Authors: Reda Alami, Abdalgader Abubaker, Mastane Achab, Mohamed El Amine Seddik, Salem Lahlou
Abstract: This paper explores the effects of various forms of regularization in the context of language model alignment via self-play. While both reinforcement learning from human feedback (RLHF) and direct preference optimization (DPO) require to collect costly human-annotated pairwise preferences, the self-play fine-tuning (SPIN) approach replaces the rejected answers by data generated from the previous iterate. However, the SPIN method presents a performance instability issue in the learning phase, which can be mitigated by playing against a mixture of the two previous iterates. In the same vein, we propose in this work to address this issue from two perspectives: first, by incorporating an additional Kullback-Leibler (KL) regularization to stay at the proximity of the reference policy; second, by using the idea of fictitious play which smoothens the opponent policy across all previous iterations. In particular, we show that the KL-based regularizer boils down to replacing the previous policy by its geometric mixture with the base policy inside of the SPIN loss function. We finally discuss empirical results on MT-Bench as well as on the Hugging Face Open LLM Leaderboard.
Authors: Sarit Maitra, Sukanya Kundu, Aishwarya Shankar
Abstract: The majority of modern consumer-level energy is generated by real-time smart metering systems. These frequently contain anomalies, which prevent reliable estimates of the series' evolution. This work introduces a hybrid modeling approach combining statistics and a Convolutional Autoencoder with a dynamic threshold. The threshold is determined based on Mahalanobis distance and moving averages. It has been tested using real-life energy consumption data collected from smart metering systems. The solution includes a real-time, meter-level anomaly detection system that connects to an advanced monitoring system. This makes a substantial contribution by detecting unusual data movements and delivering an early warning. Early detection and subsequent troubleshooting can financially benefit organizations and consumers and prevent disasters from occurring.
Authors: Mahesh Lorik Yadav, Harish Guruprasad Ramaswamy, Chandrashekar Lakshminarayanan
Abstract: There currently exist two extreme viewpoints for neural network feature learning -- (i) Neural networks simply implement a kernel method (a la NTK) and hence no features are learned (ii) Neural networks can represent (and hence learn) intricate hierarchical features suitable for the data. We argue in this paper neither interpretation is likely to be correct based on a novel viewpoint. Neural networks can be viewed as a mixture of experts, where each expert corresponds to a (number of layers length) path through a sequence of hidden units. We use this alternate interpretation to motivate a model, called the Deep Linearly Gated Network (DLGN), which sits midway between deep linear networks and ReLU networks. Unlike deep linear networks, the DLGN is capable of learning non-linear features (which are then linearly combined), and unlike ReLU networks these features are ultimately simple -- each feature is effectively an indicator function for a region compactly described as an intersection of (number of layers) half-spaces in the input space. This viewpoint allows for a comprehensive global visualization of features, unlike the local visualizations for neurons based on saliency/activation/gradient maps. Feature learning in DLGNs is shown to happen and the mechanism with which this happens is through learning half-spaces in the input space that contain smooth regions of the target function. Due to the structure of DLGNs, the neurons in later layers are fundamentally the same as those in earlier layers -- they all represent a half-space -- however, the dynamics of gradient descent impart a distinct clustering to the later layer neurons. We hypothesize that ReLU networks also have similar feature learning behaviour.
Authors: Sheng Chai, Gus Chadney
Abstract: Access to smart meter data is essential to rapid and successful transitions to electrified grids, underpinned by flexibility delivered by low carbon technologies, such as electric vehicles (EV) and heat pumps, and powered by renewable energy. Yet little of this data is available for research and modelling purposes due consumer privacy protections. Whilst many are calling for raw datasets to be unlocked through regulatory changes, we believe this approach will take too long. Synthetic data addresses these challenges directly by overcoming privacy issues. In this paper, we present Faraday, a Variational Auto-encoder (VAE)-based model trained over 300 million smart meter data readings from an energy supplier in the UK, with information such as property type and low carbon technologies (LCTs) ownership. The model produces household-level synthetic load profiles conditioned on these labels, and we compare its outputs against actual substation readings to show how the model can be used for real-world applications by grid modellers interested in modelling energy grids of the future.
Authors: Xinyu Ma, Xu Chu, Zhibang Yang, Yang Lin, Xin Gao, Junfeng Zhao
Abstract: With the increasingly powerful performances and enormous scales of Pretrained Language Models (PLMs), promoting parameter efficiency in fine-tuning has become a crucial need for effective and efficient adaptation to various downstream tasks. One representative line of fine-tuning methods is Orthogonal Fine-tuning (OFT), which rigorously preserves the angular distances within the parameter space to preserve the pretrained knowledge. Despite the empirical effectiveness, OFT still suffers low parameter efficiency at $\mathcal{O}(d^2)$ and limited capability of downstream adaptation. Inspired by Givens rotation, in this paper, we proposed quasi-Givens Orthogonal Fine-Tuning (qGOFT) to address the problems. We first use $\mathcal{O}(d)$ Givens rotations to accomplish arbitrary orthogonal transformation in $SO(d)$ with provable equivalence, reducing parameter complexity from $\mathcal{O}(d^2)$ to $\mathcal{O}(d)$. Then we introduce flexible norm and relative angular adjustments under soft orthogonality regularization to enhance the adaptation capability of downstream semantic deviations. Extensive experiments on various tasks and PLMs validate the effectiveness of our methods.
Authors: Mo Kordzanganeh, Danial Keshvary, Nariman Arian
Abstract: Latent diffusion models are the state-of-the-art for synthetic image generation. To align these models with human preferences, training the models using reinforcement learning on human feedback is crucial. Black et. al 2024 introduced denoising diffusion policy optimisation (DDPO), which accounts for the iterative denoising nature of the generation by modelling it as a Markov chain with a final reward. As the reward is a single value that determines the model's performance on the entire image, the model has to navigate a very sparse reward landscape and so requires a large sample count. In this work, we extend the DDPO by presenting the Pixel-wise Policy Optimisation (PXPO) algorithm, which can take feedback for each pixel, providing a more nuanced reward to the model.
Authors: Shanshan Wu, Zheng Xu, Yanxiang Zhang, Yuanbo Zhang, Daniel Ramage
Abstract: Pre-training on public data is an effective method to improve the performance for federated learning (FL) with differential privacy (DP). This paper investigates how large language models (LLMs) trained on public data can improve the quality of pre-training data for the on-device language models trained with DP and FL. We carefully design LLM prompts to filter and transform existing public data, and generate new data to resemble the real user data distribution. The model pre-trained on our synthetic dataset achieves relative improvement of 19.0% and 22.8% in next word prediction accuracy compared to the baseline model pre-trained on a standard public dataset, when evaluated over the real user data in Gboard (Google Keyboard, a production mobile keyboard application). Furthermore, our method achieves evaluation accuracy better than or comparable to the baseline during the DP FL fine-tuning over millions of mobile devices, and our final model outperforms the baseline in production A/B testing. Our experiments demonstrate the strengths of LLMs in synthesizing data close to the private distribution even without accessing the private data, and also suggest future research directions to further reduce the distribution gap.
Authors: Yuezhu Xu, S. Sivaranjani
Abstract: The Lipschitz constant plays a crucial role in certifying the robustness of neural networks to input perturbations and adversarial attacks, as well as the stability and safety of systems with neural network controllers. Therefore, estimation of tight bounds on the Lipschitz constant of neural networks is a well-studied topic. However, typical approaches involve solving a large matrix verification problem, the computational cost of which grows significantly for deeper networks. In this letter, we provide a compositional approach to estimate Lipschitz constants for deep feedforward neural networks by obtaining an exact decomposition of the large matrix verification problem into smaller sub-problems. We further obtain a closed-form solution that applies to most common neural network activation functions, which will enable rapid robustness and stability certificates for neural networks deployed in online control settings. Finally, we demonstrate through numerical experiments that our approach provides a steep reduction in computation time while yielding Lipschitz bounds that are very close to those achieved by state-of-the-art approaches.
Authors: Ronny Hug, Stefan Becker, Wolfgang H\"ubner, Michael Arens
Abstract: An appropriate data basis grants one of the most important aspects for training and evaluating probabilistic trajectory prediction models based on neural networks. In this regard, a common shortcoming of current benchmark datasets is their limitation to sets of sample trajectories and a lack of actual ground truth distributions, which prevents the use of more expressive error metrics, such as the Wasserstein distance for model evaluation. Towards this end, this paper proposes a novel approach to synthetic dataset generation based on composite probabilistic B\'ezier curves, which is capable of generating ground truth data in terms of probability distributions over full trajectories. This allows the calculation of arbitrary posterior distributions. The paper showcases an exemplary trajectory prediction model evaluation using generated ground truth distribution data.
Authors: Haiguang Li, Usama Pervaiz, Micha{\l} Matuszak, Robert Kamara, Gilles Roux, Trausti Thormundsson, Joseph Antognini
Abstract: On-device machine learning (ODML) enables intelligent applications on resource-constrained devices. However, power consumption poses a major challenge, forcing a trade-off between model accuracy and power efficiency that often limits model complexity. The previously established Gated Compression (GC) layers offer a solution, enabling power efficiency without sacrificing model performance by selectively gating samples that lack signals of interest. However, their reliance on ground truth labels limits GC layers to supervised tasks. This work introduces the Dynamic Switch Layer (DSL), extending the benefits of GC layers to unsupervised learning scenarios, and maintaining power efficiency without the need for labeled data. The DSL builds upon the GC architecture, leveraging a dynamic pathway selection, and adapting model complexity in response to the innate structure of the data. We integrate the DSL into the SoundStream architecture and demonstrate that by routing up to 80% of samples through a lightweight pass we achieve a 12.3x reduction in the amount of computation performed and a 20.9x reduction in model size. This reduces the on-device inference latency by up to 26.5% and improves power efficiency by up to 21.4% without impacting model performance.
