dimensions of human preferences. SPO avoids explicit reward modeling, directly optimizing the models to align with nuanced human preferences. We theoretically derive closed-form optimal SPO policy and loss function. Gradient analysis is conducted to show how SPO manages to fine-tune the LLMs while maintaining alignment on previously optimized dimensions. Empirical results on LLMs of different size and multiple evaluation datasets demonstrate that SPO successfully aligns LLMs across multiple dimensions of human preferences and significantly outperforms the baselines.
Authors: Xixi Wu, Yifei Shen, Caihua Shan, Kaitao Song, Siwei Wang, Bohang Zhang, Jiarui Feng, Hong Cheng, Wei Chen, Yun Xiong, Dongsheng Li
Abstract: Task planning in language agents is emerging as an important research topic alongside the development of large language models (LLMs). It aims to break down complex user requests in natural language into solvable sub-tasks, thereby fulfilling the original requests. In this context, the sub-tasks can be naturally viewed as a graph, where the nodes represent the sub-tasks, and the edges denote the dependencies among them. Consequently, task planning is a decision-making problem that involves selecting a connected path or subgraph within the corresponding graph and invoking it. In this paper, we explore graph learning-based methods for task planning, a direction that is orthogonal to the prevalent focus on prompt design. Our interest in graph learning stems from a theoretical discovery: the biases of attention and auto-regressive loss impede LLMs' ability to effectively navigate decision-making on graphs, which is adeptly addressed by graph neural networks (GNNs). This theoretical insight led us to integrate GNNs with LLMs to enhance overall performance. Extensive experiments demonstrate that GNN-based methods surpass existing solutions even without training, and minimal training can further enhance their performance. The performance gain increases with a larger task graph size.
Authors: Massimo Bini, Karsten Roth, Zeynep Akata, Anna Khoreva
Abstract: Parameter-efficient finetuning (PEFT) has become ubiquitous to adapt foundation models to downstream task requirements while retaining their generalization ability. However, the amount of additionally introduced parameters and compute for successful adaptation and hyperparameter searches can explode quickly, especially when deployed at scale to serve numerous individual requests. To ensure effective, parameter-efficient, and hyperparameter-robust adaptation, we propose the ETHER transformation family, which performs Efficient fineTuning via HypErplane Reflections. By design, ETHER transformations require a minimal number of parameters, are less likely to deteriorate model performance, and exhibit robustness to hyperparameter and learning rate choices. In particular, we introduce ETHER and its relaxation ETHER+, which match or outperform existing PEFT methods with significantly fewer parameters ($\sim$$10$-$100$ times lower than LoRA or OFT) across multiple image synthesis and natural language tasks without exhaustive hyperparameter tuning. Finally, we investigate the recent emphasis on Hyperspherical Energy retention for adaptation and raise questions on its practical utility. The code is available at https://github.com/mwbini/ether.
Authors: Kenji Fukumizu, Taiji Suzuki, Noboru Isobe, Kazusato Oko, Masanori Koyama
Abstract: Flow matching (FM) has gained significant attention as a simulation-free generative model. Unlike diffusion models, which are based on stochastic differential equations, FM employs a simpler approach by solving an ordinary differential equation with an initial condition from a normal distribution, thus streamlining the sample generation process. This paper discusses the convergence properties of FM for large sample size under the $p$-Wasserstein distance, a measure of distributional discrepancy. We establish that FM can achieve an almost minimax optimal convergence rate for $1 \leq p \leq 2$, presenting the first theoretical evidence that FM can reach convergence rates comparable to those of diffusion models. Our analysis extends existing frameworks by examining a broader class of mean and variance functions for the vector fields and identifies specific conditions necessary to attain almost optimal rates.
Authors: Dimitrios Samakovlis, Stefano Albini, Rub\'en Rodr\'iguez \'Alvarez, Denisa-Andreea Constantinescu, Pasquale Davide Schiavone, Miguel Pe\'on Quir\'os, David Atienza
Abstract: The design of low-power wearables for the biomedical domain has received a lot of attention in recent decades, as technological advances in chip manufacturing have allowed real-time monitoring of patients using low-complexity ML within the mW range. Despite advances in application and hardware design research, the domain lacks a systematic approach to hardware evaluation. In this work, we propose BiomedBench, a new benchmark suite composed of complete end-to-end TinyML biomedical applications for real-time monitoring of patients using wearable devices. Each application presents different requirements during typical signal acquisition and processing phases, including varying computational workloads and relations between active and idle times. Furthermore, our evaluation of five state-of-the-art low-power platforms in terms of energy efficiency shows that modern platforms cannot effectively target all types of biomedical applications. BiomedBench is released as an open-source suite to standardize hardware evaluation and guide hardware and application design in the TinyML wearable domain.
Authors: Antti Koskela, Jafar Mohammadi
Abstract: We present a novel method for accurately auditing the differential privacy (DP) guarantees of DP mechanisms. In particular, our solution is applicable to auditing DP guarantees of machine learning (ML) models. Previous auditing methods tightly capture the privacy guarantees of DP-SGD trained models in the white-box setting where the auditor has access to all intermediate models; however, the success of these methods depends on a priori information about the parametric form of the noise and the subsampling ratio used for sampling the gradients. We present a method that does not require such information and is agnostic to the randomization used for the underlying mechanism. Similarly to several previous DP auditing methods, we assume that the auditor has access to a set of independent observations from two one-dimensional distributions corresponding to outputs from two neighbouring datasets. Furthermore, our solution is based on a simple histogram-based density estimation technique to find lower bounds for the statistical distance between these distributions when measured using the hockey-stick divergence. We show that our approach also naturally generalizes the previously considered class of threshold membership inference auditing methods. We improve upon accurate auditing methods such as the $f$-DP auditing. Moreover, we address an open problem on how to accurately audit the subsampled Gaussian mechanism without any knowledge of the parameters of the underlying mechanism.
Authors: Weronika Ormaniec, Scott Sussex, Lars Lorch, Bernhard Sch\"olkopf, Andreas Krause
Abstract: Synthetic datasets generated by structural causal models (SCMs) are commonly used for benchmarking causal structure learning algorithms. However, the variances and pairwise correlations in SCM data tend to increase along the causal ordering. Several popular algorithms exploit these artifacts, possibly leading to conclusions that do not generalize to real-world settings. Existing metrics like $\operatorname{Var}$-sortability and $\operatorname{R^2}$-sortability quantify these patterns, but they do not provide tools to remedy them. To address this, we propose internally-standardized structural causal models (iSCMs), a modification of SCMs that introduces a standardization operation at each variable during the generative process. By construction, iSCMs are not $\operatorname{Var}$-sortable, and as we show experimentally, not $\operatorname{R^2}$-sortable either for commonly-used graph families. Moreover, contrary to the post-hoc standardization of data generated by standard SCMs, we prove that linear iSCMs are less identifiable from prior knowledge on the weights and do not collapse to deterministic relationships in large systems, which may make iSCMs a useful model in causal inference beyond the benchmarking problem studied here.
Authors: Jialun Cao, Zhiyong Chen, Jiarong Wu, Shing-chi Cheung, Chang Xu
Abstract: Code generation benchmarks such as HumanEval are widely adopted to evaluate LLMs' capabilities. However, after consolidating the latest 24 benchmarks, we noticed three significant imbalances. First, imbalanced programming language. 95.8% of benchmarks involve Python, while only 5 benchmarks involve Java. Second, imbalanced code granularity. Function-/statement-level benchmarks account for over 83.3% of benchmarks. Only a mere handful extends to class-/project-levels, and all are limited to Python. Third, lacking advanced features. Existing benchmarks primarily assess basic coding skills, while overlooking advanced Object-Oriented Programming (OOP) features (i.e., encapsulation, inheritance, and polymorphism). To fill these gaps, we propose JavaBench, a project-level Java benchmark that exercises OOP features. It comprises four Java projects with 389 methods in 106 Java classes. The test coverage is up to 92%, and JavaBench is attested by 282 undergraduate students, reaching a 90.93/100 average score (i.e., pass rate against the test suite), ensuring the quality of documentation, code skeleton, and tests. To better evaluate LLM's capability against JavaBench, we introduce a systematic evaluation design covering three context settings and five synthesis strategies at two granularities using three hierarchical metrics. Our extensive experiment yields several interesting findings. First, we noticed that regarding project-level Java programming, LLMs are far behind undergraduate students (no project can be correctly completed by any studied LLMs, and at most 41.17% Pass@5 in a more relaxed evaluation). Second, using method signature as prompt context may strike an ideal balance for project-level code generation. JavaBench is publicly available at https://github.com/java-bench/JavaBench.
Authors: Yapei Chang, Kalpesh Krishna, Amir Houmansadr, John Wieting, Mohit Iyyer
Abstract: The most effective techniques to detect LLM-generated text rely on inserting a detectable signature -- or watermark -- during the model's decoding process. Most existing watermarking methods require access to the underlying LLM's logits, which LLM API providers are loath to share due to fears of model distillation. As such, these watermarks must be implemented independently by each LLM provider. In this paper, we develop PostMark, a modular post-hoc watermarking procedure in which an input-dependent set of words (determined via a semantic embedding) is inserted into the text after the decoding process has completed. Critically, PostMark does not require logit access, which means it can be implemented by a third party. We also show that PostMark is more robust to paraphrasing attacks than existing watermarking methods: our experiments cover eight baseline algorithms, five base LLMs, and three datasets. Finally, we evaluate the impact of PostMark on text quality using both automated and human assessments, highlighting the trade-off between quality and robustness to paraphrasing. We release our code, outputs, and annotations at https://github.com/lilakk/PostMark.
Authors: Zhengyang Geng, Ashwini Pokle, William Luo, Justin Lin, J. Zico Kolter
Abstract: Consistency models (CMs) offer faster sampling than traditional diffusion models, but their training is resource-intensive. For example, as of 2024, training a state-of-the-art CM on CIFAR-10 takes one week on 8 GPUs. In this work, we propose an effective scheme for training CMs that largely improves the efficiency of building such models. Specifically, by expressing CM trajectories via a particular differential equation, we argue that diffusion models can be viewed as a special case of CMs. We can thus fine-tune a consistency model starting from a pretrained diffusion model and progressively approximate the full consistency condition to stronger degrees over the training process. Our resulting method, which we term Easy Consistency Tuning (ECT), achieves vastly reduced training times while improving upon the quality of previous methods: for example, ECT achieves a 2-step FID of 2.73 on CIFAR10 within 1 hour on a single A100 GPU, matching Consistency Distillation trained for hundreds of GPU hours. Owing to this computational efficiency, we investigate the scaling laws of CMs under ECT, showing that they obey the classic power law scaling, hinting at their ability to improve efficiency and performance at larger scales. Our code (https://github.com/locuslab/ect) is publicly available, making CMs more accessible to the broader community.
Authors: Ali Naseh, Jaechul Roh, Eugene Bagdasaryan, Amir Houmansadr
Abstract: Text-conditional diffusion models, i.e. text-to-image, produce eye-catching images that represent descriptions given by a user. These images often depict benign concepts but could also carry other purposes. Specifically, visual information is easy to comprehend and could be weaponized for propaganda -- a serious challenge given widespread usage and deployment of generative models. In this paper, we show that an adversary can add an arbitrary bias through a backdoor attack that would affect even benign users generating images. While a user could inspect a generated image to comply with the given text description, our attack remains stealthy as it preserves semantic information given in the text prompt. Instead, a compromised model modifies other unspecified features of the image to add desired biases (that increase by 4-8x). Furthermore, we show how the current state-of-the-art generative models make this attack both cheap and feasible for any adversary, with costs ranging between $12-$18. We evaluate our attack over various types of triggers, adversary objectives, and biases and discuss mitigations and future work. Our code is available at https://github.com/jrohsc/Backdororing_Bias.
Authors: Yifan Chen, Daniel Zhengyu Huang, Jiaoyang Huang, Sebastian Reich, Andrew M. Stuart
Abstract: In this paper, we study efficient approximate sampling for probability distributions known up to normalization constants. We specifically focus on a problem class arising in Bayesian inference for large-scale inverse problems in science and engineering applications. The computational challenges we address with the proposed methodology are: (i) the need for repeated evaluations of expensive forward models; (ii) the potential existence of multiple modes; and (iii) the fact that gradient of, or adjoint solver for, the forward model might not be feasible. While existing Bayesian inference methods meet some of these challenges individually, we propose a framework that tackles all three systematically. Our approach builds upon the Fisher-Rao gradient flow in probability space, yielding a dynamical system for probability densities that converges towards the target distribution at a uniform exponential rate. This rapid convergence is advantageous for the computational burden outlined in (i). We apply Gaussian mixture approximations with operator splitting techniques to simulate the flow numerically; the resulting approximation can capture multiple modes thus addressing (ii). Furthermore, we employ the Kalman methodology to facilitate a derivative-free update of these Gaussian components and their respective weights, addressing the issue in (iii). The proposed methodology results in an efficient derivative-free sampler flexible enough to handle multi-modal distributions: Gaussian Mixture Kalman Inversion (GMKI). The effectiveness of GMKI is demonstrated both theoretically and numerically in several experiments with multimodal target distributions, including proof-of-concept and two-dimensional examples, as well as a large-scale application: recovering the Navier-Stokes initial condition from solution data at positive times.