Authors: Shuo Xie, Zhiyuan Li
Abstract: Adam with decoupled weight decay, also known as AdamW, is widely acclaimed for its superior performance in language modeling tasks, surpassing Adam with $\ell_2$ regularization in terms of generalization and optimization. However, this advantage is not theoretically well-understood. One challenge here is that though intuitively Adam with $\ell_2$ regularization optimizes the $\ell_2$ regularized loss, it is not clear if AdamW optimizes a specific objective. In this work, we make progress toward understanding the benefit of AdamW by showing that it implicitly performs constrained optimization. More concretely, we show in the full-batch setting, if AdamW converges with any non-increasing learning rate schedule whose partial sum diverges, it must converge to a KKT point of the original loss under the constraint that the $\ell_\infty$ norm of the parameter is bounded by the inverse of the weight decay factor. This result is built on the observation that Adam can be viewed as a smoothed version of SignGD, which is the normalized steepest descent with respect to $\ell_\infty$ norm, and a surprising connection between normalized steepest descent with weight decay and Frank-Wolfe.
Authors: Yann Dubois, Bal\'azs Galambosi, Percy Liang, Tatsunori B. Hashimoto
Abstract: LLM-based auto-annotators have become a key component of the LLM development process due to their cost-effectiveness and scalability compared to human-based evaluation. However, these auto-annotators can introduce complex biases that are hard to remove. Even simple, known confounders such as preference for longer outputs remain in existing automated evaluation metrics. We propose a simple regression analysis approach for controlling biases in auto-evaluations. As a real case study, we focus on reducing the length bias of AlpacaEval, a fast and affordable benchmark for chat LLMs that uses LLMs to estimate response quality. Despite being highly correlated with human preferences, AlpacaEval is known to favor models that generate longer outputs. We introduce a length-controlled AlpacaEval that aims to answer the counterfactual question: "What would the preference be if the model's and baseline's output had the same length?". To achieve this, we first fit a generalized linear model to predict the biased output of interest (auto-annotator preferences) based on the mediators we want to control for (length difference) and other relevant features. We then obtain length-controlled preferences by predicting preferences while conditioning the GLM with a zero difference in lengths. Length-controlling not only improves the robustness of the metric to manipulations in model verbosity, we also find that it increases the Spearman correlation with LMSYS' Chatbot Arena from 0.94 to 0.98. We release the code and leaderboard at https://tatsu-lab.github.io/alpaca_eval/ .
Authors: Siddeshwar Raghavan, Jiangpeng He, Fengqing Zhu
Abstract: A significant challenge in achieving ubiquitous Artificial Intelligence is the limited ability of models to rapidly learn new information in real-world scenarios where data follows long-tailed distributions, all while avoiding forgetting previously acquired knowledge. In this work, we study the under-explored problem of Long-Tailed Online Continual Learning (LTOCL), which aims to learn new tasks from sequentially arriving class-imbalanced data streams. Each data is observed only once for training without knowing the task data distribution. We present DELTA, a decoupled learning approach designed to enhance learning representations and address the substantial imbalance in LTOCL. We enhance the learning process by adapting supervised contrastive learning to attract similar samples and repel dissimilar (out-of-class) samples. Further, by balancing gradients during training using an equalization loss, DELTA significantly enhances learning outcomes and successfully mitigates catastrophic forgetting. Through extensive evaluation, we demonstrate that DELTA improves the capacity for incremental learning, surpassing existing OCL methods. Our results suggest considerable promise for applying OCL in real-world applications.
Authors: Yan Kang, Ziyao Ren, Lixin Fan, Linghua Yang, Yongxin Tong, Qiang Yang
Abstract: SecureBoost is a tree-boosting algorithm that leverages homomorphic encryption (HE) to protect data privacy in vertical federated learning. SecureBoost and its variants have been widely adopted in fields such as finance and healthcare. However, the hyperparameters of SecureBoost are typically configured heuristically for optimizing model performance (i.e., utility) solely, assuming that privacy is secured. Our study found that SecureBoost and some of its variants are still vulnerable to label leakage. This vulnerability may lead the current heuristic hyperparameter configuration of SecureBoost to a suboptimal trade-off between utility, privacy, and efficiency, which are pivotal elements toward a trustworthy federated learning system. To address this issue, we propose the Constrained Multi-Objective SecureBoost (CMOSB) algorithm, which aims to approximate Pareto optimal solutions that each solution is a set of hyperparameters achieving an optimal trade-off between utility loss, training cost, and privacy leakage. We design measurements of the three objectives, including a novel label inference attack named instance clustering attack (ICA) to measure the privacy leakage of SecureBoost. Additionally, we provide two countermeasures against ICA. The experimental results demonstrate that the CMOSB yields superior hyperparameters over those optimized by grid search and Bayesian optimization regarding the trade-off between utility loss, training cost, and privacy leakage.
Authors: I-Hong Hou
Abstract: This paper studies multi-stage systems with end-to-end bandit feedback. In such systems, each job needs to go through multiple stages, each managed by a different agent, before generating an outcome. Each agent can only control its own action and learn the final outcome of the job. It has neither knowledge nor control on actions taken by agents in the next stage. The goal of this paper is to develop distributed online learning algorithms that achieve sublinear regret in adversarial environments. The setting of this paper significantly expands the traditional multi-armed bandit problem, which considers only one agent and one stage. In addition to the exploration-exploitation dilemma in the traditional multi-armed bandit problem, we show that the consideration of multiple stages introduces a third component, education, where an agent needs to choose its actions to facilitate the learning of agents in the next stage. To solve this newly introduced exploration-exploitation-education trilemma, we propose a simple distributed online learning algorithm, $\epsilon-$EXP3. We theoretically prove that the $\epsilon-$EXP3 algorithm is a no-regret policy that achieves sublinear regret. Simulation results show that the $\epsilon-$EXP3 algorithm significantly outperforms existing no-regret online learning algorithms for the traditional multi-armed bandit problem.
Authors: Hamidreza Montaseri, Amin Gohari
Abstract: The proliferation of algorithmic systems has fueled discussions surrounding the regulation and control of their social impact. Herein, we consider a system whose primary objective is to maximize utility by selecting the most qualified individuals. To promote demographic parity in the selection algorithm, we consider penalizing discrimination across social groups. We examine conditions under which a discrimination penalty can effectively reduce disparity in the selection. Additionally, we explore the implications of such a penalty when individual qualifications may evolve over time in response to the imposed penalizing policy. We identify scenarios where the penalty could hinder the natural attainment of equity within the population. Moreover, we propose certain conditions that can counteract this undesirable outcome, thus ensuring fairness.
Authors: Guoming Li, Jian Yang, Shangsong Liang, Dongsheng Luo
Abstract: Spectral Graph Neural Networks (GNNs) have achieved tremendous success in graph learning. As an essential part of spectral GNNs, spectral graph convolution extracts crucial frequency information in graph data, leading to superior performance of spectral GNNs in downstream tasks. However, in this paper, we show that existing spectral GNNs remain critical drawbacks in performing the spectral graph convolution. Specifically, considering the spectral graph convolution as a construction operation towards target output, we prove that existing popular convolution paradigms cannot construct the target output with mild conditions on input graph signals, causing spectral GNNs to fall into suboptimal solutions. To address the issues, we rethink the spectral graph convolution from a more general two-dimensional (2-D) signal convolution perspective and propose a new convolution paradigm, named 2-D graph convolution. We prove that 2-D graph convolution unifies existing graph convolution paradigms, and is capable to construct arbitrary target output. Based on the proposed 2-D graph convolution, we further propose ChebNet2D, an efficient and effective GNN implementation of 2-D graph convolution through applying Chebyshev interpolation. Extensive experiments on benchmark datasets demonstrate both effectiveness and efficiency of the ChebNet2D.
Authors: Zi-Hao Qiu, Siqi Guo, Mao Xu, Tuo Zhao, Lijun Zhang, Tianbao Yang
Abstract: The temperature parameter plays a profound role during training and/or inference with large foundation models (LFMs) such as large language models (LLMs) and CLIP models. Particularly, it adjusts the logits in the softmax function in LLMs, which is crucial for next token generation, and it scales the similarities in the contrastive loss for training CLIP models. A significant question remains: Is it viable to learn a neural network to predict a personalized temperature of any input data for enhancing LFMs"? In this paper, we present a principled framework for learning a small yet generalizable temperature prediction network (TempNet) to improve LFMs. Our solution is composed of a novel learning framework with a robust loss underpinned by constrained distributionally robust optimization (DRO), and a properly designed TempNet with theoretical inspiration. TempNet can be trained together with a large foundation model from scratch or learned separately given a pretrained foundation model. It is not only useful for predicting personalized temperature to promote the training of LFMs but also generalizable and transferable to new tasks. Our experiments on LLMs and CLIP models demonstrate that TempNet greatly improves the performance of existing solutions or models, e.g. Table 1. The code to reproduce the experimental results in this paper can be found at https://github.com/zhqiu/TempNet.
Authors: Adarsh Jamadandi, Celia Rubio-Madrigal, Rebekka Burkholz
Abstract: Message Passing Graph Neural Networks are known to suffer from two problems that are sometimes believed to be diametrically opposed: over-squashing and over-smoothing. The former results from topological bottlenecks that hamper the information flow from distant nodes and are mitigated by spectral gap maximization, primarily, by means of edge additions. However, such additions often promote over-smoothing that renders nodes of different classes less distinguishable. Inspired by the Braess phenomenon, we argue that deleting edges can address over-squashing and over-smoothing simultaneously. This insight explains how edge deletions can improve generalization, thus connecting spectral gap optimization to a seemingly disconnected objective of reducing computational resources by pruning graphs for lottery tickets. To this end, we propose a more effective spectral gap optimization framework to add or delete edges and demonstrate its effectiveness on large heterophilic datasets.
Authors: Yongding Tian, Zaid Al-Ars, Maksim Kitsak, Peter Hofstee
Abstract: Federated learning and gossip learning are emerging methodologies designed to mitigate data privacy concerns by retaining training data on client devices and exclusively sharing locally-trained machine learning (ML) models with others. The primary distinction between the two lies in their approach to model aggregation: federated learning employs a centralized parameter server, whereas gossip learning adopts a fully decentralized mechanism, enabling direct model exchanges among nodes. This decentralized nature often positions gossip learning as less efficient compared to federated learning. Both methodologies involve a critical step: computing a representation of received ML models and integrating this representation into the existing model. Conventionally, this representation is derived by averaging the received models, exemplified by the FedAVG algorithm. Our findings suggest that this averaging approach inherently introduces a potential delay in model convergence. We identify the underlying cause and refer to it as the "vanishing variance" problem, where averaging across uncorrelated ML models undermines the optimal variance established by the Xavier weight initialization. Unlike federated learning where the central server ensures model correlation, and unlike traditional gossip learning which circumvents this problem through model partitioning and sampling, our research introduces a variance-corrected model averaging algorithm. This novel algorithm preserves the optimal variance needed during model averaging, irrespective of network topology or non-IID data distributions. Our extensive simulation results demonstrate that our approach enables gossip learning to achieve convergence efficiency comparable to that of federated learning.
Authors: Abhishek Sahu, Peter H. Aaen, Praveen Damacharla
Abstract: In this paper, we present a machine learning based architecture for microwave characterization of inkjet printed components on flexible substrates. Our proposed architecture uses several machine learning algorithms and automatically selects the best algorithm to extract the material parameters (ink conductivity and dielectric properties) from on-wafer measurements. Initially, the mutual dependence between material parameters of the inkjet printed coplanar waveguides (CPWs) and EM-simulated propagation constants is utilized to train the machine learning models. Next, these machine learning models along with measured propagation constants are used to extract the ink conductivity and dielectric properties of the test prototypes. To demonstrate the applicability of our proposed approach, we compare and contrast four heuristic based machine learning models. It is shown that eXtreme Gradient Boosted Trees Regressor (XGB) and Light Gradient Boosting (LGB) algorithms perform best for the characterization problem under study.
Authors: Seungjae Jung, Gunsoo Han, Daniel Wontae Nam, Kyoung-Woon On
Abstract: Aligning Large Language Models (LLMs) to human preferences through preference optimization has been crucial but labor-intensive, necessitating for each prompt a comparison of both a chosen and a rejected text completion by evaluators. Recently, Kahneman-Tversky Optimization (KTO) has demonstrated that LLMs can be aligned using merely binary "thumbs-up" or "thumbs-down" signals on each prompt-completion pair. In this paper, we present theoretical foundations to explain the successful alignment achieved through these binary signals. Our analysis uncovers a new perspective: optimizing a binary classifier, whose logit is a reward, implicitly induces minimizing the Direct Preference Optimization (DPO) loss. In the process of this discovery, we identified two techniques for effective alignment: reward shift and underlying distribution matching. Consequently, we propose a new algorithm, \textit{Binary Classifier Optimization}, that integrates the techniques. We validate our methodology in two settings: first, on a paired preference dataset, where our method performs on par with DPO and KTO; and second, on binary signal datasets simulating real-world conditions with divergent underlying distributions between thumbs-up and thumbs-down data. Our model consistently demonstrates effective and robust alignment across two base LLMs and three different binary signal datasets, showcasing the strength of our approach to learning from binary feedback.
Authors: Tianle Pu, Changjun Fan, Mutian Shen, Yizhou Lu, Li Zeng, Zohar Nussinov, Chao Chen, Zhong Liu
Abstract: Many complex problems encountered in both production and daily life can be conceptualized as combinatorial optimization problems (COPs) over graphs. Recent years, reinforcement learning (RL) based models have emerged as a promising direction, which treat the COPs solving as a heuristic learning problem. However, current finite-horizon-MDP based RL models have inherent limitations. They are not allowed to explore adquately for improving solutions at test time, which may be necessary given the complexity of NP-hard optimization tasks. Some recent attempts solve this issue by focusing on reward design and state feature engineering, which are tedious and ad-hoc. In this work, we instead propose a much simpler but more effective technique, named gauge transformation (GT). The technique is originated from physics, but is very effective in enabling RL agents to explore to continuously improve the solutions during test. Morever, GT is very simple, which can be implemented with less than 10 lines of Python codes, and can be applied to a vast majority of RL models. Experimentally, we show that traditional RL models with GT technique produce the state-of-the-art performances on the MaxCut problem. Furthermore, since GT is independent of any RL models, it can be seamlessly integrated into various RL frameworks, paving the way of these models for more effective explorations in the solving of general COPs.
Authors: Chuqin Geng, Zhaoyue Wang, Haolin Ye, Saifei Liao, Xujie Si
Abstract: Specifications play a crucial role in neural network verification. They define the precise input regions we aim to verify, typically represented as L-infinity norm balls. While recent research suggests using neural activation patterns (NAPs) as specifications for verifying unseen test set data, it focuses on computing the most refined NAPs, often limited to very small regions in the input space. In this paper, we study the following problem: Given a neural network, find a minimal (coarsest) NAP that is sufficient for formal verification of the network's robustness. Finding the minimal NAP specification not only expands verifiable bounds but also provides insights into which neurons contribute to the model's robustness. To address this problem, we propose several exact and approximate approaches. Our exact approaches leverage the verification tool to find minimal NAP specifications in either a deterministic or statistical manner. Whereas the approximate methods efficiently estimate minimal NAPs using adversarial examples and local gradients, without making calls to the verification tool. This allows us to inspect potential causal links between neurons and the robustness of state-of-the-art neural networks, a task for which existing verification frameworks fail to scale. Our experimental results suggest that minimal NAP specifications require much smaller fractions of neurons compared to the most refined NAP specifications, yet they can significantly expand the verifiable boundaries to several orders of magnitude larger.
Authors: Anurag Singh, Siu Lun Chau, Shahine Bouabid, Krikamol Muandet
Abstract: Out-of-distribution (OOD) generalisation is challenging because it involves not only learning from empirical data, but also deciding among various notions of generalisation, e.g., optimising the average-case risk, worst-case risk, or interpolations thereof. While this choice should in principle be made by the model operator like medical doctors, this information might not always be available at training time. The institutional separation between machine learners and model operators leads to arbitrary commitments to specific generalisation strategies by machine learners due to these deployment uncertainties. We introduce the Imprecise Domain Generalisation framework to mitigate this, featuring an imprecise risk optimisation that allows learners to stay imprecise by optimising against a continuous spectrum of generalisation strategies during training, and a model framework that allows operators to specify their generalisation preference at deployment. Supported by both theoretical and empirical evidence, our work showcases the benefits of integrating imprecision into domain generalisation.
Authors: Philipp Andelfinger, Justin N. Kreikemeyer
Abstract: Recently proposed gradient estimators enable gradient descent over stochastic programs with discrete jumps in the response surface, which are not covered by automatic differentiation (AD) alone. Although these estimators' capability to guide a swift local search has been shown for certain problems, their applicability to models relevant to real-world applications remains largely unexplored. As the gradients governing the choice in candidate solutions are calculated from sampled simulation trajectories, the optimization procedure bears similarities to metaheuristics such as particle swarm optimization, which puts the focus on the different methods' calibration progress per function evaluation. Here, we consider the calibration of force-based crowd evacuation models based on the popular Social Force model augmented by discrete decision making. After studying the ability of an AD-based estimator for branching programs to capture the simulation's rugged response surface, calibration problems are tackled using gradient descent and two metaheuristics. As our main insights, we find 1) that the estimation's fidelity benefits from disregarding jumps of large magnitude inherent to the Social Force model, and 2) that the common problem of calibration by adjusting a simulation input distribution obviates the need for AD across the Social Force calculations, allowing gradient descent to excel.
Authors: Yeda Song, Dongwook Lee, Gunhee Kim
Abstract: Offline reinforcement learning (RL) is a compelling framework for learning optimal policies from past experiences without additional interaction with the environment. Nevertheless, offline RL inevitably faces the problem of distributional shifts, where the states and actions encountered during policy execution may not be in the training dataset distribution. A common solution involves incorporating conservatism into the policy or the value function to safeguard against uncertainties and unknowns. In this work, we focus on achieving the same objectives of conservatism but from a different perspective. We propose COmpositional COnservatism with Anchor-seeking (COCOA) for offline RL, an approach that pursues conservatism in a compositional manner on top of the transductive reparameterization (Netanyahu et al., 2023), which decomposes the input variable (the state in our case) into an anchor and its difference from the original input. Our COCOA seeks both in-distribution anchors and differences by utilizing the learned reverse dynamics model, encouraging conservatism in the compositional input space for the policy or value function. Such compositional conservatism is independent of and agnostic to the prevalent behavioral conservatism in offline RL. We apply COCOA to four state-of-the-art offline RL algorithms and evaluate them on the D4RL benchmark, where COCOA generally improves the performance of each algorithm. The code is available at https://github.com/runamu/compositional-conservatism.
Authors: Taminul Islam, Md. Alif Sheakh, Mst. Sazia Tahosin, Most. Hasna Hena, Shopnil Akash, Yousef A. Bin Jardan, Gezahign Fentahun Wondmie, Hiba-Allah Nafidi, Mohammed Bourhia
Abstract: Breast cancer has rapidly increased in prevalence in recent years, making it one of the leading causes of mortality worldwide. Among all cancers, it is by far the most common. Diagnosing this illness manually requires significant time and expertise. Since detecting breast cancer is a time-consuming process, preventing its further spread can be aided by creating machine-based forecasts. Machine learning and Explainable AI are crucial in classification as they not only provide accurate predictions but also offer insights into how the model arrives at its decisions, aiding in the understanding and trustworthiness of the classification results. In this study, we evaluate and compare the classification accuracy, precision, recall, and F-1 scores of five different machine learning methods using a primary dataset (500 patients from Dhaka Medical College Hospital). Five different supervised machine learning techniques, including decision tree, random forest, logistic regression, naive bayes, and XGBoost, have been used to achieve optimal results on our dataset. Additionally, this study applied SHAP analysis to the XGBoost model to interpret the model's predictions and understand the impact of each feature on the model's output. We compared the accuracy with which several algorithms classified the data, as well as contrasted with other literature in this field. After final evaluation, this study found that XGBoost achieved the best model accuracy, which is 97%.
Authors: Mohammed A. A. Elmaleeh
Abstract: This paper presents different neural network-based classifier algorithms for diagnosing and classifying Anemia. The study compares these classifiers with established models such as Feed Forward Neural Network (FFNN), Elman network, and Non-linear Auto-Regressive Exogenous model (NARX). Experimental evaluations were conducted using data from clinical laboratory test results for 230 patients. The proposed neural network features nine inputs (age, gender, RBC, HGB, HCT, MCV, MCH, MCHC, WBCs) and one output. The simulation outcomes for diverse patients demonstrate that the suggested artificial neural network rapidly and accurately detects the presence of the disease. Consequently, the network could be seamlessly integrated into clinical laboratories for automatic generation of Anemia patients' reports Additionally, the suggested method is affordable and can be deployed on hardware at low costs.