Authors: Xinghao Wu, Xuefeng Liu, Jianwei Niu, Haolin Wang, Shaojie Tang, Guogang Zhu, Hao Su
Abstract: To address data heterogeneity, the key strategy of Personalized Federated Learning (PFL) is to decouple general knowledge (shared among clients) and client-specific knowledge, as the latter can have a negative impact on collaboration if not removed. Existing PFL methods primarily adopt a parameter partitioning approach, where the parameters of a model are designated as one of two types: parameters shared with other clients to extract general knowledge and parameters retained locally to learn client-specific knowledge. However, as these two types of parameters are put together like a jigsaw puzzle into a single model during the training process, each parameter may simultaneously absorb both general and client-specific knowledge, thus struggling to separate the two types of knowledge effectively. In this paper, we introduce FedDecomp, a simple but effective PFL paradigm that employs parameter additive decomposition to address this issue. Instead of assigning each parameter of a model as either a shared or personalized one, FedDecomp decomposes each parameter into the sum of two parameters: a shared one and a personalized one, thus achieving a more thorough decoupling of shared and personalized knowledge compared to the parameter partitioning method. In addition, as we find that retaining local knowledge of specific clients requires much lower model capacity compared with general knowledge across all clients, we let the matrix containing personalized parameters be low rank during the training process. Moreover, a new alternating training strategy is proposed to further improve the performance. Experimental results across multiple datasets and varying degrees of data heterogeneity demonstrate that FedDecomp outperforms state-of-the-art methods up to 4.9\%. The code is available at https://github.com/XinghaoWu/FedDecomp.
Authors: Antti Koskela
Abstract: The hidden state threat model of differential privacy (DP) assumes that the adversary has access only to the final trained machine learning (ML) model, without seeing intermediate states during training. Current privacy analyses under this model, however, are limited to convex optimization problems, reducing their applicability to multi-layer neural networks, which are essential in modern deep learning applications. Additionally, the most successful applications of the hidden state privacy analyses in classification tasks have been for logistic regression models. We demonstrate that it is possible to privately train convex problems with privacy-utility trade-offs comparable to those of one hidden-layer ReLU networks trained with DP stochastic gradient descent (DP-SGD). We achieve this through a stochastic approximation of a dual formulation of the ReLU minimization problem which results in a strongly convex problem. This enables the use of existing hidden state privacy analyses, providing accurate privacy bounds also for the noisy cyclic mini-batch gradient descent (NoisyCGD) method with fixed disjoint mini-batches. Our experiments on benchmark classification tasks show that NoisyCGD can achieve privacy-utility trade-offs comparable to DP-SGD applied to one-hidden-layer ReLU networks. Additionally, we provide theoretical utility bounds that highlight the speed-ups gained through the convex approximation.
Authors: Kamran Chitsaz, Quentin Fournier, Gon\c{c}alo Mordido, Sarath Chandar
Abstract: The increasing scale of Transformer models has led to an increase in their pre-training computational requirements. While quantization has proven to be effective after pre-training and during fine-tuning, applying quantization in Transformers during pre-training has remained largely unexplored at scale for language modeling. This study aims to explore the impact of quantization for efficient pre-training of Transformers, with a focus on linear layer components. By systematically applying straightforward linear quantization to weights, activations, gradients, and optimizer states, we assess its effects on model efficiency, stability, and performance during training. By offering a comprehensive recipe of effective quantization strategies to be applied during the pre-training of Transformers, we promote high training efficiency from scratch while retaining language modeling ability. Code is available at https://github.com/chandar-lab/EfficientLLMs.
Authors: Enqiang Zhu, Xiang Li, Chanjuan Liu, Nikhil R. Pal
Abstract: Uncovering new therapeutic uses of existing drugs, drug repositioning offers a fast and cost-effective strategy and holds considerable significance in the realm of drug discovery and development. In recent years, deep learning techniques have emerged as powerful tools in drug repositioning due to their ability to analyze large and complex datasets. However, many existing methods focus on extracting feature information from nearby nodes in the network to represent drugs and diseases, without considering the potential inter-relationships between the features of drugs and diseases, which may lead to inaccurate representations. To address this limitation, we use two features (similarity and association) to capture the potential relationships between the features of drugs and diseases, proposing a Dual-Feature Drug Repositioning Neural Network (DFDRNN) model. DFDRNN uses a self-attention mechanism to extract neighbor features and incorporates two dual-feature extraction modules: the intra-domain dual-feature extraction (IntraDDFE) module for extracting features within a single domain (drugs or diseases) and the inter-domain dual-feature extraction (InterDDFE) module for extracting features across domains. By utilizing these modules, we ensure more appropriate encoding of drugs and diseases. Additionally, a cross-dual-domain decoder is designed to predict drug-disease associations in both domains. Our proposed DFDRNN model outperforms six state-of-the-art methods on four benchmark datasets, achieving an average AUROC of 0.946 and an average AUPR of 0.597. Case studies on two diseases show that the proposed DFDRNN model can be applied in real-world scenarios, demonstrating its significant potential in drug repositioning.
Authors: Ayush Kaushal, Tejas Vaidhya, Arnab Kumar Mondal, Tejas Pandey, Aaryan Bhagat, Irina Rish
Abstract: Rapid advancements in GPU computational power has outpaced memory capacity and bandwidth growth, creating bottlenecks in Large Language Model (LLM) inference. Post-training quantization is the leading method for addressing memory-related bottlenecks in LLM inference, but it suffers from significant performance degradation below 4-bit precision. This paper addresses these challenges by investigating the pretraining of low-bitwidth models specifically Ternary Language Models (TriLMs) as an alternative to traditional floating-point models (FloatLMs) and their post-training quantized versions (QuantLMs). We present Spectra LLM suite, the first open suite of LLMs spanning multiple bit-widths, including FloatLMs, QuantLMs, and TriLMs, ranging from 99M to 3.9B parameters trained on 300B tokens. Our comprehensive evaluation demonstrates that TriLMs offer superior scaling behavior in terms of model size (in bits). Surprisingly, at scales exceeding one billion parameters, TriLMs consistently outperform their QuantLM and FloatLM counterparts for a given bit size across various benchmarks. Notably, the 3.9B parameter TriLM matches the performance of the FloatLM 3.9B across all benchmarks, despite having fewer bits than FloatLM 830M. Overall, this research provides valuable insights into the feasibility and scalability of low-bitwidth language models, paving the way for the development of more efficient LLMs. To enhance understanding of low-bitwidth models, we are releasing 500+ intermediate checkpoints of the Spectra suite at https://github.com/NolanoOrg/SpectraSuite.
Authors: Philipp Petersen, Jakob Zech
Abstract: This book provides an introduction to the mathematical analysis of deep learning. It covers fundamental results in approximation theory, optimization theory, and statistical learning theory, which are the three main pillars of deep neural network theory. Serving as a guide for students and researchers in mathematics and related fields, the book aims to equip readers with foundational knowledge on the topic. It prioritizes simplicity over generality, and presents rigorous yet accessible results to help build an understanding of the essential mathematical concepts underpinning deep learning.
Authors: Shu Ishida, Jo\~ao F. Henriques
Abstract: This work compares ways of extending Reinforcement Learning algorithms to Partially Observed Markov Decision Processes (POMDPs) with options. One view of options is as temporally extended action, which can be realized as a memory that allows the agent to retain historical information beyond the policy's context window. While option assignment could be handled using heuristics and hand-crafted objectives, learning temporally consistent options and associated sub-policies without explicit supervision is a challenge. Two algorithms, PPOEM and SOAP, are proposed and studied in depth to address this problem. PPOEM applies the forward-backward algorithm (for Hidden Markov Models) to optimize the expected returns for an option-augmented policy. However, this learning approach is unstable during on-policy rollouts. It is also unsuited for learning causal policies without the knowledge of future trajectories, since option assignments are optimized for offline sequences where the entire episode is available. As an alternative approach, SOAP evaluates the policy gradient for an optimal option assignment. It extends the concept of the generalized advantage estimation (GAE) to propagate option advantages through time, which is an analytical equivalent to performing temporal back-propagation of option policy gradients. This option policy is only conditional on the history of the agent, not future actions. Evaluated against competing baselines, SOAP exhibited the most robust performance, correctly discovering options for POMDP corridor environments, as well as on standard benchmarks including Atari and MuJoCo, outperforming PPOEM, as well as LSTM and Option-Critic baselines. The open-sourced code is available at https://github.com/shuishida/SoapRL.
Authors: Steve Yuwono, Dorothea Schwung, Andreas Schwung
Abstract: This paper presents a novel transfer learning approach in state-based potential games (TL-SbPGs) for enhancing distributed self-optimization in manufacturing systems. The approach focuses on the practical relevant industrial setting where sharing and transferring gained knowledge among similar-behaved players improves the self-learning mechanism in large-scale systems. With TL-SbPGs, the gained knowledge can be reused by other players to optimize their policies, thereby improving the learning outcomes of the players and accelerating the learning process. To accomplish this goal, we develop transfer learning concepts and similarity criteria for players, which offer two distinct settings: (a) predefined similarities between players and (b) dynamically inferred similarities between players during training. We formally prove the applicability of the SbPG framework in transfer learning. Additionally, we introduce an efficient method to determine the optimal timing and weighting of the transfer learning procedure during the training phase. Through experiments on a laboratory-scale testbed, we demonstrate that TL-SbPGs significantly boost production efficiency while reducing power consumption of the production schedules while also outperforming native SbPGs.
Authors: Lunjun Zhang, Arian Hosseini, Hritik Bansal, Mehran Kazemi, Aviral Kumar, Rishabh Agarwal
Abstract: Verifiers or reward models are often used to enhance the reasoning performance of large language models (LLMs). A common approach is the Best-of-N method, where N candidate solutions generated by the LLM are ranked by a verifier, and the best one is selected. While LLM-based verifiers are typically trained as discriminative classifiers to score solutions, they do not utilize the text generation capabilities of pretrained LLMs. To overcome this limitation, we instead propose training verifiers using the ubiquitous next-token prediction objective, jointly on verification and solution generation. Compared to standard verifiers, such generative verifiers (GenRM) can benefit from several advantages of LLMs: they integrate seamlessly with instruction tuning, enable chain-of-thought reasoning, and can utilize additional test-time compute via majority voting for better verification. We demonstrate that GenRM outperforms discriminative, DPO verifiers, and LLM-as-a-Judge, resulting in a 16-40% improvement in the number of problems solved with Best-of-N on algorithmic and math reasoning tasks. Furthermore, we find that training GenRM with synthetic verification rationales is sufficient to pick out subtle errors on math problems. Finally, we demonstrate that generative verifiers scale favorably with model size and inference-time compute.
Authors: Pritthijit Nath, Henry Moss, Emily Shuckburgh, Mark Webb
Abstract: This study explores integrating reinforcement learning (RL) with idealised climate models to address key parameterisation challenges in climate science. Current climate models rely on complex mathematical parameterisations to represent sub-grid scale processes, which can introduce substantial uncertainties. RL offers capabilities to enhance these parameterisation schemes, including direct interaction, handling sparse or delayed feedback, continuous online learning, and long-term optimisation. We evaluate the performance of eight RL algorithms on two idealised environments: one for temperature bias correction, another for radiative-convective equilibrium (RCE) imitating real-world computational constraints. Results show different RL approaches excel in different climate scenarios with exploration algorithms performing better in bias correction, while exploitation algorithms proving more effective for RCE. These findings support the potential of RL-based parameterisation schemes to be integrated into global climate models, improving accuracy and efficiency in capturing complex climate dynamics. Overall, this work represents an important first step towards leveraging RL to enhance climate model accuracy, critical for improving climate understanding and predictions. Code accessible at https://github.com/p3jitnath/climate-rl.
Authors: Anton Kozyriev, Vladimir Norkin
Abstract: This paper addresses the limitations of conventional vector quantization algorithms, particularly K-Means and its variant K-Means++, and investigates the Stochastic Quantization (SQ) algorithm as a scalable alternative for high-dimensional unsupervised and semi-supervised learning tasks. Traditional clustering algorithms often suffer from inefficient memory utilization during computation, necessitating the loading of all data samples into memory, which becomes impractical for large-scale datasets. While variants such as Mini-Batch K-Means partially mitigate this issue by reducing memory usage, they lack robust theoretical convergence guarantees due to the non-convex nature of clustering problems. In contrast, the Stochastic Quantization algorithm provides strong theoretical convergence guarantees, making it a robust alternative for clustering tasks. We demonstrate the computational efficiency and rapid convergence of the algorithm on an image classification problem with partially labeled data, comparing model accuracy across various ratios of labeled to unlabeled data. To address the challenge of high dimensionality, we employ a Triplet Network to encode images into low-dimensional representations in a latent space, which serve as a basis for comparing the efficiency of both the Stochastic Quantization algorithm and traditional quantization algorithms. Furthermore, we enhance the algorithm's convergence speed by introducing modifications with an adaptive learning rate.
Authors: Brian Hyeongseok Kim, Jingbo Wang, Chao Wang
Abstract: We propose a method for formally certifying and quantifying individual fairness of deep neural networks (DNN). Individual fairness guarantees that any two individuals who are identical except for a legally protected attribute (e.g., gender or race) receive the same treatment. While there are existing techniques that provide such a guarantee, they tend to suffer from lack of scalability or accuracy as the size and input dimension of the DNN increase. Our method overcomes this limitation by applying abstraction to a symbolic interval based analysis of the DNN followed by iterative refinement guided by the fairness property. Furthermore, our method lifts the symbolic interval based analysis from conventional qualitative certification to quantitative certification, by computing the percentage of individuals whose classification outputs are provably fair, instead of merely deciding if the DNN is fair. We have implemented our method and evaluated it on deep neural networks trained on four popular fairness research datasets. The experimental results show that our method is not only more accurate than state-of-the-art techniques but also several orders-of-magnitude faster.
Authors: Nirmalya Thakur
Abstract: The world is currently experiencing an outbreak of mpox, which has been declared a Public Health Emergency of International Concern by WHO. No prior work related to social media mining has focused on the development of a dataset of Instagram posts about the mpox outbreak. The work presented in this paper aims to address this research gap and makes two scientific contributions to this field. First, it presents a multilingual dataset of 60,127 Instagram posts about mpox, published between July 23, 2022, and September 5, 2024. The dataset, available at https://dx.doi.org/10.21227/7fvc-y093, contains Instagram posts about mpox in 52 languages. For each of these posts, the Post ID, Post Description, Date of publication, language, and translated version of the post (translation to English was performed using the Google Translate API) are presented as separate attributes in the dataset. After developing this dataset, sentiment analysis, hate speech detection, and anxiety or stress detection were performed. This process included classifying each post into (i) one of the sentiment classes, i.e., fear, surprise, joy, sadness, anger, disgust, or neutral, (ii) hate or not hate, and (iii) anxiety/stress detected or no anxiety/stress detected. These results are presented as separate attributes in the dataset. Second, this paper presents the results of performing sentiment analysis, hate speech analysis, and anxiety or stress analysis. The variation of the sentiment classes - fear, surprise, joy, sadness, anger, disgust, and neutral were observed to be 27.95%, 2.57%, 8.69%, 5.94%, 2.69%, 1.53%, and 50.64%, respectively. In terms of hate speech detection, 95.75% of the posts did not contain hate and the remaining 4.25% of the posts contained hate. Finally, 72.05% of the posts did not indicate any anxiety/stress, and the remaining 27.95% of the posts represented some form of anxiety/stress.
Authors: Zhiyue Xia, Kathleen Stewart
Abstract: Synthetic opioids are the most common drugs involved in drug-involved overdose mortalities in the U.S. The Center for Disease Control and Prevention reported that in 2018, about 70% of all drug overdose deaths involved opioids and 67% of all opioid-involved deaths were accounted for by synthetic opioids. In this study, we investigated the spread of synthetic opioids between 2013 and 2020 in the U.S. We analyzed the relationship between the spatiotemporal pattern of synthetic opioid-involved deaths and another key opioid, heroin, and compared patterns of deaths involving these two types of drugs during this period. Spatial connections and human mobility between counties were incorporated into a graph convolutional neural network model to represent and analyze the spread of synthetic opioid-involved deaths in the context of previous heroin-involved death patterns.
Authors: Xinrui Wang, Chuanxing Geng, Wenhai Wan, Shao-yuan Li, Songcan Chen
Abstract: Online continual learning requires the models to learn from constant, endless streams of data. While significant efforts have been made in this field, most were focused on mitigating the catastrophic forgetting issue to achieve better classification ability, at the cost of a much heavier training workload. They overlooked that in real-world scenarios, e.g., in high-speed data stream environments, data do not pause to accommodate slow models. In this paper, we emphasize that model throughput -- defined as the maximum number of training samples that a model can process within a unit of time -- is equally important. It directly limits how much data a model can utilize and presents a challenging dilemma for current methods. With this understanding, we revisit key challenges in OCL from both empirical and theoretical perspectives, highlighting two critical issues beyond the well-documented catastrophic forgetting: Model's ignorance: the single-pass nature of OCL challenges models to learn effective features within constrained training time and storage capacity, leading to a trade-off between effective learning and model throughput; Model's myopia: the local learning nature of OCL on the current task leads the model to adopt overly simplified, task-specific features and excessively sparse classifier, resulting in the gap between the optimal solution for the current task and the global objective. To tackle these issues, we propose the Non-sparse Classifier Evolution framework (NsCE) to facilitate effective global discriminative feature learning with minimal time cost. NsCE integrates non-sparse maximum separation regularization and targeted experience replay techniques with the help of pre-trained models, enabling rapid acquisition of new globally discriminative features.
Authors: Mary Monroe, Anish Thilagar, Melody Hsu, Rafael Frongillo
Abstract: In forecasting competitions, the traditional mechanism scores the predictions of each contestant against the outcome of each event, and the contestant with the highest total score wins. While it is well-known that this traditional mechanism can suffer from incentive issues, it is folklore that contestants will still be roughly truthful as the number of events grows. Yet thus far the literature lacks a formal analysis of this traditional mechanism. This paper gives the first such analysis. We first demonstrate that the ''long-run truthfulness'' folklore is false: even for arbitrary numbers of events, the best forecaster can have an incentive to hedge, reporting more moderate beliefs to increase their win probability. On the positive side, however, we show that two contestants will be approximately truthful when they have sufficient uncertainty over the relative quality of their opponent and the outcomes of the events, a case which may arise in practice.
Authors: Zezhou Wang, Yaxin Du, Zhuzhong Qian, Siheng Chen
Abstract: Federated Domain-specific Instruction Tuning (FedDIT) utilizes limited cross-client private data together with server-side public data for instruction augmentation, ultimately boosting model performance within specific domains. To date, the factors affecting FedDIT remain unclear, and existing instruction augmentation methods primarily focus on the centralized setting without considering distributed environments. Our experiments reveal that the cross-client domain coverage, rather than data heterogeneity, drives model performance in FedDIT. In response, we propose FedDCA, which optimizes domain coverage through greedy client center selection and retrieval-based augmentation. For client-side computational efficiency and system scalability, FedDCA$^*$, the variant of FedDCA, utilizes heterogeneous encoders with server-side feature alignment. Extensive experiments across four distinct domains (code, medical, financial, and mathematical) substantiate the effectiveness of both methods. Additionally, we investigate privacy preservation against memory extraction attacks utilizing various amounts of public data. Results show that there is no significant correlation between the volume of public data and the privacy-preserving capability. However, as the fine-tuning rounds increase, the risk of privacy leakage reduces or converges.
Authors: A. Yark{\i}n Y{\i}ld{\i}z, Asli Kalayci
Abstract: Medical diagnosis is a crucial task in the medical field, in terms of providing accurate classification and respective treatments. Having near-precise decisions based on correct diagnosis can affect a patient's life itself, and may extremely result in a catastrophe if not classified correctly. Several traditional machine learning (ML), such as support vector machines (SVMs) and logistic regression, and state-of-the-art tabular deep learning (DL) methods, including TabNet and TabTransformer, have been proposed and used over tabular medical datasets. Additionally, due to the superior performances, lower computational costs, and easier optimization over different tasks, ensemble methods have been used in the field more recently. They offer a powerful alternative in terms of providing successful medical decision-making processes in several diagnosis tasks. In this study, we investigated the benefits of ensemble methods, especially the Gradient Boosting Decision Tree (GBDT) algorithms in medical classification tasks over tabular data, focusing on XGBoost, CatBoost, and LightGBM. The experiments demonstrate that GBDT methods outperform traditional ML and deep neural network architectures and have the highest average rank over several benchmark tabular medical diagnosis datasets. Furthermore, they require much less computational power compared to DL models, creating the optimal methodology in terms of high performance and lower complexity.
Authors: Elie Attias, Cengiz Pehlevan, Dina Obeid
Abstract: Convolutional Neural Networks (CNNs) excel in many visual tasks, but they tend to be sensitive to slight input perturbations that are imperceptible to the human eye, often resulting in task failures. Recent studies indicate that training CNNs with regularizers that promote brain-like representations, using neural recordings, can improve model robustness. However, the requirement to use neural data severely restricts the utility of these methods. Is it possible to develop regularizers that mimic the computational function of neural regularizers without the need for neural recordings, thereby expanding the usability and effectiveness of these techniques? In this work, we inspect a neural regularizer introduced in Li et al. (2019) to extract its underlying strength. The regularizer uses neural representational similarities, which we find also correlate with pixel similarities. Motivated by this finding, we introduce a new regularizer that retains the essence of the original but is computed using image pixel similarities, eliminating the need for neural recordings. We show that our regularization method 1) significantly increases model robustness to a range of black box attacks on various datasets and 2) is computationally inexpensive and relies only on original datasets. Our work explores how biologically motivated loss functions can be used to drive the performance of artificial neural networks.
Authors: Irene Cannistraci, Emanuele Rodol\`a, Bastian Rieck
Abstract: Deep neural networks often learn similar internal representations, both across different models and within their own layers. While inter-network similarities have enabled techniques such as model stitching and merging, intra-network similarities present new opportunities for designing more efficient architectures. In this paper, we investigate the emergence of these internal similarities across different layers in diverse neural architectures, showing that similarity patterns emerge independently of the datataset used. We introduce a simple metric, Block Redundancy, to detect redundant blocks, providing a foundation for future architectural optimization methods. Building on this, we propose Redundant Blocks Approximation (RBA), a general framework that identifies and approximates one or more redundant computational blocks using simpler transformations. We show that the transformation $\mathcal{T}$ between two representations can be efficiently computed in closed-form, and it is enough to replace the redundant blocks from the network. RBA reduces model parameters and time complexity while maintaining good performance. We validate our method on classification tasks in the vision domain using a variety of pretrained foundational models and datasets.
Authors: Dianzhi Yu, Xinni Zhang, Yankai Chen, Aiwei Liu, Yifei Zhang, Philip S. Yu, Irwin King
Abstract: Continual learning (CL) aims to empower machine learning models to learn continually from new data, while building upon previously acquired knowledge without forgetting. As machine learning models have evolved from small to large pre-trained architectures, and from supporting unimodal to multimodal data, multimodal continual learning (MMCL) methods have recently emerged. The primary challenge of MMCL is that it goes beyond a simple stacking of unimodal CL methods, as such straightforward approaches often yield unsatisfactory performance. In this work, we present the first comprehensive survey on MMCL. We provide essential background knowledge and MMCL settings, as well as a structured taxonomy of MMCL methods. We categorize existing MMCL methods into four categories, i.e., regularization-based, architecture-based, replay-based, and prompt-based methods, explaining their methodologies and highlighting their key innovations. Additionally, to prompt further research in this field, we summarize open MMCL datasets and benchmarks, and discuss several promising future directions for investigation and development. We have also created a GitHub repository for indexing relevant MMCL papers and open resources available at https://github.com/LucyDYu/Awesome-Multimodal-Continual-Learning.
URLs: https://github.com/LucyDYu/Awesome-Multimodal-Continual-Learning.
Authors: Amber Cassimon, Phil Reiter, Siegfried Mercelis, Kevin Mets
Abstract: This paper showcases the use of a reinforcement learning-based Neural Architecture Search (NAS) agent to design a small neural network to perform active fire detection on multispectral satellite imagery. Specifically, we aim to design a neural network that can determine if a single multispectral pixel is a part of a fire, and do so within the constraints of a Low Earth Orbit (LEO) nanosatellite with a limited power budget, to facilitate on-board processing of sensor data. In order to use reinforcement learning, a reward function is needed. We supply this reward function in the shape of a regression model that predicts the F1 score obtained by a particular architecture, following quantization to INT8 precision, from purely architectural features. This model is trained by collecting a random sample of neural network architectures, training these architectures, and collecting their classification performance statistics. Besides the F1 score, we also include the total number of trainable parameters in our reward function to limit the size of the designed model and ensure it fits within the resource constraints imposed by nanosatellite platforms. Finally, we deployed the best neural network to the Google Coral Micro Dev Board and evaluated its inference latency and power consumption. This neural network consists of 1,716 trainable parameters, takes on average 984{\mu}s to inference, and consumes around 800mW to perform inference. These results show that our reinforcement learning-based NAS approach can be successfully applied to novel problems not tackled before.