Authors: Elita Lobo, Harvineet Singh, Marek Petrik, Cynthia Rudin, Himabindu Lakkaraju
Abstract: Off-policy Evaluation (OPE) methods are a crucial tool for evaluating policies in high-stakes domains such as healthcare, where exploration is often infeasible, unethical, or expensive. However, the extent to which such methods can be trusted under adversarial threats to data quality is largely unexplored. In this work, we make the first attempt at investigating the sensitivity of OPE methods to marginal adversarial perturbations to the data. We design a generic data poisoning attack framework leveraging influence functions from robust statistics to carefully construct perturbations that maximize error in the policy value estimates. We carry out extensive experimentation with multiple healthcare and control datasets. Our results demonstrate that many existing OPE methods are highly prone to generating value estimates with large errors when subject to data poisoning attacks, even for small adversarial perturbations. These findings question the reliability of policy values derived using OPE methods and motivate the need for developing OPE methods that are statistically robust to train-time data poisoning attacks.
Authors: Zihao Wang, Shaoduo Gan
Abstract: Optimizing the Key-Value (KV) cache of the Large Language Model (LLM) has been considered critical to saving the cost of inference. Most of the existing KV-cache compression algorithms attempted to sparsify the sequence of tokens by taking advantage of the different importance of tokens. In this work, we found that by identifying the importance of attention layers, we could optimize the KV-cache jointly from two dimensions. Based on our observations regarding layer-wise importance in inference, we propose SqueezeAttention to precisely optimize the allocation of KV-cache budget among layers on-the-fly and then incorporate three representative token sparsification algorithms to compress the KV-cache for each layer with its very own budget. By optimizing the KV-cache from both sequence's and layer's dimensions, SqueezeAttention achieves around 30% to 70% of the memory reductions and up to 2.2 times of throughput improvements in a wide range of LLMs and benchmarks. The code is available at https://github.com/hetailang/SqueezeAttention.
Authors: Yukun Yang, Naihao Wang, Haixin Yang, Ruirui Li
Abstract: Label noise is a common issue in real-world datasets that inevitably impacts the generalization of models. This study focuses on robust classification tasks where the label noise is instance-dependent. Estimating the transition matrix accurately in this task is challenging, and methods based on sample selection often exhibit confirmation bias to varying degrees. Sparse over-parameterized training (SOP) has been theoretically effective in estimating and recovering label noise, offering a novel solution for noise-label learning. However, this study empirically observes and verifies a technical flaw of SOP: the lack of coordination between model predictions and noise recovery leads to increased generalization error. To address this, we propose a method called Coordinated Sparse Recovery (CSR). CSR introduces a collaboration matrix and confidence weights to coordinate model predictions and noise recovery, reducing error leakage. Based on CSR, this study designs a joint sample selection strategy and constructs a comprehensive and powerful learning framework called CSR+. CSR+ significantly reduces confirmation bias, especially for datasets with more classes and a high proportion of instance-specific noise. Experimental results on simulated and real-world noisy datasets demonstrate that both CSR and CSR+ achieve outstanding performance compared to methods at the same level.
Authors: Yi Zhang, Jitao Sang
Abstract: Fairness is critical for artificial intelligence systems, especially for those deployed in high-stakes applications such as hiring and justice. Existing efforts toward fairness in machine learning fairness require retraining or fine-tuning the neural network weights to meet the fairness criteria. However, this is often not feasible in practice for regular model users due to the inability to access and modify model weights. In this paper, we propose a more flexible fairness paradigm, Inference-Time Rule Eraser, or simply Eraser, which considers the case where model weights can not be accessed and tackles fairness issues from the perspective of biased rules removal at inference-time. We first verified the feasibility of modifying the model output to wipe the biased rule through Bayesian analysis, and deduced Inference-Time Rule Eraser via subtracting the logarithmic value associated with unfair rules (i.e., the model's response to biased features) from the model's logits output as a means of removing biased rules. Moreover, we present a specific implementation of Rule Eraser that involves two stages: (1) limited queries are performed on the model with inaccessible weights to distill its biased rules into an additional patched model, and (2) during inference time, the biased rules already distilled into the patched model are excluded from the output of the original model, guided by the removal strategy outlined in Rule Eraser. Exhaustive experimental evaluation demonstrates the effectiveness and superior performance of the proposed Rule Eraser in addressing fairness concerns.
Authors: Muhammad Tanzil Furqon, Mahardhika Pratama, Lin Liu, Habibullah, Kutluyil Dogancay
Abstract: Remaining Useful Life (RUL) predictions play vital role for asset planning and maintenance leading to many benefits to industries such as reduced downtime, low maintenance costs, etc. Although various efforts have been devoted to study this topic, most existing works are restricted for i.i.d conditions assuming the same condition of the training phase and the deployment phase. This paper proposes a solution to this problem where a mix-up domain adaptation (MDAN) is put forward. MDAN encompasses a three-staged mechanism where the mix-up strategy is not only performed to regularize the source and target domains but also applied to establish an intermediate mix-up domain where the source and target domains are aligned. The self-supervised learning strategy is implemented to prevent the supervision collapse problem. Rigorous evaluations have been performed where MDAN is compared to recently published works for dynamic RUL predictions. MDAN outperforms its counterparts with substantial margins in 12 out of 12 cases. In addition, MDAN is evaluated with the bearing machine dataset where it beats prior art with significant gaps in 8 of 12 cases. Source codes of MDAN are made publicly available in \url{https://github.com/furqon3009/MDAN}.
Authors: Atoosa Parsa, Corey S. O'Hern, Rebecca Kramer-Bottiglio, Josh Bongard
Abstract: There is growing interest in engineering unconventional computing devices that leverage the intrinsic dynamics of physical substrates to perform fast and energy-efficient computations. Granular metamaterials are one such substrate that has emerged as a promising platform for building wave-based information processing devices with the potential to integrate sensing, actuation, and computation. Their high-dimensional and nonlinear dynamics result in nontrivial and sometimes counter-intuitive wave responses that can be shaped by the material properties, geometry, and configuration of individual grains. Such highly tunable rich dynamics can be utilized for mechanical computing in special-purpose applications. However, there are currently no general frameworks for the inverse design of large-scale granular materials. Here, we build upon the similarity between the spatiotemporal dynamics of wave propagation in material and the computational dynamics of Recurrent Neural Networks to develop a gradient-based optimization framework for harmonically driven granular crystals. We showcase how our framework can be utilized to design basic logic gates where mechanical vibrations carry the information at predetermined frequencies. We compare our design methodology with classic gradient-free methods and find that our approach discovers higher-performing configurations with less computational effort. Our findings show that a gradient-based optimization method can greatly expand the design space of metamaterials and provide the opportunity to systematically traverse the parameter space to find materials with the desired functionalities.
Authors: David D. Nguyen, David Liebowitz, Surya Nepal, Salil S. Kanhere, Sharif Abuadbba
Abstract: Honeyfiles are security assets designed to attract and detect intruders on compromised systems. Honeyfiles are a type of honeypot that mimic real, sensitive documents, creating the illusion of the presence of valuable data. Interaction with a honeyfile reveals the presence of an intruder, and can provide insights into their goals and intentions. Their practical use, however, is limited by the time, cost and effort associated with manually creating realistic content. The introduction of large language models has made high-quality text generation accessible, but honeyfiles contain a variety of content including charts, tables and images. This content needs to be plausible and realistic, as well as semantically consistent both within honeyfiles and with the real documents they mimic, to successfully deceive an intruder. In this paper, we focus on an important component of the honeyfile content generation problem: document charts. Charts are ubiquitous in corporate documents and are commonly used to communicate quantitative and scientific data. Existing image generation models, such as DALL-E, are rather prone to generating charts with incomprehensible text and unconvincing data. We take a multi-modal approach to this problem by combining two purpose-built generative models: a multitask Transformer and a specialized multi-head autoencoder. The Transformer generates realistic captions and plot text, while the autoencoder generates the underlying tabular data for the plot. To advance the field of automated honeyplot generation, we also release a new document-chart dataset and propose a novel metric Keyword Semantic Matching (KSM). This metric measures the semantic consistency between keywords of a corpus and a smaller bag of words. Extensive experiments demonstrate excellent performance against multiple large language models, including ChatGPT and GPT4.
Authors: Yuqing Li, Tao Luo, Qixuan Zhou
Abstract: In this paper, we advance the understanding of neural network training dynamics by examining the intricate interplay of various factors introduced by weight parameters in the initialization process. Motivated by the foundational work of Luo et al. (J. Mach. Learn. Res., Vol. 22, Iss. 1, No. 71, pp 3327-3373), we explore the gradient descent dynamics of neural networks through the lens of macroscopic limits, where we analyze its behavior as width $m$ tends to infinity. Our study presents a unified approach with refined techniques designed for multi-layer fully connected neural networks, which can be readily extended to other neural network architectures. Our investigation reveals that gradient descent can rapidly drive deep neural networks to zero training loss, irrespective of the specific initialization schemes employed by weight parameters, provided that the initial scale of the output function $\kappa$ surpasses a certain threshold. This regime, characterized as the theta-lazy area, accentuates the predominant influence of the initial scale $\kappa$ over other factors on the training behavior of neural networks. Furthermore, our approach draws inspiration from the Neural Tangent Kernel (NTK) paradigm, and we expand its applicability. While NTK typically assumes that $\lim_{m\to\infty}\frac{\log \kappa}{\log m}=\frac{1}{2}$, and imposes each weight parameters to scale by the factor $\frac{1}{\sqrt{m}}$, in our theta-lazy regime, we discard the factor and relax the conditions to $\lim_{m\to\infty}\frac{\log \kappa}{\log m}>0$. Similar to NTK, the behavior of overparameterized neural networks within the theta-lazy regime trained by gradient descent can be effectively described by a specific kernel. Through rigorous analysis, our investigation illuminates the pivotal role of $\kappa$ in governing the training dynamics of neural networks.