Authors: Xing Lei, Xuetao Zhang, Zifeng Zhuang, Donglin Wang
Abstract: A novel class of advanced algorithms, termed Goal-Conditioned Weighted Supervised Learning (GCWSL), has recently emerged to tackle the challenges posed by sparse rewards in goal-conditioned reinforcement learning (RL). GCWSL consistently delivers strong performance across a diverse set of goal-reaching tasks due to its simplicity, effectiveness, and stability. However, GCWSL methods lack a crucial capability known as trajectory stitching, which is essential for learning optimal policies when faced with unseen skills during testing. This limitation becomes particularly pronounced when the replay buffer is predominantly filled with sub-optimal trajectories. In contrast, traditional TD-based RL methods, such as Q-learning, which utilize Dynamic Programming, do not face this issue but often experience instability due to the inherent difficulties in value function approximation. In this paper, we propose Q-learning Weighted Supervised Learning (Q-WSL), a novel framework designed to overcome the limitations of GCWSL by incorporating the strengths of Dynamic Programming found in Q-learning. Q-WSL leverages Dynamic Programming results to output the optimal action of (state, goal) pairs across different trajectories within the replay buffer. This approach synergizes the strengths of both Q-learning and GCWSL, effectively mitigating their respective weaknesses and enhancing overall performance. Empirical evaluations on challenging goal-reaching tasks demonstrate that Q-WSL surpasses other goal-conditioned approaches in terms of both performance and sample efficiency. Additionally, Q-WSL exhibits notable robustness in environments characterized by binary reward structures and environmental stochasticity.
Authors: Lise Le Boudec, Emmanuel de Bezenac, Louis Serrano, Ramon Daniel Regueiro-Espino, Yuan Yin, Patrick Gallinari
Abstract: Physics-informed deep learning often faces optimization challenges due to the complexity of solving partial differential equations (PDEs), which involve exploring large solution spaces, require numerous iterations, and can lead to unstable training. These challenges arise particularly from the ill-conditioning of the optimization problem, caused by the differential terms in the loss function. To address these issues, we propose learning a solver, i.e., solving PDEs using a physics-informed iterative algorithm trained on data. Our method learns to condition a gradient descent algorithm that automatically adapts to each PDE instance, significantly accelerating and stabilizing the optimization process and enabling faster convergence of physics-aware models. Furthermore, while traditional physics-informed methods solve for a single PDE instance, our approach addresses parametric PDEs. Specifically, our method integrates the physical loss gradient with the PDE parameters to solve over a distribution of PDE parameters, including coefficients, initial conditions, or boundary conditions. We demonstrate the effectiveness of our method through empirical experiments on multiple datasets, comparing training and test-time optimization performance.
Authors: Seongjin Choi, Zhixiong Jin, Seung Woo Ham, Jiwon Kim, Lijun Sun
Abstract: Deep Generative Models (DGMs) have rapidly advanced in recent years, becoming essential tools in various fields due to their ability to learn complex data distributions and generate synthetic data. Their importance in transportation research is increasingly recognized, particularly for applications like traffic data generation, prediction, and feature extraction. This paper offers a comprehensive introduction and tutorial on DGMs, with a focus on their applications in transportation. It begins with an overview of generative models, followed by detailed explanations of fundamental models, a systematic review of the literature, and practical tutorial code to aid implementation. The paper also discusses current challenges and opportunities, highlighting how these models can be effectively utilized and further developed in transportation research. This paper serves as a valuable reference, guiding researchers and practitioners from foundational knowledge to advanced applications of DGMs in transportation research.
Authors: Samy Wu Fung, Benjamin Berkels
Abstract: Implicit networks are a class of neural networks whose outputs are defined by the fixed point of a parameterized operator. They have enjoyed success in many applications including natural language processing, image processing, and numerous other applications. While they have found abundant empirical success, theoretical work on its generalization is still under-explored. In this work, we consider a large family of implicit networks defined parameterized contractive fixed point operators. We show a generalization bound for this class based on a covering number argument for the Rademacher complexity of these architectures.
Authors: Han Shen, Pin-Yu Chen, Payel Das, Tianyi Chen
Abstract: Fine-tuning on task-specific data to boost downstream performance is a crucial step for leveraging Large Language Models (LLMs). However, previous studies have demonstrated that fine-tuning the models on several adversarial samples or even benign data can greatly comprise the model's pre-equipped alignment and safety capabilities. In this work, we propose SEAL, a novel framework to enhance safety in LLM fine-tuning. SEAL learns a data ranker based on the bilevel optimization to up rank the safe and high-quality fine-tuning data and down rank the unsafe or low-quality ones. Models trained with SEAL demonstrate superior quality over multiple baselines, with 8.5% and 9.7% win rate increase compared to random selection respectively on Llama-3-8b-Instruct and Merlinite-7b models. Our code is available on github https://github.com/hanshen95/SEAL.
Authors: Sambit Panda, Cencheng Shen, Joshua T. Vogelstein
Abstract: Decision forests are widely used for classification and regression tasks. A lesser known property of tree-based methods is that one can construct a proximity matrix from the tree(s), and these proximity matrices are induced kernels. While there has been extensive research on the applications and properties of kernels, there is relatively little research on kernels induced by decision forests. We construct Kernel Mean Embedding Random Forests (KMERF), which induce kernels from random trees and/or forests using leaf-node proximity. We introduce the notion of an asymptotically characteristic kernel, and prove that KMERF kernels are asymptotically characteristic for both discrete and continuous data. Because KMERF is data-adaptive, we suspected it would outperform kernels selected a priori on finite sample data. We illustrate that KMERF nearly dominates current state-of-the-art kernel-based tests across a diverse range of high-dimensional two-sample and independence testing settings. Furthermore, our forest-based approach is interpretable, and provides feature importance metrics that readily distinguish important dimensions, unlike other high-dimensional non-parametric testing procedures. Hence, this work demonstrates the decision forest-based kernel can be more powerful and more interpretable than existing methods, flying in the face of conventional wisdom of the trade-off between the two.
Authors: Brian Zhang, Gabriele Farina, Andrea Celli, Tuomas Sandholm
Abstract: We study the problem of finding optimal correlated equilibria of various sorts in extensive-form games: normal-form coarse correlated equilibrium (NFCCE), extensive-form coarse correlated equilibrium (EFCCE), and extensive-form correlated equilibrium (EFCE). We make two primary contributions. First, we introduce a new algorithm for computing optimal equilibria in all three notions. Its runtime depends exponentially only on a parameter related to the information structure of the game. We also prove a fundamental complexity gap: while our size bounds for NFCCE are similar to those achieved in the case of team games by Zhang et al., this is impossible to achieve for the other two concepts under standard complexity assumptions. Second, we propose a two-sided column generation approach for use when the runtime or memory usage of the previous algorithm is prohibitive. Our algorithm improves upon the one-sided approach of Farina et al. by means of a new decomposition of correlated strategies which allows players to re-optimize their sequence-form strategies with respect to correlation plans which were previously added to the support. Experiments show that our techniques outperform the prior state of the art for computing optimal general-sum correlated equilibria.
Authors: Kaizhong Zheng, Shujian Yu, Baojuan Li, Robert Jenssen, Badong Chen
Abstract: Developing a new diagnostic models based on the underlying biological mechanisms rather than subjective symptoms for psychiatric disorders is an emerging consensus. Recently, machine learning-based classifiers using functional connectivity (FC) for psychiatric disorders and healthy controls are developed to identify brain markers. However, existing machine learning-based diagnostic models are prone to over-fitting (due to insufficient training samples) and perform poorly in new test environment. Furthermore, it is difficult to obtain explainable and reliable brain biomarkers elucidating the underlying diagnostic decisions. These issues hinder their possible clinical applications. In this work, we propose BrainIB, a new graph neural network (GNN) framework to analyze functional magnetic resonance images (fMRI), by leveraging the famed Information Bottleneck (IB) principle. BrainIB is able to identify the most informative edges in the brain (i.e., subgraph) and generalizes well to unseen data. We evaluate the performance of BrainIB against 3 baselines and 7 state-of-the-art brain network classification methods on three psychiatric datasets and observe that our BrainIB always achieves the highest diagnosis accuracy. It also discovers the subgraph biomarkers which are consistent to clinical and neuroimaging findings. The source code and implementation details of BrainIB are freely available at GitHub repository (https://github.com/SJYuCNEL/brain-and-Information-Bottleneck/).
URLs: https://github.com/SJYuCNEL/brain-and-Information-Bottleneck/).
Authors: Antonino Marciano, Emanuele Zappala, Tommaso Torda, Matteo Lulli, Stefano Giagu, Chris Fields, Deen Chen, Filippo Fabrocini
Abstract: Deep Neural Networks miss a principled model of their operation. A novel framework for supervised learning based on Topological Quantum Field Theory that looks particularly well suited for implementation on quantum processors has been recently explored. We propose using this framework to understand the problem of generalisation in Deep Neural Networks. More specifically, in this approach, Deep Neural Networks are viewed as the semi-classical limit of Topological Quantum Neural Networks. A framework of this kind explains the overfitting behavior of Deep Neural Networks during the training step and the corresponding generalisation capabilities. We explore the paradigmatic case of the perceptron, which we implement as the semiclassical limit of Topological Quantum Neural Networks. We apply a novel algorithm we developed, showing that it obtains similar results to standard neural networks, but without the need for training (optimisation).
Authors: Pierre Tholoniat, Kelly Kostopoulou, Mosharaf Chowdhury, Asaf Cidon, Roxana Geambasu, Mathias L\'ecuyer, Junfeng Yang
Abstract: Machine learning (ML) models can leak information about users, and differential privacy (DP) provides a rigorous way to bound that leakage under a given budget. This DP budget can be regarded as a new type of compute resource in workloads of multiple ML models training on user data. Once it is used, the DP budget is forever consumed. Therefore, it is crucial to allocate it most efficiently to train as many models as possible. This paper presents the scheduler for privacy that optimizes for efficiency. We formulate privacy scheduling as a new type of multidimensional knapsack problem, called privacy knapsack, which maximizes DP budget efficiency. We show that privacy knapsack is NP-hard, hence practical algorithms are necessarily approximate. We develop an approximation algorithm for privacy knapsack, DPack, and evaluate it on microbenchmarks and on a new, synthetic private-ML workload we developed from the Alibaba ML cluster trace. We show that DPack: (1) often approaches the efficiency-optimal schedule, (2) consistently schedules more tasks compared to a state-of-the-art privacy scheduling algorithm that focused on fairness (1.3-1.7x in Alibaba, 1.0-2.6x in microbenchmarks), but (3) sacrifices some level of fairness for efficiency. Therefore, using DPack, DP ML operators should be able to train more models on the same amount of user data while offering the same privacy guarantee to their users.
Authors: Chiara Mauri, Stefano Cerri, Oula Puonti, Mark M\"uhlau, Koen Van Leemput
Abstract: Recent years have seen a growing interest in methods for predicting an unknown variable of interest, such as a subject's diagnosis, from medical images depicting its anatomical-functional effects. Methods based on discriminative modeling excel at making accurate predictions, but are challenged in their ability to explain their decisions in anatomically meaningful terms. In this paper, we propose a simple technique for single-subject prediction that is inherently interpretable. It augments the generative models used in classical human brain mapping techniques, in which the underlying cause-effect relations can be encoded, with a multivariate noise model that captures dominant spatial correlations. Experiments demonstrate that the resulting model can be efficiently inverted to make accurate subject-level predictions, while at the same time offering intuitive visual explanations of its inner workings. The method is easy to use: training is fast for typical training set sizes, and only a single hyperparameter needs to be set by the user. Our code is available at https://github.com/chiara-mauri/Interpretable-subject-level-prediction.
URLs: https://github.com/chiara-mauri/Interpretable-subject-level-prediction.
Authors: Iosif Sakos, Antonios Varvitsiotis, Georgios Piliouras
Abstract: Understanding how players adjust their strategies in games, based on their experience, is a crucial tool for policymakers. It enables them to forecast the system's eventual behavior, exert control over the system, and evaluate counterfactual scenarios. The task becomes increasingly difficult when only a limited number of observations are available or difficult to acquire. In this work, we introduce the Side-Information Assisted Regression (SIAR) framework, designed to identify game dynamics in multiplayer normal-form games only using data from a short run of a single system trajectory. To enhance system recovery in the face of scarce data, we integrate side-information constraints into SIAR, which restrict the set of feasible solutions to those satisfying game-theoretic properties and common assumptions about strategic interactions. SIAR is solved using sum-of-squares (SOS) optimization, resulting in a hierarchy of approximations that provably converge to the true dynamics of the system. We showcase that the SIAR framework accurately predicts player behavior across a spectrum of normal-form games, widely-known families of game dynamics, and strong benchmarks, even if the unknown system is chaotic.
Authors: Kevin Kloos, Julian D. Karch, Quinten A. Meertens, Mark de Rooij
Abstract: A quantifier is a supervised machine learning algorithm, focused on estimating the class prevalence in a dataset rather than labeling its individual observations. We introduce Continuous Sweep, a new parametric binary quantifier inspired by the well-performing Median Sweep, which is an ensemble method based on Adjusted Count estimators. We modified two aspects of Median Sweep: 1) using parametric class distributions instead of empirical distributions for the true and false positive rate; 2) using the mean instead of the median of a set of Adjusted Count estimates. These two modifications allow for a theoretical analysis of the bias and variance of Continuous Sweep. Furthermore, the expressions of bias and variance can be used to define optimal decision boundaries of the set of Adjusted count estimates to be used in the ensemble. We show in three simulation studies that Continuous Sweep outperforms the quantifiers in the group Classify, Count, and Correct, including Median Sweep, and is competitive with the two best quantifiers from the group Distribution Matchers. Also an empirical data set is analysed with these quantifiers showing similar performances.
Authors: Hossein Shakibania, Sina Raoufi, Hassan Khotanlou
Abstract: Low-light images, characterized by inadequate illumination, pose challenges of diminished clarity, muted colors, and reduced details. Low-light image enhancement, an essential task in computer vision, aims to rectify these issues by improving brightness, contrast, and overall perceptual quality, thereby facilitating accurate analysis and interpretation. This paper introduces the Convolutional Dense Attention-guided Network (CDAN), a novel solution for enhancing low-light images. CDAN integrates an autoencoder-based architecture with convolutional and dense blocks, complemented by an attention mechanism and skip connections. This architecture ensures efficient information propagation and feature learning. Furthermore, a dedicated post-processing phase refines color balance and contrast. Our approach demonstrates notable progress compared to state-of-the-art results in low-light image enhancement, showcasing its robustness across a wide range of challenging scenarios. Our model performs remarkably on benchmark datasets, effectively mitigating under-exposure and proficiently restoring textures and colors in diverse low-light scenarios. This achievement underscores CDAN's potential for diverse computer vision tasks, notably enabling robust object detection and recognition in challenging low-light conditions.
Authors: Weirui Ye, Yunsheng Zhang, Haoyang Weng, Xianfan Gu, Shengjie Wang, Tong Zhang, Mengchen Wang, Pieter Abbeel, Yang Gao
Abstract: Reinforcement learning (RL) is a promising approach for solving robotic manipulation tasks. However, it is challenging to apply the RL algorithms directly in the real world. For one thing, RL is data-intensive and typically requires millions of interactions with environments, which are impractical in real scenarios. For another, it is necessary to make heavy engineering efforts to design reward functions manually. To address these issues, we leverage foundation models in this paper. We propose Reinforcement Learning with Foundation Priors (RLFP) to utilize guidance and feedback from policy, value, and success-reward foundation models. Within this framework, we introduce the Foundation-guided Actor-Critic (FAC) algorithm, which enables embodied agents to explore more efficiently with automatic reward functions. The benefits of our framework are threefold: (1) \textit{sample efficient}; (2) \textit{minimal and effective reward engineering}; (3) \textit{agnostic to foundation model forms and robust to noisy priors}. Our method achieves remarkable performances in various manipulation tasks on both real robots and in simulation. Across 5 dexterous tasks with real robots, FAC achieves an average success rate of 86\% after one hour of real-time learning. Across 8 tasks in the simulated Meta-world, FAC achieves 100\% success rates in 7/8 tasks under less than 100k frames (about 1-hour training), outperforming baseline methods with manual-designed rewards in 1M frames. We believe the RLFP framework can enable future robots to explore and learn autonomously in the physical world for more tasks. Visualizations and code are available at \url{https://yewr.github.io/rlfp}.
Authors: Bahareh Nikpour, Narges Armanfard
Abstract: Attention mechanisms have exhibited promising potential in enhancing learning models by identifying salient portions of input data. This is particularly valuable in scenarios where limited training samples are accessible due to challenges in data collection and labeling. Drawing inspiration from human recognition processes, we posit that an AI baseline's performance could be more accurate and dependable if it is exposed to essential segments of raw data rather than the entire input dataset, akin to human perception. However, the task of selecting these informative data segments, referred to as hard attention finding, presents a formidable challenge. In situations with few training samples, existing studies struggle to locate such informative regions due to the large number of training parameters that cannot be effectively learned from the available limited samples. In this study, we introduce a novel and practical framework for achieving explainable hard attention finding, specifically tailored for few-shot learning scenarios, called FewXAT. Our approach employs deep reinforcement learning to implement the concept of hard attention, directly impacting raw input data and thus rendering the process interpretable for human understanding. Through extensive experimentation across various benchmark datasets, we demonstrate the efficacy of our proposed method.
Authors: Mohammad Dehghani, Mobin Mohammadi, Diyana Tehrany Dehkordy
Abstract: It is crucial for emergency physicians to identify patients at higher risk of mortality to effectively prioritize hospital resources, particularly in regions with limited medical services. This became even more critical during global pandemics, which have disrupted lives in unprecedented ways and caused widespread morbidity and mortality. The collected data from patients is beneficial to predict the outcome, although there is a question about which data makes the most accurate predictions. Therefore, this study aimed to achieve two main objectives during the pandemic, using data and experiments from the most recent global health crisis, COVID-19. First, we want to examine whether deep learning algorithms can predict a patient's morality. Second, we investigated the impact of Clinical and RT-PCR on prediction to determine which one is more reliable. We defined four stages with different feature sets and used 9 machine learning and deep learning methods to build appropriate model. Based on results, the deep neural decision forest, as an interpretable deep learning methods, performed the best across all stages and proved its capability to predict the recovery and death of patients. Additionally, results indicate that Clinical alone (without the use of RT-PCR) is the most effective method of diagnosis, with an accuracy of 80%. This study can provide guidance for medical professionals in the event of a crisis or outbreak similar to COVID-19. Moreover, the proposed deep learning method demonstrates exceptional suitability for mortality prediction.
Authors: Rongxiang Wang, Felix Xiaozhu Lin
Abstract: Modern speech understanding (SU) runs a sophisticated pipeline: ingesting streaming voice input, the pipeline executes encoder-decoder based deep neural networks repeatedly; by doing so, the pipeline generates tentative outputs (called hypotheses), and periodically scores the hypotheses. This paper sets to accelerate SU on resource-constrained edge devices. It takes a hybrid approach: to speed up on-device execution; to offload inputs that are beyond the device's capacity. While the approach is well-known, we address SU's unique challenges with novel techniques: (1) late contextualization, which executes a model's attentive encoder in parallel to the input ingestion; (2) pilot inference, which mitigates the SU pipeline's temporal load imbalance; (3) autoregression offramps, which evaluate offloading decisions based on pilot inferences and hypotheses. Our techniques are compatible with existing speech models, pipelines, and frameworks; they can be applied independently or in combination. Our prototype, called PASU, is tested on Arm platforms with 6 - 8 cores: it delivers SOTA accuracy; it reduces the end-to-end latency by 2x and reduces the offloading needs by 2x.
Authors: Jiwon Kim, Byeongho Heo, Sangdoo Yun, Seungryong Kim, Dongyoon Han
Abstract: Semantic correspondence methods have advanced to obtaining high-quality correspondences employing complicated networks, aiming to maximize the model capacity. However, despite the performance improvements, they may remain constrained by the scarcity of training keypoint pairs, a consequence of the limited training images and the sparsity of keypoints. This paper builds on the hypothesis that there is an inherent data-hungry matter in learning semantic correspondences and uncovers the models can be more trained by employing densified training pairs. We demonstrate a simple machine annotator reliably enriches paired key points via machine supervision, requiring neither extra labeled key points nor trainable modules from unlabeled images. Consequently, our models surpass current state-of-the-art models on semantic correspondence learning benchmarks like SPair-71k, PF-PASCAL, and PF-WILLOW and enjoy further robustness on corruption benchmarks. Our code is available at https://github.com/naver-ai/matchme.
Authors: Carles Domingo-Enrich, Jiequn Han, Brandon Amos, Joan Bruna, Ricky T. Q. Chen
Abstract: Stochastic optimal control, which has the goal of driving the behavior of noisy systems, is broadly applicable in science, engineering and artificial intelligence. Our work introduces Stochastic Optimal Control Matching (SOCM), a novel Iterative Diffusion Optimization (IDO) technique for stochastic optimal control that stems from the same philosophy as the conditional score matching loss for diffusion models. That is, the control is learned via a least squares problem by trying to fit a matching vector field. The training loss, which is closely connected to the cross-entropy loss, is optimized with respect to both the control function and a family of reparameterization matrices which appear in the matching vector field. The optimization with respect to the reparameterization matrices aims at minimizing the variance of the matching vector field. Experimentally, our algorithm achieves lower error than all the existing IDO techniques for stochastic optimal control for three out of four control problems, in some cases by an order of magnitude. The key idea underlying SOCM is the path-wise reparameterization trick, a novel technique that may be of independent interest. Code at https://github.com/facebookresearch/SOC-matching
Authors: Haoyu Wei, Runzhe Wan, Lei Shi, Rui Song
Abstract: Many real applications of bandits have sparse non-zero rewards, leading to slow learning speed. Using problem-specific structures for careful distribution modeling is known as critical to estimation efficiency in statistics, yet is under-explored in bandits. We initiate the study of zero-inflated bandits, where the reward is modeled as a classic semi-parametric distribution called zero-inflated distribution. We design Upper Confidence Bound- and Thompson Sampling-type algorithms for this specific structure. We derive the regret bounds under both multi-armed bandits with general reward assumptions and contextual generalized linear bandit with sub-Gaussian rewards. In many settings, the regret rates of our algorithms are either minimax optimal or state-of-the-art. The superior empirical performance of our methods is demonstrated via numerical studies.
Authors: Nathan Painchaud, J\'er\'emie Stym-Popper, Pierre-Yves Courand, Nicolas Thome, Pierre-Marc Jodoin, Nicolas Duchateau, Olivier Bernard
Abstract: Deep learning enables automatic and robust extraction of cardiac function descriptors from echocardiographic sequences, such as ejection fraction or strain. These descriptors provide fine-grained information that physicians consider, in conjunction with more global variables from the clinical record, to assess patients' condition. Drawing on novel transformer models applied to tabular data, we propose a method that considers all descriptors extracted from medical records and echocardiograms to learn the representation of a cardiovascular pathology with a difficult-to-characterize continuum, namely hypertension. Our method first projects each variable into its own representation space using modality-specific approaches. These standardized representations of multimodal data are then fed to a transformer encoder, which learns to merge them into a comprehensive representation of the patient through the task of predicting a clinical rating. This stratification task is formulated as an ordinal classification to enforce a pathological continuum in the representation space. We observe the major trends along this continuum on a cohort of 239 hypertensive patients, providing unprecedented details in the description of hypertension's impact on various cardiac function descriptors. Our analysis shows that i) the XTab foundation model's architecture allows to reach outstanding performance (98% AUROC) even with limited data (less than 200 training samples), ii) stratification across the population is reproducible between trainings (within 3.6% MAE), and iii) patterns emerge in descriptors, some of which align with established physiological knowledge about hypertension, while others could pave the way for a more comprehensive understanding of this pathology.
Authors: Ollie Liu, Deqing Fu, Dani Yogatama, Willie Neiswanger
Abstract: The potential of large language models (LLMs) as decision support tools is increasingly being explored in fields such as business, engineering, and medicine, which often face challenging tasks of decision-making under uncertainty. In this paper, we show that directly prompting LLMs on these types of decision-making problems can yield poor results, especially as the problem complexity increases. To aid in these tasks, we propose DeLLMa (Decision-making Large Language Model assistant), a framework designed to enhance decision-making accuracy in uncertain environments. DeLLMa involves a multi-step reasoning procedure that integrates recent best practices in scaling inference-time reasoning, drawing upon principles from decision theory and utility theory, to provide an accurate and human-auditable decision-making process. We validate our procedure on multiple realistic decision-making environments, demonstrating that DeLLMa can consistently enhance the decision-making performance of leading language models, and achieve up to a 40% increase in accuracy over competing methods. Additionally, we show how performance improves when scaling compute at test time, and carry out human evaluations to benchmark components of DeLLMa.
Authors: Abdoulaye Sakho (LPSM), Emmanuel Malherbe (LPSM), Erwan Scornet (LPSM)
Abstract: Synthetic Minority Oversampling Technique (SMOTE) is a common rebalancing strategy for handling imbalanced tabular data sets. However, few works analyze SMOTE theoretically. In this paper, we prove that SMOTE (with default parameter) tends to copy the original minority samples asymptotically. We also prove that SMOTE exhibits boundary artifacts, thus justifying existing SMOTE variants. Then we introduce two new SMOTE-related strategies, and compare them with state-of-the-art rebalancing procedures. Surprisingly, for most data sets, we observe that applying no rebalancing strategy is competitive in terms of predictive performances, with tuned random forests, logistic regression or LightGBM. For highly imbalanced data sets, our new methods, named CV-SMOTE and Multivariate Gaussian SMOTE, are competitive. Besides, our analysis sheds some lights on the behavior of common rebalancing strategies, when used in conjunction with random forests.