Authors: Zhen Fang, Yixuan Li, Feng Liu, Bo Han, Jie Lu
Abstract: Supervised learning aims to train a classifier under the assumption that training and test data are from the same distribution. To ease the above assumption, researchers have studied a more realistic setting: out-of-distribution (OOD) detection, where test data may come from classes that are unknown during training (i.e., OOD data). Due to the unavailability and diversity of OOD data, good generalization ability is crucial for effective OOD detection algorithms, and corresponding learning theory is still an open problem. To study the generalization of OOD detection, this paper investigates the probably approximately correct (PAC) learning theory of OOD detection that fits the commonly used evaluation metrics in the literature. First, we find a necessary condition for the learnability of OOD detection. Then, using this condition, we prove several impossibility theorems for the learnability of OOD detection under some scenarios. Although the impossibility theorems are frustrating, we find that some conditions of these impossibility theorems may not hold in some practical scenarios. Based on this observation, we next give several necessary and sufficient conditions to characterize the learnability of OOD detection in some practical scenarios. Lastly, we offer theoretical support for representative OOD detection works based on our OOD theory.
Authors: Jaesung Choi, Pilwon Kim
Abstract: Removing noise from a signal without knowing the characteristics of the noise is a challenging task. This paper introduces a signal-noise separation method based on time series prediction. We use Reservoir Computing (RC) to extract the maximum portion of "predictable information" from a given signal. Reproducing the deterministic component of the signal using RC, we estimate the noise distribution from the difference between the original signal and reconstructed one. The method is based on a machine learning approach and requires no prior knowledge of either the deterministic signal or the noise distribution. It provides a way to identify additivity/multiplicativity of noise and to estimate the signal-to-noise ratio (SNR) indirectly. The method works successfully for combinations of various signal and noise, including chaotic signal and highly oscillating sinusoidal signal which are corrupted by non-Gaussian additive/ multiplicative noise. The separation performances are robust and notably outstanding for signals with strong noise, even for those with negative SNR.
Authors: Yu-Hsi Chen
Abstract: In the realm of continual learning, the presence of noisy labels within data streams represents a notable obstacle to model reliability and fairness. We focus on the data stream scenario outlined in pertinent literature, characterized by fuzzy task boundaries and noisy labels. To address this challenge, we introduce a novel and intuitive sampling method called Noisy Test Debiasing (NTD) to mitigate noisy labels in evolving data streams and establish a fair and robust continual learning algorithm. NTD is straightforward to implement, making it feasible across various scenarios. Our experiments benchmark four datasets, including two synthetic noise datasets (CIFAR10 and CIFAR100) and real-world noise datasets (mini-WebVision and Food-101N). The results validate the efficacy of NTD for online continual learning in scenarios with noisy labels in data streams. Compared to the previous leading approach, NTD achieves a training speedup enhancement over two times while maintaining or surpassing accuracy levels. Moreover, NTD utilizes less than one-fifth of the GPU memory resources compared to previous leading methods.
Authors: Moshe Eliasof, Eldad Haber
Abstract: This paper investigates a link between Graph Neural Networks (GNNs) and Binary Programming (BP) problems, laying the groundwork for GNNs to approximate solutions for these computationally challenging problems. By analyzing the sensitivity of BP problems, we are able to frame the solution of BP problems as a heterophilic node classification task. We then propose Binary-Programming GNN (BPGNN), an architecture that integrates graph representation learning techniques with BP-aware features to approximate BP solutions efficiently. Additionally, we introduce a self-supervised data generation mechanism, to enable efficient and tractable training data acquisition even for large-scale BP problems. Experimental evaluations of BPGNN across diverse BP problem sizes showcase its superior performance compared to exhaustive search and heuristic approaches. Finally, we discuss open challenges in the under-explored field of BP problems with GNNs.
Authors: Wenlong Liao, Fernando Porte-Agel, Jiannong Fang, Christian Rehtanz, Shouxiang Wang, Dechang Yang, Zhe Yang
Abstract: Machine learning models have made significant progress in load forecasting, but their forecast accuracy is limited in cases where historical load data is scarce. Inspired by the outstanding performance of large language models (LLMs) in computer vision and natural language processing, this paper aims to discuss the potential of large time series models in load forecasting with scarce historical data. Specifically, the large time series model is constructed as a time series generative pre-trained transformer (TimeGPT), which is trained on massive and diverse time series datasets consisting of 100 billion data points (e.g., finance, transportation, banking, web traffic, weather, energy, healthcare, etc.). Then, the scarce historical load data is used to fine-tune the TimeGPT, which helps it to adapt to the data distribution and characteristics associated with load forecasting. Simulation results show that TimeGPT outperforms the benchmarks (e.g., popular machine learning models and statistical models) for load forecasting on several real datasets with scarce training samples, particularly for short look-ahead times. However, it cannot be guaranteed that TimeGPT is always superior to benchmarks for load forecasting with scarce data, since the performance of TimeGPT may be affected by the distribution differences between the load data and the training data. In practical applications, we can divide the historical data into a training set and a validation set, and then use the validation set loss to decide whether TimeGPT is the best choice for a specific dataset.
Authors: Rohit Agarwal, Arijit Das, Alexander Horsch, Krishna Agarwal, Dilip K. Prasad
Abstract: The domain of online learning has experienced multifaceted expansion owing to its prevalence in real-life applications. Nonetheless, this progression operates under the assumption that the input feature space of the streaming data remains constant. In this survey paper, we address the topic of online learning in the context of haphazard inputs, explicitly foregoing such an assumption. We discuss, classify, evaluate, and compare the methodologies that are adept at modeling haphazard inputs, additionally providing the corresponding code implementations and their carbon footprint. Moreover, we classify the datasets related to the field of haphazard inputs and introduce evaluation metrics specifically designed for datasets exhibiting imbalance. The code of each methodology can be found at https://github.com/Rohit102497/HaphazardInputsReview
Authors: Xudong Yu, Chenjia Bai, Haoran He, Changhong Wang, Xuelong Li
Abstract: Sequential decision-making is desired to align with human intents and exhibit versatility across various tasks. Previous methods formulate it as a conditional generation process, utilizing return-conditioned diffusion models to directly model trajectory distributions. Nevertheless, the return-conditioned paradigm relies on pre-defined reward functions, facing challenges when applied in multi-task settings characterized by varying reward functions (versatility) and showing limited controllability concerning human preferences (alignment). In this work, we adopt multi-task preferences as a unified condition for both single- and multi-task decision-making, and propose preference representations aligned with preference labels. The learned representations are used to guide the conditional generation process of diffusion models, and we introduce an auxiliary objective to maximize the mutual information between representations and corresponding generated trajectories, improving alignment between trajectories and preferences. Extensive experiments in D4RL and Meta-World demonstrate that our method presents favorable performance in single- and multi-task scenarios, and exhibits superior alignment with preferences.
Authors: Roi Livni
Abstract: We analyze the sample complexity of full-batch Gradient Descent (GD) in the setup of non-smooth Stochastic Convex Optimization. We show that the generalization error of GD, with (minmax) optimal choice of hyper-parameters, can be $\tilde \Theta(d/m + 1/\sqrt{m})$, where $d$ is the dimension and $m$ is the sample size. This matches the sample complexity of \emph{worst-case} empirical risk minimizers. That means that, in contrast with other algorithms, GD has no advantage over naive ERMs. Our bound follows from a new generalization bound that depends on both the dimension as well as the learning rate and number of iterations. Our bound also shows that, for general hyper-parameters, when the dimension is strictly larger than number of samples, $T=\Omega(1/\epsilon^4)$ iterations are necessary to avoid overfitting. This resolves an open problem by \citet*{schliserman2024dimension, amir2021sgd}, and improves over previous lower bounds that demonstrated that the sample size must be at least square root of the dimension.
Authors: Han Lu, Fangfang Li, Quanxue Gao, Cheng Deng, Chris Ding, Qianqian Wang
Abstract: Fuzzy K-Means clustering is a critical technique in unsupervised data analysis. However, the performance of popular Fuzzy K-Means algorithms is sensitive to the selection of initial cluster centroids and is also affected by noise when updating mean cluster centroids. To address these challenges, this paper proposes a novel Fuzzy K-Means clustering algorithm that entirely eliminates the reliance on cluster centroids, obtaining membership matrices solely through distance matrix computation. This innovation enhances flexibility in distance measurement between sample points, thus improving the algorithm's performance and robustness. The paper also establishes theoretical connections between the proposed model and popular Fuzzy K-Means clustering techniques. Experimental results on several real datasets demonstrate the effectiveness of the algorithm.
Authors: Zhihui Deng, Yuanyuan Duan, Leilai Shao, Xiaolei Zhu
Abstract: Chiplet-based systems, integrating various silicon dies manufactured at different integrated circuit technology nodes on a carrier interposer, have garnered significant attention in recent years due to their cost-effectiveness and competitive performance. The widespread adoption of reinforcement learning as a sequential placement method has introduced a new challenge in determining the optimal placement order for each chiplet. The order in which chiplets are placed on the interposer influences the spatial resources available for earlier and later placed chiplets, making the placement results highly sensitive to the sequence of chiplet placement. To address these challenges, we propose a learning to rank approach with graph representation, building upon the reinforcement learning framework RLPlanner. This method aims to select the optimal chiplet placement order for each chiplet-based system. Experimental results demonstrate that compared to placement order obtained solely based on the descending order of the chiplet area and the number of interconnect wires between the chiplets, utilizing the placement order obtained from the learning to rank network leads to further improvements in system temperature and inter-chiplet wirelength. Specifically, applying the top-ranked placement order obtained from the learning to rank network results in a 10.05% reduction in total inter-chiplet wirelength and a 1.01% improvement in peak system temperature during the chiplet placement process.