Authors: Junyou Li, Qin Zhang, Yangbin Yu, Qiang Fu, Deheng Ye
Abstract: We find that, simply via a sampling-and-voting method, the performance of large language models (LLMs) scales with the number of agents instantiated. Also, this method, termed as Agent Forest, is orthogonal to existing complicated methods to further enhance LLMs, while the degree of enhancement is correlated to the task difficulty. We conduct comprehensive experiments on a wide range of LLM benchmarks to verify the presence of our finding, and to study the properties that can facilitate its occurrence. Our code is publicly available at: https://github.com/MoreAgentsIsAllYouNeed/AgentForest
Authors: Nutan Chen, Botond Cseke, Elie Aljalbout, Alexandros Paraschos, Marvin Alles, Patrick van der Smagt
Abstract: We present a novel motion generation approach for robot arms, with high degrees of freedom, in complex settings that can adapt online to obstacles or new via points. Learning from Demonstration facilitates rapid adaptation to new tasks and optimizes the utilization of accumulated expertise by allowing robots to learn and generalize from demonstrated trajectories. We train a transformer architecture, based on conditional variational autoencoder, on a large dataset of simulated trajectories used as demonstrations. Our architecture learns essential motion generation skills from these demonstrations and is able to adapt them to meet auxiliary tasks. Additionally, our approach implements auto-regressive motion generation to enable real-time adaptations, as, for example, introducing or changing via-points, and velocity and acceleration constraints. Using beam search, we present a method for further adaption of our motion generator to avoid obstacles. We show that our model successfully generates motion from different initial and target points and that is capable of generating trajectories that navigate complex tasks across different robotic platforms.
Authors: Zezhi Wang, Jin Zhu, Peng Chen, Huiyang Peng, Xiaoke Zhang, Anran Wang, Junxian Zhu, Xueqin Wang
Abstract: Applying iterative solvers on sparsity-constrained optimization (SCO) requires tedious mathematical deduction and careful programming/debugging that hinders these solvers' broad impact. In the paper, the library skscope is introduced to overcome such an obstacle. With skscope, users can solve the SCO by just programming the objective function. The convenience of skscope is demonstrated through two examples in the paper, where sparse linear regression and trend filtering are addressed with just four lines of code. More importantly, skscope's efficient implementation allows state-of-the-art solvers to quickly attain the sparse solution regardless of the high dimensionality of parameter space. Numerical experiments reveal the available solvers in skscope can achieve up to 80x speedup on the competing relaxation solutions obtained via the benchmarked convex solver. skscope is published on the Python Package Index (PyPI) and Conda, and its source code is available at: https://github.com/abess-team/skscope.
Authors: Gianluca Barone, Aashrit Cunchala, Rudy Nunez
Abstract: Standard classification theory assumes that the distribution of images in the test and training sets are identical. Unfortunately, real-life scenarios typically feature unseen data (``out-of-distribution data") which is different from data in the training distribution (``in-distribution"). This issue is most prevalent in social justice problems where data from under-represented groups may appear in the test data without representing an equal proportion of the training data. This may result in a model returning confidently wrong decisions and predictions. We are interested in the following question: Can the performance of a neural network improve on facial images of out-of-distribution data when it is trained simultaneously on multiple datasets of in-distribution data? We approach this problem by incorporating the Outlier Exposure model and investigate how the model's performance changes when other datasets of facial images were implemented. We observe that the accuracy and other metrics of the model can be increased by applying Outlier Exposure, incorporating a trainable weight parameter to increase the machine's emphasis on outlier images, and by re-weighting the importance of different class labels. We also experimented with whether sorting the images and determining outliers via image features would have more of an effect on the metrics than sorting by average pixel value, and found no conclusive results. Our goal was to make models not only more accurate but also more fair by scanning a more expanded range of images. Utilizing Python and the Pytorch package, we found models utilizing outlier exposure could result in more fair classification.
Authors: SIMA Team, Maria Abi Raad, Arun Ahuja, Catarina Barros, Frederic Besse, Andrew Bolt, Adrian Bolton, Bethanie Brownfield, Gavin Buttimore, Max Cant, Sarah Chakera, Stephanie C. Y. Chan, Jeff Clune, Adrian Collister, Vikki Copeman, Alex Cullum, Ishita Dasgupta, Dario de Cesare, Julia Di Trapani, Yani Donchev, Emma Dunleavy, Martin Engelcke, Ryan Faulkner, Frankie Garcia, Charles Gbadamosi, Zhitao Gong, Lucy Gonzales, Kshitij Gupta, Karol Gregor, Arne Olav Hallingstad, Tim Harley, Sam Haves, Felix Hill, Ed Hirst, Drew A. Hudson, Jony Hudson, Steph Hughes-Fitt, Danilo J. Rezende, Mimi Jasarevic, Laura Kampis, Rosemary Ke, Thomas Keck, Junkyung Kim, Oscar Knagg, Kavya Kopparapu, Rory Lawton, Andrew Lampinen, Shane Legg, Alexander Lerchner, Marjorie Limont, Yulan Liu, Maria Loks-Thompson, Joseph Marino, Kathryn Martin Cussons, Loic Matthey, Siobhan Mcloughlin, Piermaria Mendolicchio, Hamza Merzic, Anna Mitenkova, Alexandre Moufarek, Valeria Oliveira, Yanko Oliveira, Hannah Openshaw, Renke Pan, Aneesh Pappu, Alex Platonov, Ollie Purkiss, David Reichert, John Reid, Pierre Harvey Richemond, Tyson Roberts, Giles Ruscoe, Jaume Sanchez Elias, Tasha Sandars, Daniel P. Sawyer, Tim Scholtes, Guy Simmons, Daniel Slater, Hubert Soyer, Heiko Strathmann, Peter Stys, Allison C. Tam, Denis Teplyashin, Tayfun Terzi, Davide Vercelli, Bojan Vujatovic, Marcus Wainwright, Jane X. Wang, Zhengdong Wang, Daan Wierstra, Duncan Williams, Nathaniel Wong, Sarah York, Nick Young
Abstract: Building embodied AI systems that can follow arbitrary language instructions in any 3D environment is a key challenge for creating general AI. Accomplishing this goal requires learning to ground language in perception and embodied actions, in order to accomplish complex tasks. The Scalable, Instructable, Multiworld Agent (SIMA) project tackles this by training agents to follow free-form instructions across a diverse range of virtual 3D environments, including curated research environments as well as open-ended, commercial video games. Our goal is to develop an instructable agent that can accomplish anything a human can do in any simulated 3D environment. Our approach focuses on language-driven generality while imposing minimal assumptions. Our agents interact with environments in real-time using a generic, human-like interface: the inputs are image observations and language instructions and the outputs are keyboard-and-mouse actions. This general approach is challenging, but it allows agents to ground language across many visually complex and semantically rich environments while also allowing us to readily run agents in new environments. In this paper we describe our motivation and goal, the initial progress we have made, and promising preliminary results on several diverse research environments and a variety of commercial video games.
Authors: Gautham Vinod, Jiangpeng He, Zeman Shao, Fengqing Zhu
Abstract: Image-based methods to analyze food images have alleviated the user burden and biases associated with traditional methods. However, accurate portion estimation remains a major challenge due to the loss of 3D information in the 2D representation of foods captured by smartphone cameras or wearable devices. In this paper, we propose a new framework to estimate both food volume and energy from 2D images by leveraging the power of 3D food models and physical reference in the eating scene. Our method estimates the pose of the camera and the food object in the input image and recreates the eating occasion by rendering an image of a 3D model of the food with the estimated poses. We also introduce a new dataset, SimpleFood45, which contains 2D images of 45 food items and associated annotations including food volume, weight, and energy. Our method achieves an average error of 31.10 kCal (17.67%) on this dataset, outperforming existing portion estimation methods. The dataset can be accessed at: https://lorenz.ecn.purdue.edu/~gvinod/simplefood45/ and the code can be accessed at: https://gitlab.com/viper-purdue/monocular-food-volume-3d
URLs: https://lorenz.ecn.purdue.edu/, https://gitlab.com/viper-purdue/monocular-food-volume-3d
Authors: Ryan Thompson, Edwin V. Bonilla, Robert Kohn
Abstract: Directed acyclic graph (DAG) learning is a rapidly expanding field of research. Though the field has witnessed remarkable advances over the past few years, it remains statistically and computationally challenging to learn a single (point estimate) DAG from data, let alone provide uncertainty quantification. Our article addresses the difficult task of quantifying graph uncertainty by developing a Bayesian variational inference framework based on novel distributions that have support directly on the space of DAGs. The distributions, which we use to form our prior and variational posterior, are induced by a projection operation, whereby an arbitrary continuous distribution is projected onto the space of sparse weighted acyclic adjacency matrices (matrix representations of DAGs) with probability mass on exact zeros. Though the projection constitutes a combinatorial optimization problem, it is solvable at scale via recently developed techniques that reformulate acyclicity as a continuous constraint. We empirically demonstrate that our method, ProDAG, can deliver accurate inference and often outperforms existing state-of-the-art alternatives.
Authors: Zebin You, Xinyu Zhang, Hanzhong Guo, Jingdong Wang, Chongxuan Li
Abstract: The ultimate goal of generative models is to perfectly capture the data distribution. For image generation, common metrics of visual quality (e.g., FID) and the perceived truthfulness of generated images seem to suggest that we are nearing this goal. However, through distribution classification tasks, we reveal that, from the perspective of neural network-based classifiers, even advanced diffusion models are still far from this goal. Specifically, classifiers are able to consistently and effortlessly distinguish real images from generated ones across various settings. Moreover, we uncover an intriguing discrepancy: classifiers can easily differentiate between diffusion models with comparable performance (e.g., U-ViT-H vs. DiT-XL), but struggle to distinguish between models within the same family but of different scales (e.g., EDM2-XS vs. EDM2-XXL). Our methodology carries several important implications. First, it naturally serves as a diagnostic tool for diffusion models by analyzing specific features of generated data. Second, it sheds light on the model autophagy disorder and offers insights into the use of generated data: augmenting real data with generated data is more effective than replacing it.
Authors: John Hawkins, David Tivey
Abstract: Identifying critical research within the growing body of academic work is an intrinsic aspect of conducting quality research. Systematic review processes used in evidence-based medicine formalise this as a procedure that must be followed in a research program. However, it comes with an increasing burden in terms of the time required to identify the important articles of research for a given topic. In this work, we develop a method for building a general-purpose filtering system that matches a research question, posed as a natural language description of the required content, against a candidate set of articles obtained via the application of broad search terms. Our results demonstrate that transformer models, pre-trained on biomedical literature, and then fine tuned for the specific task, offer a promising solution to this problem. The model can remove large volumes of irrelevant articles for most research questions. Furthermore, analysis of the specific research questions in our training data suggest natural avenues for further improvement.
Authors: Ruichen Jiang, Devyani Maladkar, Aryan Mokhtari
Abstract: Adaptive gradient methods, such as AdaGrad, are among the most successful optimization algorithms for neural network training. While these methods are known to achieve better dimensional dependence than stochastic gradient descent (SGD) under favorable geometry for stochastic convex optimization, the theoretical justification for their success in stochastic non-convex optimization remains elusive. In fact, under standard assumptions of Lipschitz gradients and bounded noise variance, it is known that SGD is worst-case optimal (up to absolute constants) in terms of finding a near-stationary point with respect to the $\ell_2$-norm, making further improvements impossible. Motivated by this limitation, we introduce refined assumptions on the smoothness structure of the objective and the gradient noise variance, which better suit the coordinate-wise nature of adaptive gradient methods. Moreover, we adopt the $\ell_1$-norm of the gradient as the stationarity measure, as opposed to the standard $\ell_2$-norm, to align with the coordinate-wise analysis and obtain tighter convergence guarantees for AdaGrad. Under these new assumptions and the $\ell_1$-norm stationarity measure, we establish an upper bound on the convergence rate of AdaGrad and a corresponding lower bound for SGD. In particular, for certain configurations of problem parameters, we show that the iteration complexity of AdaGrad outperforms SGD by a factor of $d$. To the best of our knowledge, this is the first result to demonstrate a provable gain of adaptive gradient methods over SGD in a non-convex setting. We also present supporting lower bounds, including one specific to AdaGrad and one applicable to general deterministic first-order methods, showing that our upper bound for AdaGrad is tight and unimprovable up to a logarithmic factor under certain conditions.