Authors: Bin Lu, Tingyan Ma, Xiaoying Gan, Xinbing Wang, Yunqiang Zhu, Chenghu Zhou, Shiyu Liang
Abstract: Graph Neural Networks (GNNs) are widely deployed in vast fields, but they often struggle to maintain accurate representations as graphs evolve. We theoretically establish a lower bound, proving that under mild conditions, representation distortion inevitably occurs over time. To estimate the temporal distortion without human annotation after deployment, one naive approach is to pre-train a recurrent model (e.g., RNN) before deployment and use this model afterwards, but the estimation is far from satisfactory. In this paper, we analyze the representation distortion from an information theory perspective, and attribute it primarily to inaccurate feature extraction during evolution. Consequently, we introduce Smart, a straightforward and effective baseline enhanced by an adaptive feature extractor through self-supervised graph reconstruction. In synthetic random graphs, we further refine the former lower bound to show the inevitable distortion over time and empirically observe that Smart achieves good estimation performance. Moreover, we observe that Smart consistently shows outstanding generalization estimation on four real-world evolving graphs. The ablation studies underscore the necessity of graph reconstruction. For example, on OGB-arXiv dataset, the estimation metric MAPE deteriorates from 2.19% to 8.00% without reconstruction.
Authors: Jinyi Xu, Zuowei Zhang, Ze Lin, Yixiang Chen, Zhe Liu, Weiping Ding
Abstract: It is still challenging to cluster multi-view data since existing methods can only assign an object to a specific (singleton) cluster when combining different view information. As a result, it fails to characterize imprecision of objects in overlapping regions of different clusters, thus leading to a high risk of errors. In this paper, we thereby want to answer the question: how to characterize imprecision in multi-view clustering? Correspondingly, we propose a multi-view low-rank evidential c-means based on entropy constraint (MvLRECM). The proposed MvLRECM can be considered as a multi-view version of evidential c-means based on the theory of belief functions. In MvLRECM, each object is allowed to belong to different clusters with various degrees of support (masses of belief) to characterize uncertainty when decision-making. Moreover, if an object is in the overlapping region of several singleton clusters, it can be assigned to a meta-cluster, defined as the union of these singleton clusters, to characterize the local imprecision in the result. In addition, entropy-weighting and low-rank constraints are employed to reduce imprecision and improve accuracy. Compared to state-of-the-art methods, the effectiveness of MvLRECM is demonstrated based on several toy and UCI real datasets.
Authors: Cangqing Wang, Yutian Yang, Ruisi Li, Dan Sun, Ruicong Cai, Yuzhu Zhang, Chengqian Fu, Lillian Floyd
Abstract: The rapid advancement of Large Language Models (LLMs) has inaugurated a transformative epoch in natural language processing, fostering unprecedented proficiency in text generation, comprehension, and contextual scrutiny. Nevertheless, effectively handling extensive contexts, crucial for myriad applications, poses a formidable obstacle owing to the intrinsic constraints of the models' context window sizes and the computational burdens entailed by their operations. This investigation presents an innovative framework that strategically tailors LLMs for streamlined context processing by harnessing the synergies among natural language summarization, soft prompt compression, and augmented utility preservation mechanisms. Our methodology, dubbed SoftPromptComp, amalgamates natural language prompts extracted from summarization methodologies with dynamically generated soft prompts to forge a concise yet semantically robust depiction of protracted contexts. This depiction undergoes further refinement via a weighting mechanism optimizing information retention and utility for subsequent tasks. We substantiate that our framework markedly diminishes computational overhead and enhances LLMs' efficacy across various benchmarks, while upholding or even augmenting the caliber of the produced content. By amalgamating soft prompt compression with sophisticated summarization, SoftPromptComp confronts the dual challenges of managing lengthy contexts and ensuring model scalability. Our findings point towards a propitious trajectory for augmenting LLMs' applicability and efficiency, rendering them more versatile and pragmatic for real-world applications. This research enriches the ongoing discourse on optimizing language models, providing insights into the potency of soft prompts and summarization techniques as pivotal instruments for the forthcoming generation of NLP solutions.
Authors: Weilin Cai, Juyong Jiang, Le Qin, Junwei Cui, Sunghun Kim, Jiayi Huang
Abstract: Expert parallelism has been introduced as a strategy to distribute the computational workload of sparsely-gated mixture-of-experts (MoE) models across multiple computing devices, facilitating the execution of these increasingly large-scale models. However, the All-to-All communication intrinsic to expert parallelism constitutes a significant overhead, diminishing the MoE models' efficiency. Current optimization approaches offer some relief, yet they are constrained by the sequential interdependence of communication and computation operations. To address this limitation, we present a novel shortcut-connected MoE architecture with overlapping parallel strategy, designated as ScMoE, which effectively decouples communication from its conventional sequence, allowing for a substantial overlap of 70% to 100% with computation. When compared with the prevalent top-2 MoE architecture, ScMoE demonstrates training speed improvements of 30% and 11%, and inference improvements of 40% and 15%, in our PCIe and NVLink hardware environments, respectively, where communication constitutes 60% and 15% of the total MoE time consumption. On the other hand, extensive experiments and theoretical analyses indicate that ScMoE not only achieves comparable but in some instances surpasses the model quality of existing approaches in vision and language tasks.
Authors: Bishwas Mandal, George Amariucai, Shuangqing Wei
Abstract: We propose a novel problem formulation to address the privacy-utility tradeoff, specifically when dealing with two distinct user groups characterized by unique sets of private and utility attributes. Unlike previous studies that primarily focus on scenarios where all users share identical private and utility attributes and often rely on auxiliary datasets or manual annotations, we introduce a collaborative data-sharing mechanism between two user groups through a trusted third party. This third party uses adversarial privacy techniques with our proposed data-sharing mechanism to internally sanitize data for both groups and eliminates the need for manual annotation or auxiliary datasets. Our methodology ensures that private attributes cannot be accurately inferred while enabling highly accurate predictions of utility features. Importantly, even if analysts or adversaries possess auxiliary datasets containing raw data, they are unable to accurately deduce private features. Additionally, our data-sharing mechanism is compatible with various existing adversarially trained privacy techniques. We empirically demonstrate the effectiveness of our approach using synthetic and real-world datasets, showcasing its ability to balance the conflicting goals of privacy and utility.
Authors: Bishwas Mandal, George Amariucai, Shuangqing Wei
Abstract: We investigate the application of large language models (LLMs), specifically GPT-4, to scenarios involving the tradeoff between privacy and utility in tabular data. Our approach entails prompting GPT-4 by transforming tabular data points into textual format, followed by the inclusion of precise sanitization instructions in a zero-shot manner. The primary objective is to sanitize the tabular data in such a way that it hinders existing machine learning models from accurately inferring private features while allowing models to accurately infer utility-related attributes. We explore various sanitization instructions. Notably, we discover that this relatively simple approach yields performance comparable to more complex adversarial optimization methods used for managing privacy-utility tradeoffs. Furthermore, while the prompts successfully obscure private features from the detection capabilities of existing machine learning models, we observe that this obscuration alone does not necessarily meet a range of fairness metrics. Nevertheless, our research indicates the potential effectiveness of LLMs in adhering to these fairness metrics, with some of our experimental results aligning with those achieved by well-established adversarial optimization techniques.
Authors: Haitong Ma, Zhaolin Ren, Bo Dai, Na Li
Abstract: We study sim-to-real skill transfer and discovery in the context of robotics control using representation learning. We draw inspiration from spectral decomposition of Markov decision processes. The spectral decomposition brings about representation that can linearly represent the state-action value function induced by any policies, thus can be regarded as skills. The skill representations are transferable across arbitrary tasks with the same transition dynamics. Moreover, to handle the sim-to-real gap in the dynamics, we propose a skill discovery algorithm that learns new skills caused by the sim-to-real gap from real-world data. We promote the discovery of new skills by enforcing orthogonal constraints between the skills to learn and the skills from simulators, and then synthesize the policy using the enlarged skill sets. We demonstrate our methodology by transferring quadrotor controllers from simulators to Crazyflie 2.1 quadrotors. We show that we can learn the skill representations from a single simulator task and transfer these to multiple different real-world tasks including hovering, taking off, landing and trajectory tracking. Our skill discovery approach helps narrow the sim-to-real gap and improve the real-world controller performance by up to 30.2%.
Authors: Elita A. Lobo, Cyrus Cousins, Yair Zick, Marek Petrik
Abstract: In reinforcement learning, robust policies for high-stakes decision-making problems with limited data are usually computed by optimizing the \emph{percentile criterion}. The percentile criterion is approximately solved by constructing an \emph{ambiguity set} that contains the true model with high probability and optimizing the policy for the worst model in the set. Since the percentile criterion is non-convex, constructing ambiguity sets is often challenging. Existing work uses \emph{Bayesian credible regions} as ambiguity sets, but they are often unnecessarily large and result in learning overly conservative policies. To overcome these shortcomings, we propose a novel Value-at-Risk based dynamic programming algorithm to optimize the percentile criterion without explicitly constructing any ambiguity sets. Our theoretical and empirical results show that our algorithm implicitly constructs much smaller ambiguity sets and learns less conservative robust policies.
Authors: Zhiyu Liang, Chen Liang, Zheng Liang, Hongzhi Wang, Bo Zheng
Abstract: Unsupervised (a.k.a. Self-supervised) representation learning (URL) has emerged as a new paradigm for time series analysis, because it has the ability to learn generalizable time series representation beneficial for many downstream tasks without using labels that are usually difficult to obtain. Considering that existing approaches have limitations in the design of the representation encoder and the learning objective, we have proposed Contrastive Shapelet Learning (CSL), the first URL method that learns the general-purpose shapelet-based representation through unsupervised contrastive learning, and shown its superior performance in several analysis tasks, such as time series classification, clustering, and anomaly detection. In this paper, we develop TimeCSL, an end-to-end system that makes full use of the general and interpretable shapelets learned by CSL to achieve explorable time series analysis in a unified pipeline. We introduce the system components and demonstrate how users interact with TimeCSL to solve different analysis tasks in the unified pipeline, and gain insight into their time series by exploring the learned shapelets and representation.
Authors: Wenlu Tang, Zicheng Liu
Abstract: The performance of machine learning models can be impacted by changes in data over time. A promising approach to address this challenge is invariant learning, with a particular focus on a method known as invariant risk minimization (IRM). This technique aims to identify a stable data representation that remains effective with out-of-distribution (OOD) data. While numerous studies have developed IRM-based methods adaptive to data augmentation scenarios, there has been limited attention on directly assessing how well these representations preserve their invariant performance under varying conditions. In our paper, we propose a novel method to evaluate invariant performance, specifically tailored for IRM-based methods. We establish a bridge between the conditional expectation of an invariant predictor across different environments through the likelihood ratio. Our proposed criterion offers a robust basis for evaluating invariant performance. We validate our approach with theoretical support and demonstrate its effectiveness through extensive numerical studies.These experiments illustrate how our method can assess the invariant performance of various representation techniques.
Authors: Zhiqiang Cai, Tong Ding, Min Liu, Xinyu Liu, Jianlin Xia
Abstract: In this paper, we propose a structure-guided Gauss-Newton (SgGN) method for solving least squares problems using a shallow ReLU neural network. The method effectively takes advantage of both the least squares structure and the neural network structure of the objective function. By categorizing the weights and biases of the hidden and output layers of the network as nonlinear and linear parameters, respectively, the method iterates back and forth between the nonlinear and linear parameters. The nonlinear parameters are updated by a damped Gauss-Newton method and the linear ones are updated by a linear solver. Moreover, at the Gauss-Newton step, a special form of the Gauss-Newton matrix is derived for the shallow ReLU neural network and is used for efficient iterations. It is shown that the corresponding mass and Gauss-Newton matrices in the respective linear and nonlinear steps are symmetric and positive definite under reasonable assumptions. Thus, the SgGN method naturally produces an effective search direction without the need of additional techniques like shifting in the Levenberg-Marquardt method to achieve invertibility of the Gauss-Newton matrix. The convergence and accuracy of the method are demonstrated numerically for several challenging function approximation problems, especially those with discontinuities or sharp transition layers that pose significant challenges for commonly used training algorithms in machine learning.
Authors: Sri Harsha Dumpala, Chandramouli Shama Sastry, Rudolf Uher, Sageev Oore
Abstract: Previous works on depression detection use datasets collected in similar environments to train and test the models. In practice, however, the train and test distributions cannot be guaranteed to be identical. Distribution shifts can be introduced due to variations such as recording environment (e.g., background noise) and demographics (e.g., gender, age, etc). Such distributional shifts can surprisingly lead to severe performance degradation of the depression detection models. In this paper, we analyze the application of test-time training (TTT) to improve robustness of models trained for depression detection. When compared to regular testing of the models, we find TTT can significantly improve the robustness of the model under a variety of distributional shifts introduced due to: (a) background-noise, (b) gender-bias, and (c) data collection and curation procedure (i.e., train and test samples are from separate datasets).
Authors: Vlad Fomenko, Han Yu, Jongho Lee, Stanley Hsieh, Weizhu Chen
Abstract: LoRA (Low-Rank Adaptation) has emerged as a preferred method for efficiently adapting Large Language Models (LLMs) with remarkable simplicity and efficacy. This note extends the original LoRA paper by offering new perspectives that were not initially discussed and presents a series of insights for deploying LoRA at scale. Without introducing new experiments, we aim to improve the understanding and application of LoRA.
Authors: Mohamed El Amine Seddik, Suei-Wen Chen, Soufiane Hayou, Pierre Youssef, Merouane Debbah
Abstract: The phenomenon of model collapse, introduced in (Shumailov et al., 2023), refers to the deterioration in performance that occurs when new models are trained on synthetic data generated from previously trained models. This recursive training loop makes the tails of the original distribution disappear, thereby making future-generation models forget about the initial (real) distribution. With the aim of rigorously understanding model collapse in language models, we consider in this paper a statistical model that allows us to characterize the impact of various recursive training scenarios. Specifically, we demonstrate that model collapse cannot be avoided when training solely on synthetic data. However, when mixing both real and synthetic data, we provide an estimate of a maximal amount of synthetic data below which model collapse can eventually be avoided. Our theoretical conclusions are further supported by empirical validations.
Authors: Shurui Gui, Xiner Li, Shuiwang Ji
Abstract: Test-time adaptation (TTA) addresses distribution shifts for streaming test data in unsupervised settings. Currently, most TTA methods can only deal with minor shifts and rely heavily on heuristic and empirical studies. To advance TTA under domain shifts, we propose the novel problem setting of active test-time adaptation (ATTA) that integrates active learning within the fully TTA setting. We provide a learning theory analysis, demonstrating that incorporating limited labeled test instances enhances overall performances across test domains with a theoretical guarantee. We also present a sample entropy balancing for implementing ATTA while avoiding catastrophic forgetting (CF). We introduce a simple yet effective ATTA algorithm, known as SimATTA, using real-time sample selection techniques. Extensive experimental results confirm consistency with our theoretical analyses and show that the proposed ATTA method yields substantial performance improvements over TTA methods while maintaining efficiency and shares similar effectiveness to the more demanding active domain adaptation (ADA) methods. Our code is available at https://github.com/divelab/ATTA
Authors: Ali Mortazavi, Junhao Lin, Nishant A. Mehta
Abstract: In one view of the classical game of prediction with expert advice with binary outcomes, in each round, each expert maintains an adversarially chosen belief and honestly reports this belief. We consider a recently introduced, strategic variant of this problem with selfish (reputation-seeking) experts, where each expert strategically reports in order to maximize their expected future reputation based on their belief. In this work, our goal is to design an algorithm for the selfish experts problem that is incentive-compatible (IC, or \emph{truthful}), meaning each expert's best strategy is to report truthfully, while also ensuring the algorithm enjoys sublinear regret with respect to the expert with the best belief. Freeman et al. (2020) recently studied this problem in the full information and bandit settings and obtained truthful, no-regret algorithms by leveraging prior work on wagering mechanisms. While their results under full information match the minimax rate for the classical ("honest experts") problem, the best-known regret for their bandit algorithm WSU-UX is $O(T^{2/3})$, which does not match the minimax rate for the classical ("honest bandits") setting. It was unclear whether the higher regret was an artifact of their analysis or a limitation of WSU-UX. We show, via explicit construction of loss sequences, that the algorithm suffers a worst-case $\Omega(T^{2/3})$ lower bound. Left open is the possibility that a different IC algorithm obtains $O(\sqrt{T})$ regret. Yet, WSU-UX was a natural choice for such an algorithm owing to the limited design room for IC algorithms in this setting.
Authors: Tham Yik Foong, Heng Zhang, Mao Po Yuan, Danilo Vasconcellos Vargas
Abstract: Input distribution shift presents a significant problem in many real-world systems. Here we present Xenovert, an adaptive algorithm that can dynamically adapt to changes in input distribution. It is a perfect binary tree that adaptively divides a continuous input space into several intervals of uniform density while receiving a continuous stream of input. This process indirectly maps the source distribution to the shifted target distribution, preserving the data's relationship with the downstream decoder/operation, even after the shift occurs. In this paper, we demonstrated how a neural network integrated with Xenovert achieved better results in 4 out of 5 shifted datasets, saving the hurdle of retraining a machine learning model. We anticipate that Xenovert can be applied to many more applications that require adaptation to unforeseen input distribution shifts, even when the distribution shift is drastic.
Authors: Chao Gao, Sai Qian Zhang
Abstract: To enhance the performance of large language models (LLM) on downstream tasks, one solution is to fine-tune certain LLM parameters and make it better align with the characteristics of the training dataset. This process is commonly known as parameter-efficient fine-tuning (PEFT). Due to the scale of LLM, PEFT operations are usually executed in the public environment (e.g., cloud server). This necessitates the sharing of sensitive user data across public environments, thereby raising potential privacy concerns. To tackle these challenges, we propose a distributed PEFT framework called DLoRA. DLoRA enables scalable PEFT operations to be performed collaboratively between the cloud and user devices. Coupled with the proposed Kill and Revive algorithm, the evaluation results demonstrate that DLoRA can significantly reduce the computation and communication workload over the user devices while achieving superior accuracy and privacy protection.
Authors: Hengyu Ye, Jiadong Chen, Shijin Gong, Fuxin Jiang, Tieying Zhang, Jianjun Chen, Xiaofeng Gao
Abstract: The intricate nature of time series data analysis benefits greatly from the distinct advantages offered by time and frequency domain representations. While the time domain is superior in representing local dependencies, particularly in non-periodic series, the frequency domain excels in capturing global dependencies, making it ideal for series with evident periodic patterns. To capitalize on both of these strengths, we propose ATFNet, an innovative framework that combines a time domain module and a frequency domain module to concurrently capture local and global dependencies in time series data. Specifically, we introduce Dominant Harmonic Series Energy Weighting, a novel mechanism for dynamically adjusting the weights between the two modules based on the periodicity of the input time series. In the frequency domain module, we enhance the traditional Discrete Fourier Transform (DFT) with our Extended DFT, designed to address the challenge of discrete frequency misalignment. Additionally, our Complex-valued Spectrum Attention mechanism offers a novel approach to discern the intricate relationships between different frequency combinations. Extensive experiments across multiple real-world datasets demonstrate that our ATFNet framework outperforms current state-of-the-art methods in long-term time series forecasting.
Authors: Naveen Karunanayake, Ravin Gunawardena, Suranga Seneviratne, Sanjay Chawla
Abstract: Deep neural networks (DNNs) deployed in real-world applications can encounter out-of-distribution (OOD) data and adversarial examples. These represent distinct forms of distributional shifts that can significantly impact DNNs' reliability and robustness. Traditionally, research has addressed OOD detection and adversarial robustness as separate challenges. This survey focuses on the intersection of these two areas, examining how the research community has investigated them together. Consequently, we identify two key research directions: robust OOD detection and unified robustness. Robust OOD detection aims to differentiate between in-distribution (ID) data and OOD data, even when they are adversarially manipulated to deceive the OOD detector. Unified robustness seeks a single approach to make DNNs robust against both adversarial attacks and OOD inputs. Accordingly, first, we establish a taxonomy based on the concept of distributional shifts. This framework clarifies how robust OOD detection and unified robustness relate to other research areas addressing distributional shifts, such as OOD detection, open set recognition, and anomaly detection. Subsequently, we review existing work on robust OOD detection and unified robustness. Finally, we highlight the limitations of the existing work and propose promising research directions that explore adversarial and OOD inputs within a unified framework.