Authors: Yifei Chen, Zhu Zhu, Shenghao Zhu, Linwei Qiu, Binfeng Zou, Fan Jia, Yunpeng Zhu, Chenyan Zhang, Zhaojie Fang, Feiwei Qin, Jin Fan, Changmiao Wang, Yu Gao, Gang Yu
Abstract: The incidence and mortality rates of malignant tumors, such as acute leukemia, have risen significantly. Clinically, hospitals rely on cytological examination of peripheral blood and bone marrow smears to diagnose malignant tumors, with accurate blood cell counting being crucial. Existing automated methods face challenges such as low feature expression capability, poor interpretability, and redundant feature extraction when processing high-dimensional microimage data. We propose a novel fine-grained classification model, SCKansformer, for bone marrow blood cells, which addresses these challenges and enhances classification accuracy and efficiency. The model integrates the Kansformer Encoder, SCConv Encoder, and Global-Local Attention Encoder. The Kansformer Encoder replaces the traditional MLP layer with the KAN, improving nonlinear feature representation and interpretability. The SCConv Encoder, with its Spatial and Channel Reconstruction Units, enhances feature representation and reduces redundancy. The Global-Local Attention Encoder combines Multi-head Self-Attention with a Local Part module to capture both global and local features. We validated our model using the Bone Marrow Blood Cell Fine-Grained Classification Dataset (BMCD-FGCD), comprising over 10,000 samples and nearly 40 classifications, developed with a partner hospital. Comparative experiments on our private dataset, as well as the publicly available PBC and ALL-IDB datasets, demonstrate that SCKansformer outperforms both typical and advanced microcell classification methods across all datasets. Our source code and private BMCD-FGCD dataset are available at https://github.com/JustlfC03/SCKansformer.
Authors: Navid Rajabi, Jana Kosecka
Abstract: The ability to understand and reason about spatial relationships between objects in images is an important component of visual reasoning. This skill rests on the ability to recognize and localize objects of interest and determine their spatial relation. Early vision and language models (VLMs) have been shown to struggle to recognize spatial relations. We extend the previously released What'sUp dataset and propose a novel comprehensive evaluation for spatial relationship understanding that highlights the strengths and weaknesses of 27 different models. In addition to the VLMs evaluated in What'sUp, our extensive evaluation encompasses 3 classes of Multimodal LLMs (MLLMs) that vary in their parameter sizes (ranging from 7B to 110B), training/instruction-tuning methods, and visual resolution to benchmark their performances and scrutinize the scaling laws in this task.
Authors: Michael Wornow, Avanika Narayan, Ben Viggiano, Ishan S. Khare, Tathagat Verma, Tibor Thompson, Miguel Angel Fuentes Hernandez, Sudharsan Sundar, Chloe Trujillo, Krrish Chawla, Rongfei Lu, Justin Shen, Divya Nagaraj, Joshua Martinez, Vardhan Agrawal, Althea Hudson, Nigam H. Shah, Christopher Re
Abstract: Existing ML benchmarks lack the depth and diversity of annotations needed for evaluating models on business process management (BPM) tasks. BPM is the practice of documenting, measuring, improving, and automating enterprise workflows. However, research has focused almost exclusively on one task - full end-to-end automation using agents based on multimodal foundation models (FMs) like GPT-4. This focus on automation ignores the reality of how most BPM tools are applied today - simply documenting the relevant workflow takes 60% of the time of the typical process optimization project. To address this gap we present WONDERBREAD, the first benchmark for evaluating multimodal FMs on BPM tasks beyond automation. Our contributions are: (1) a dataset containing 2928 documented workflow demonstrations; (2) 6 novel BPM tasks sourced from real-world applications ranging from workflow documentation to knowledge transfer to process improvement; and (3) an automated evaluation harness. Our benchmark shows that while state-of-the-art FMs can automatically generate documentation (e.g. recalling 88% of the steps taken in a video demonstration of a workflow), they struggle to re-apply that knowledge towards finer-grained validation of workflow completion (F1 < 0.3). We hope WONDERBREAD encourages the development of more "human-centered" AI tooling for enterprise applications and furthers the exploration of multimodal FMs for the broader universe of BPM tasks. We publish our dataset and experiments here: https://github.com/HazyResearch/wonderbread
Authors: Sungbin Shin, Wonpyo Park, Jaeho Lee, Namhoon Lee
Abstract: This work suggests fundamentally rethinking the current practice of pruning large language models (LLMs). The way it is done is by divide and conquer: split the model into submodels, sequentially prune them, and reconstruct predictions of the dense counterparts on small calibration data one at a time; the final model is obtained simply by putting the resulting sparse submodels together. While this approach enables pruning under memory constraints, it generates high reconstruction errors. In this work, we first present an array of reconstruction techniques that can significantly reduce this error by more than $90\%$. Unwittingly, however, we discover that minimizing reconstruction error is not always ideal and can overfit the given calibration data, resulting in rather increased language perplexity and poor performance at downstream tasks. We find out that a strategy of self-generating calibration data can mitigate this trade-off between reconstruction and generalization, suggesting new directions in the presence of both benefits and pitfalls of reconstruction for pruning LLMs.
Authors: Boyi Wei, Weijia Shi, Yangsibo Huang, Noah A. Smith, Chiyuan Zhang, Luke Zettlemoyer, Kai Li, Peter Henderson
Abstract: Language models (LMs) derive their capabilities from extensive training on diverse data, including potentially copyrighted material. These models can memorize and generate content similar to their training data, posing potential concerns. Therefore, model creators are motivated to develop mitigation methods that prevent generating protected content. We term this procedure as copyright takedowns for LMs, noting the conceptual similarity to (but legal distinction from) the DMCA takedown This paper introduces the first evaluation of the feasibility and side effects of copyright takedowns for LMs. We propose CoTaEval, an evaluation framework to assess the effectiveness of copyright takedown methods, the impact on the model's ability to retain uncopyrightable factual knowledge from the training data whose recitation is embargoed, and how well the model maintains its general utility and efficiency. We examine several strategies, including adding system prompts, decoding-time filtering interventions, and unlearning approaches. Our findings indicate that no tested method excels across all metrics, showing significant room for research in this unique problem setting and indicating potential unresolved challenges for live policy proposals.
Authors: Sheridan Feucht, David Atkinson, Byron Wallace, David Bau
Abstract: LLMs process text as sequences of tokens that roughly correspond to words, where less common words are represented by multiple tokens. However, individual tokens are often semantically unrelated to the meanings of the words/concepts they comprise. For example, Llama-2-7b's tokenizer splits the word "northeastern" into the tokens ['_n', 'ort', 'he', 'astern'], none of which correspond to semantically meaningful units like "north" or "east." Similarly, the overall meanings of named entities like "Neil Young" and multi-word expressions like "break a leg" cannot be directly inferred from their constituent tokens. Mechanistically, how do LLMs convert such arbitrary groups of tokens into useful higher-level representations? In this work, we find that last token representations of named entities and multi-token words exhibit a pronounced "erasure" effect, where information about previous and current tokens is rapidly forgotten in early layers. Using this observation, we propose a method to "read out" the implicit vocabulary of an autoregressive LLM by examining differences in token representations across layers, and present results of this method for Llama-2-7b and Llama-3-8B. To our knowledge, this is the first attempt to probe the implicit vocabulary of an LLM.
Authors: Khai-Nguyen Nguyen, Khai Le-Duc, Bach Phan Tat, Duy Le, Long Vo-Dang, Truong-Son Hy
Abstract: Transparency in AI healthcare decision-making is crucial for building trust among AI and users. Incorporating reasoning capabilities enables Large Language Models (LLMs) to understand emotions in context, handle nuanced language, and infer unstated sentiments. In this work, we introduce a new task -- Sentiment Reasoning -- for both speech and text modalities, along with our proposed multimodal multitask framework and dataset. Sentiment Reasoning is an auxiliary task in sentiment analysis where the model predicts both the sentiment label and generates the rationale behind it based on the input transcript. Our study conducted on both human transcripts and Automatic Speech Recognition (ASR) transcripts shows that Sentiment Reasoning helps improve model transparency by providing rationale for model prediction with quality semantically comparable to humans while also improving model performance (1% increase in both accuracy and macro-F1) via rationale-augmented fine-tuning. Also, no significant difference in the semantic quality of generated rationales between human and ASR transcripts. All code, data (English-translated and Vietnamese) and models are published online: https://github.com/leduckhai/MultiMed.
Authors: Karl El Hajal, Ajinkya Kulkarni, Enno Hermann, Mathew Magimai. -Doss
Abstract: While recent zero-shot multispeaker text-to-speech (TTS) models achieve impressive results, they typically rely on extensive transcribed speech datasets from numerous speakers and intricate training pipelines. Meanwhile, self-supervised learning (SSL) speech features have emerged as effective intermediate representations for TTS. It was also observed that SSL features from different speakers that are linearly close share phonetic information while maintaining individual speaker identity, which enables straight-forward and robust voice cloning. In this study, we introduce SSL-TTS, a lightweight and efficient zero-shot TTS framework trained on transcribed speech from a single speaker. SSL-TTS leverages SSL features and retrieval methods for simple and robust zero-shot multi-speaker synthesis. Objective and subjective evaluations show that our approach achieves performance comparable to state-of-the-art models that require significantly larger training datasets. The low training data requirements mean that SSL-TTS is well suited for the development of multi-speaker TTS systems for low-resource domains and languages. We also introduce an interpolation parameter which enables fine control over the output speech by blending voices. Demo samples are available at https://idiap.github.io/ssl-tts
Authors: Qiyao Liang, Ziming Liu, Mitchell Ostrow, Ila Fiete
Abstract: Diffusion models are capable of generating photo-realistic images that combine elements which likely do not appear together in the training set, demonstrating the ability to \textit{compositionally generalize}. Nonetheless, the precise mechanism of compositionality and how it is acquired through training remains elusive. Inspired by cognitive neuroscientific approaches, we consider a highly reduced setting to examine whether and when diffusion models learn semantically meaningful and factorized representations of composable features. We performed extensive controlled experiments on conditional Denoising Diffusion Probabilistic Models (DDPMs) trained to generate various forms of 2D Gaussian bump images. We found that the models learn factorized but not fully continuous manifold representations for encoding continuous features of variation underlying the data. With such representations, models demonstrate superior feature compositionality but limited ability to interpolate over unseen values of a given feature. Our experimental results further demonstrate that diffusion models can attain compositionality with few compositional examples, suggesting a more efficient way to train DDPMs. Finally, we connect manifold formation in diffusion models to percolation theory in physics, offering insight into the sudden onset of factorized representation learning. Our thorough toy experiments thus contribute a deeper understanding of how diffusion models capture compositional structure in data.
Authors: Li-Heng Lin, Yuchen Cui, Amber Xie, Tianyu Hua, Dorsa Sadigh
Abstract: Few-shot imitation learning relies on only a small amount of task-specific demonstrations to efficiently adapt a policy for a given downstream tasks. Retrieval-based methods come with a promise of retrieving relevant past experiences to augment this target data when learning policies. However, existing data retrieval methods fall under two extremes: they either rely on the existence of exact behaviors with visually similar scenes in the prior data, which is impractical to assume; or they retrieve based on semantic similarity of high-level language descriptions of the task, which might not be that informative about the shared low-level behaviors or motions across tasks that is often a more important factor for retrieving relevant data for policy learning. In this work, we investigate how we can leverage motion similarity in the vast amount of cross-task data to improve few-shot imitation learning of the target task. Our key insight is that motion-similar data carries rich information about the effects of actions and object interactions that can be leveraged during few-shot adaptation. We propose FlowRetrieval, an approach that leverages optical flow representations for both extracting similar motions to target tasks from prior data, and for guiding learning of a policy that can maximally benefit from such data. Our results show FlowRetrieval significantly outperforms prior methods across simulated and real-world domains, achieving on average 27% higher success rate than the best retrieval-based prior method. In the Pen-in-Cup task with a real Franka Emika robot, FlowRetrieval achieves 3.7x the performance of the baseline imitation learning technique that learns from all prior and target data. Website: https://flow-retrieval.github.io
Authors: Yuancheng Wang, Haoyue Zhan, Liwei Liu, Ruihong Zeng, Haotian Guo, Jiachen Zheng, Qiang Zhang, Xueyao Zhang, Shunsi Zhang, Zhizheng Wu
Abstract: The recent large-scale text-to-speech (TTS) systems are usually grouped as autoregressive and non-autoregressive systems. The autoregressive systems implicitly model duration but exhibit certain deficiencies in robustness and lack of duration controllability. Non-autoregressive systems require explicit alignment information between text and speech during training and predict durations for linguistic units (e.g. phone), which may compromise their naturalness. In this paper, we introduce Masked Generative Codec Transformer (MaskGCT), a fully non-autoregressive TTS model that eliminates the need for explicit alignment information between text and speech supervision, as well as phone-level duration prediction. MaskGCT is a two-stage model: in the first stage, the model uses text to predict semantic tokens extracted from a speech self-supervised learning (SSL) model, and in the second stage, the model predicts acoustic tokens conditioned on these semantic tokens. MaskGCT follows the mask-and-predict learning paradigm. During training, MaskGCT learns to predict masked semantic or acoustic tokens based on given conditions and prompts. During inference, the model generates tokens of a specified length in a parallel manner. Experiments with 100K hours of in-the-wild speech demonstrate that MaskGCT outperforms the current state-of-the-art zero-shot TTS systems in terms of quality, similarity, and intelligibility. Audio samples are available at https://maskgct.github.io/.
Authors: Xiaoyuan Zhang, Liang Zhao, Yingying Yu, Xi Lin, Yifan Chen, Han Zhao, Qingfu Zhang
Abstract: Multiobjective optimization problems (MOPs) are prevalent in machine learning, with applications in multi-task learning, learning under fairness or robustness constraints, etc. Instead of reducing multiple objective functions into a scalar objective, MOPs aim to optimize for the so-called Pareto optimality or Pareto set learning, which involves optimizing more than one objective function simultaneously, over models with thousands / millions of parameters. Existing benchmark libraries for MOPs mainly focus on evolutionary algorithms, most of which are zeroth-order / meta-heuristic methods that do not effectively utilize higher-order information from objectives and cannot scale to large-scale models with thousands / millions of parameters. In light of the above gap, this paper introduces LibMOON, the first multiobjective optimization library that supports state-of-the-art gradient-based methods, provides a fair benchmark, and is open-sourced for the community.
Authors: Nicholas Kr\"amer
Abstract: Automatic differentiation is everywhere, but there exists only minimal documentation of how it works in complex arithmetic beyond stating "derivatives in $\mathbb{C}^d$" $\cong$ "derivatives in $\mathbb{R}^{2d}$" and, at best, shallow references to Wirtinger calculus. Unfortunately, the equivalence $\mathbb{C}^d \cong \mathbb{R}^{2d}$ becomes insufficient as soon as we need to derive custom gradient rules, e.g., to avoid differentiating "through" expensive linear algebra functions or differential equation simulators. To combat such a lack of documentation, this article surveys forward- and reverse-mode automatic differentiation with complex numbers, covering topics such as Wirtinger derivatives, a modified chain rule, and different gradient conventions while explicitly avoiding holomorphicity and the Cauchy--Riemann equations (which would be far too restrictive). To be precise, we will derive, explain, and implement a complex version of Jacobian-vector and vector-Jacobian products almost entirely with linear algebra without relying on complex analysis or differential geometry. This tutorial is a call to action, for users and developers alike, to take complex values seriously when implementing custom gradient propagation rules -- the manuscript explains how.
Authors: Haoyi Niu, Qimao Chen, Tenglong Liu, Jianxiong Li, Guyue Zhou, Yi Zhang, Jianming Hu, Xianyuan Zhan
Abstract: Reusing pre-collected data from different domains is an appealing solution for decision-making tasks that have insufficient data in the target domain but are relatively abundant in other related domains. Existing cross-domain policy transfer methods mostly aim at learning domain correspondences or corrections to facilitate policy learning, such as learning domain/task-specific discriminators, representations, or policies. This design philosophy often results in heavy model architectures or task/domain-specific modeling, lacking flexibility. This reality makes us wonder: can we directly bridge the domain gaps universally at the data level, instead of relying on complex downstream cross-domain policy transfer models? In this study, we propose the Cross-Domain Trajectory EDiting (xTED) framework that employs a specially designed diffusion model for cross-domain trajectory adaptation. Our proposed model architecture effectively captures the intricate dependencies among states, actions, and rewards, as well as the dynamics patterns within target data. By utilizing the pre-trained diffusion as a prior, source domain trajectories can be transformed to match with target domain properties while preserving original semantic information. This process implicitly corrects underlying domain gaps, enhancing state realism and dynamics reliability in the source data, and allowing flexible incorporation with various downstream policy learning methods. Despite its simplicity, xTED demonstrates superior performance in extensive simulation and real-robot experiments.
Authors: Rabia Asghar, Arslan Shaukat, Usman Akram, Rimsha Tariq
Abstract: Human immune system contains white blood cells (WBC) that are good indicator of many diseases like bacterial infections, AIDS, cancer, spleen, etc. White blood cells have been sub classified into four types: monocytes, lymphocytes, eosinophils and neutrophils on the basis of their nucleus, shape and cytoplasm. Traditionally in laboratories, pathologists and hematologists analyze these blood cells through microscope and then classify them manually. This manual process takes more time and increases the chance of human error. Hence, there is a need to automate this process. In this paper, first we have used different CNN pre-train models such as ResNet-50, InceptionV3, VGG16 and MobileNetV2 to automatically classify the white blood cells. These pre-train models are applied on Kaggle dataset of microscopic images. Although we achieved reasonable accuracy ranging between 92 to 95%, still there is need to enhance the performance. Hence, inspired by these architectures, a framework has been proposed to automatically categorize the four kinds of white blood cells with increased accuracy. The aim is to develop a convolution neural network (CNN) based classification system with decent generalization ability. The proposed CNN model has been tested on white blood cells images from Kaggle and LISC datasets. Accuracy achieved is 99.57% and 98.67% for both datasets respectively. Our proposed convolutional neural network-based model provides competitive performance as compared to previous results reported in literature.
Authors: Kevin Wang, Junbo Li, Neel P. Bhatt, Yihan Xi, Qiang Liu, Ufuk Topcu, Zhangyang Wang
Abstract: Recent advancements in Large Language Models (LLMs) have showcased their ability to perform complex reasoning tasks, but their effectiveness in planning remains underexplored. In this study, we evaluate the planning capabilities of OpenAI's o1 models across a variety of benchmark tasks, focusing on three key aspects: feasibility, optimality, and generalizability. Through empirical evaluations on constraint-heavy tasks (e.g., $\textit{Barman}$, $\textit{Tyreworld}$) and spatially complex environments (e.g., $\textit{Termes}$, $\textit{Floortile}$), we highlight o1-preview's strengths in self-evaluation and constraint-following, while also identifying bottlenecks in decision-making and memory management, particularly in tasks requiring robust spatial reasoning. Our results reveal that o1-preview outperforms GPT-4 in adhering to task constraints and managing state transitions in structured environments. However, the model often generates suboptimal solutions with redundant actions and struggles to generalize effectively in spatially complex tasks. This pilot study provides foundational insights into the planning limitations of LLMs, offering key directions for future research on improving memory management, decision-making, and generalization in LLM-based planning. Code available at: $\href{https://github.com/VITA-Group/o1-planning}{\text{https://github.com/VITA-Group/o1-planning}}$.
URLs: https://github.com/VITA-Group/o1-planning, https://github.com/VITA-Group/o1-planning
Authors: Roman Plaud, Matthieu Labeau, Antoine Saillenfest, Thomas Bonald
Abstract: Hierarchical text classification (HTC) is the task of assigning labels to a text within a structured space organized as a hierarchy. Recent works treat HTC as a conventional multilabel classification problem, therefore evaluating it as such. We instead propose to evaluate models based on specifically designed hierarchical metrics and we demonstrate the intricacy of metric choice and prediction inference method. We introduce a new challenging dataset and we evaluate fairly, recent sophisticated models, comparing them with a range of simple but strong baselines, including a new theoretically motivated loss. Finally, we show that those baselines are very often competitive with the latest models. This highlights the importance of carefully considering the evaluation methodology when proposing new methods for HTC. Code implementation and dataset are available at \url{https://github.com/RomanPlaud/revisitingHTC}.
Authors: Luca Castri, Sariah Mghames, Marc Hanheide, Nicola Bellotto
Abstract: The study of cause-and-effect is of the utmost importance in many branches of science, but also for many practical applications of intelligent systems. In particular, identifying causal relationships in situations that include hidden factors is a major challenge for methods that rely solely on observational data for building causal models. This paper proposes CAnDOIT, a causal discovery method to reconstruct causal models using both observational and interventional time-series data. The use of interventional data in the causal analysis is crucial for real-world applications, such as robotics, where the scenario is highly complex and observational data alone are often insufficient to uncover the correct causal structure. Validation of the method is performed initially on randomly generated synthetic models and subsequently on a well-known benchmark for causal structure learning in a robotic manipulation environment. The experiments demonstrate that the approach can effectively handle data from interventions and exploit them to enhance the accuracy of the causal analysis. A Python implementation of CAnDOIT has also been developed and is publicly available on GitHub: https://github.com/lcastri/causalflow.
Authors: Francisco M. Castro-Mac\'ias, Pablo Morales-\'Alvarez, Yunan Wu, Rafael Molina, Aggelos K. Katsaggelos
Abstract: Multiple Instance Learning (MIL) is widely used in medical imaging classification to reduce the labeling effort. While only bag labels are available for training, one typically seeks predictions at both bag and instance levels (classification and localization tasks, respectively). Early MIL methods treated the instances in a bag independently. Recent methods account for global and local dependencies among instances. Although they have yielded excellent results in classification, their performance in terms of localization is comparatively limited. We argue that these models have been designed to target the classification task, while implications at the instance level have not been deeply investigated. Motivated by a simple observation -- that neighboring instances are likely to have the same label -- we propose a novel, principled, and flexible mechanism to model local dependencies. It can be used alone or combined with any mechanism to model global dependencies (e.g., transformers). A thorough empirical validation shows that our module leads to state-of-the-art performance in localization while being competitive or superior in classification. Our code is at https://github.com/Franblueee/SmMIL.
Authors: Zhengting Chen, Lei Cheng, Lianghui Ding, Quanshi Zhang
Abstract: This paper presents a method to explain the internal representation structure of a neural network for image generation. Specifically, our method disentangles primitive feature components from the intermediate-layer feature of the neural network, which ensures that each feature component is exclusively used to generate a specific set of image regions. In this way, the generation of the entire image can be considered as the superposition of different pre-encoded primitive regional patterns, each being generated by a feature component. We find that the feature component can be represented as an OR relationship between the demands for generating different image regions, which is encoded by the neural network. Therefore, we extend the Harsanyi interaction to represent such an OR interaction to disentangle the feature component. Experiments show a clear correspondence between each feature component and the generation of specific image regions.
Authors: Peipei Yuan, Xinge You, Hong Chen, Xuelin Zhang, Qinmu Peng
Abstract: Generalized additive models (GAM) have been successfully applied to high dimensional data analysis. However, most existing methods cannot simultaneously estimate the link function, the component functions and the variable interaction. To alleviate this problem, we propose a new sparse additive model, named generalized sparse additive model with unknown link function (GSAMUL), in which the component functions are estimated by B-spline basis and the unknown link function is estimated by a multi-layer perceptron (MLP) network. Furthermore, $\ell_{2,1}$-norm regularizer is used for variable selection. The proposed GSAMUL can realize both variable selection and hidden interaction. We integrate this estimation into a bilevel optimization problem, where the data is split into training set and validation set. In theory, we provide the guarantees about the convergence of the approximate procedure. In applications, experimental evaluations on both synthetic and real world data sets consistently validate the effectiveness of the proposed approach.
Authors: Deok-Kyeong Jang, Dongseok Yang, Deok-Yun Jang, Byeoli Choi, Donghoon Shin, Sung-hee Lee
Abstract: This paper introduces ELMO, a real-time upsampling motion capture framework designed for a single LiDAR sensor. Modeled as a conditional autoregressive transformer-based upsampling motion generator, ELMO achieves 60 fps motion capture from a 20 fps LiDAR point cloud sequence. The key feature of ELMO is the coupling of the self-attention mechanism with thoughtfully designed embedding modules for motion and point clouds, significantly elevating the motion quality. To facilitate accurate motion capture, we develop a one-time skeleton calibration model capable of predicting user skeleton offsets from a single-frame point cloud. Additionally, we introduce a novel data augmentation technique utilizing a LiDAR simulator, which enhances global root tracking to improve environmental understanding. To demonstrate the effectiveness of our method, we compare ELMO with state-of-the-art methods in both image-based and point cloud-based motion capture. We further conduct an ablation study to validate our design principles. ELMO's fast inference time makes it well-suited for real-time applications, exemplified in our demo video featuring live streaming and interactive gaming scenarios. Furthermore, we contribute a high-quality LiDAR-mocap synchronized dataset comprising 20 different subjects performing a range of motions, which can serve as a valuable resource for future research. The dataset and evaluation code are available at {\blue \url{https://movin3d.github.io/ELMO_SIGASIA2024/}}
Authors: Haoyi Zhu, Honghui Yang, Yating Wang, Jiange Yang, Limin Wang, Tong He
Abstract: In this paper, we introduce SPA, a novel representation learning framework that emphasizes the importance of 3D spatial awareness in embodied AI. Our approach leverages differentiable neural rendering on multi-view images to endow a vanilla Vision Transformer (ViT) with intrinsic spatial understanding. We present the most comprehensive evaluation of embodied representation learning to date, covering 268 tasks across 8 simulators with diverse policies in both single-task and language-conditioned multi-task scenarios. The results are compelling: SPA consistently outperforms more than 10 state-of-the-art representation methods, including those specifically designed for embodied AI, vision-centric tasks, and multi-modal applications, while using less training data. Furthermore, we conduct a series of real-world experiments to confirm its effectiveness in practical scenarios. These results highlight the critical role of 3D spatial awareness for embodied representation learning. Our strongest model takes more than 6000 GPU hours to train and we are committed to open-sourcing all code and model weights to foster future research in embodied representation learning. Project Page: https://haoyizhu.github.io/spa/.