Authors: Yi Yu, Brendan P. Malone, Luigi J. Renzullo
Abstract: The evaluation of modelled or satellite-derived soil moisture (SM) estimates is usually dependent on comparisons against in-situ SM measurements. However, the inherent mismatch in spatial support (i.e., scale) necessitates a cautious interpretation of point-to-pixel comparisons. The upscaling of the in-situ measurements to a commensurate resolution to that of the modelled or retrieved SM will lead to a fairer comparison and statistically more defensible evaluation. In this study, we presented an upscaling approach that combines spatiotemporal fusion with machine learning to extrapolate point-scale SM measurements from 28 in-situ sites to a 100 m resolution for an agricultural area of 100 km by 100 km. We conducted a four-fold cross-validation, which consistently demonstrated comparable correlation performance across folds, ranging from 0.6 to 0.9. The proposed approach was further validated based on a cross-cluster strategy by using two spatial subsets within the study area, denoted as cluster A and B, each of which equally comprised of 12 in-situ sites. The cross-cluster validation underscored the capability of the upscaling approach to map the spatial variability of SM within areas that were not covered by in-situ sites, with correlation performance ranging between 0.6 and 0.8. In general, our proposed upscaling approach offers an avenue to extrapolate point measurements of SM to a spatial scale more akin to climatic model grids or remotely sensed observations. Future investigations should delve into a further evaluation of the upscaling approach using independent data, such as model simulations, satellite retrievals or field campaign data.
Authors: Amin Aminifar, Baichuan Huang, Azra Abtahi, Amir Aminifar
Abstract: The human brain performs tasks with an outstanding energy-efficiency, i.e., with approximately 20 Watts. The state-of-the-art Artificial/Deep Neural Networks (ANN/DNN), on the other hand, have recently been shown to consume massive amounts of energy. The training of these ANNs/DNNs is done almost exclusively based on the back-propagation algorithm, which is known to be biologically implausible. This has led to a new generation of forward-only techniques, including the Forward-Forward algorithm. In this paper, we propose a lightweight inference scheme specifically designed for DNNs trained using the Forward-Forward algorithm. We have evaluated our proposed lightweight inference scheme in the case of the MNIST and CIFAR datasets, as well as two real-world applications, namely, epileptic seizure detection and cardiac arrhythmia classification using wearable technologies, where complexity overheads/energy consumption is a major constraint, and demonstrate its relevance.
Authors: Viet Quoc Vo, Ehsan Abbasnejad, Damith C. Ranasinghe
Abstract: We study the unique, less-well understood problem of generating sparse adversarial samples simply by observing the score-based replies to model queries. Sparse attacks aim to discover a minimum number-the l0 bounded-perturbations to model inputs to craft adversarial examples and misguide model decisions. But, in contrast to query-based dense attack counterparts against black-box models, constructing sparse adversarial perturbations, even when models serve confidence score information to queries in a score-based setting, is non-trivial. Because, such an attack leads to i) an NP-hard problem; and ii) a non-differentiable search space. We develop the BruSLeAttack-a new, faster (more query-efficient) Bayesian algorithm for the problem. We conduct extensive attack evaluations including an attack demonstration against a Machine Learning as a Service (MLaaS) offering exemplified by Google Cloud Vision and robustness testing of adversarial training regimes and a recent defense against black-box attacks. The proposed attack scales to achieve state-of-the-art attack success rates and query efficiency on standard computer vision tasks such as ImageNet across different model architectures. Our artefacts and DIY attack samples are available on GitHub. Importantly, our work facilitates faster evaluation of model vulnerabilities and raises our vigilance on the safety, security and reliability of deployed systems.
Authors: Tim K. Smit, Hajo A. Reijers, Xixi Lu
Abstract: Predictive Process Monitoring focuses on predicting future states of ongoing process executions, such as forecasting the remaining time. Recent developments in Object-Centric Process Mining have enriched event data with objects and their explicit relations between events. To leverage this enriched data, we propose the Heterogeneous Object Event Graph encoding (HOEG), which integrates events and objects into a graph structure with diverse node types. It does so without aggregating object features, thus creating a more nuanced and informative representation. We then adopt a heterogeneous Graph Neural Network architecture, which incorporates these diverse object features in prediction tasks. We evaluate the performance and scalability of HOEG in predicting remaining time, benchmarking it against two established graph-based encodings and two baseline models. Our evaluation uses three Object-Centric Event Logs (OCELs), including one from a real-life process at a major Dutch financial institution. The results indicate that HOEG competes well with existing models and surpasses them when OCELs contain informative object attributes and event-object interactions.
Authors: Hao Ma, Melanie Zeilinger, Michael Muehlebach
Abstract: We propose a novel gradient-based online optimization framework for solving stochastic programming problems that frequently arise in the context of cyber-physical and robotic systems. Our problem formulation accommodates constraints that model the evolution of a cyber-physical system, which has, in general, a continuous state and action space, is nonlinear, and where the state is only partially observed. We also incorporate an approximate model of the dynamics as prior knowledge into the learning process and show that even rough estimates of the dynamics can significantly improve the convergence of our algorithms. Our online optimization framework encompasses both gradient descent and quasi-Newton methods, and we provide a unified convergence analysis of our algorithms in a non-convex setting. We also characterize the impact of modeling errors in the system dynamics on the convergence rate of the algorithms. Finally, we evaluate our algorithms in simulations of a flexible beam, a four-legged walking robot, and in real-world experiments with a ping-pong playing robot.
Authors: Antonio Giganti, Sara Mandelli, Paolo Bestagini, Umberto Giuriato, Alessandro D'Ausilio, Marco Marcon, Stefano Tubaro
Abstract: Due to the latest environmental concerns in keeping at bay contaminants emissions in urban areas, air pollution forecasting has been rising the forefront of all researchers around the world. When predicting pollutant concentrations, it is common to include the effects of environmental factors that influence these concentrations within an extended period, like traffic, meteorological conditions and geographical information. Most of the existing approaches exploit this information as past covariates, i.e., past exogenous variables that affected the pollutant but were not affected by it. In this paper, we present a novel forecasting methodology to predict NO$_2$ concentration via both past and future covariates. Future covariates are represented by weather forecasts and future calendar events, which are already known at prediction time. In particular, we deal with air quality observations in a city-wide network of ground monitoring stations, modeling the data structure and estimating the predictions with a Spatiotemporal Graph Neural Network (STGNN). We propose a conditioning block that embeds past and future covariates into the current observations. After extracting meaningful spatiotemporal representations, these are fused together and projected into the forecasting horizon to generate the final prediction. To the best of our knowledge, it is the first time that future covariates are included in time series predictions in a structured way. Remarkably, we find that conditioning on future weather information has a greater impact than considering past traffic conditions. We release our code implementation at https://github.com/polimi-ispl/MAGCRN.
Authors: Chengyan Fu, Wenjie Wang
Abstract: Randomized smoothing is the primary certified robustness method for accessing the robustness of deep learning models to adversarial perturbations in the l2-norm, by adding isotropic Gaussian noise to the input image and returning the majority votes over the base classifier. Theoretically, it provides a certified norm bound, ensuring predictions of adversarial examples are stable within this bound. A notable constraint limiting widespread adoption is the necessity to retrain base models entirely from scratch to attain a robust version. This is because the base model fails to learn the noise-augmented data distribution to give an accurate vote. One intuitive way to overcome this challenge is to involve a custom-trained denoiser to eliminate the noise. However, this approach is inefficient and sub-optimal. Inspired by recent large model training procedures, we explore an alternative way named PEFTSmoothing to adapt the base model to learn the Gaussian noise-augmented data with Parameter-Efficient Fine-Tuning (PEFT) methods in both white-box and black-box settings. Extensive results demonstrate the effectiveness and efficiency of PEFTSmoothing, which allow us to certify over 98% accuracy for ViT on CIFAR-10, 20% higher than SoTA denoised smoothing, and over 61% accuracy on ImageNet which is 30% higher than CNN-based denoiser and comparable to the Diffusion-based denoiser.
Authors: Qi Li, Xianjun Zeng, Shuliang Wang, Wenhao Zhu, Shijie Ruan, Zhimeng Yuan
Abstract: Missing datasets, in which some objects have missing values in certain dimensions, are prevalent in the Real-world. Existing clustering algorithms for missing datasets first impute the missing values and then perform clustering. However, both the imputation and clustering processes require input parameters. Too many input parameters inevitably increase the difficulty of obtaining accurate clustering results. Although some studies have shown that decision graphs can replace the input parameters of clustering algorithms, current decision graphs require equivalent dimensions among objects and are therefore not suitable for missing datasets. To this end, we propose a Single-Dimensional Clustering algorithm, i.e., SDC. SDC, which removes the imputation process and adapts the decision graph to the missing datasets by splitting dimension and partition intersection fusion, can obtain valid clustering results on the missing datasets without input parameters. Experiments demonstrate that, across three evaluation metrics, SDC outperforms baseline algorithms by at least 13.7%(NMI), 23.8%(ARI), and 8.1%(Purity).
Authors: Jintu Borah, Tanujit Chakraborty, Md. Shahrul Md. Nadzir, Mylene G. Cayetano, Shubhankar Majumdar
Abstract: Accurate and reliable air quality forecasting is essential for protecting public health, sustainable development, pollution control, and enhanced urban planning. This letter presents a novel WaveCatBoost architecture designed to forecast the real-time concentrations of air pollutants by combining the maximal overlapping discrete wavelet transform (MODWT) with the CatBoost model. This hybrid approach efficiently transforms time series into high-frequency and low-frequency components, thereby extracting signal from noise and improving prediction accuracy and robustness. Evaluation of two distinct regional datasets, from the Central Air Pollution Control Board (CPCB) sensor network and a low-cost air quality sensor system (LAQS), underscores the superior performance of our proposed methodology in real-time forecasting compared to the state-of-the-art statistical and deep learning architectures. Moreover, we employ a conformal prediction strategy to provide probabilistic bands with our forecasts.
Authors: Xin Li
Abstract: The conventional wisdom of manifold learning is based on nonlinear dimensionality reduction techniques such as IsoMAP and locally linear embedding (LLE). We challenge this paradigm by exploiting the blessing of dimensionality. Our intuition is simple: it is easier to untangle a low-dimensional manifold in a higher-dimensional space due to its vastness, as guaranteed by Whitney embedding theorem. A new insight brought by this work is to introduce class labels as the context variables in the lifted higher-dimensional space (so supervised learning becomes unsupervised learning). We rigorously show that manifold untangling leads to linearly separable classifiers in the lifted space. To correct the inevitable overfitting, we consider the dual process of manifold untangling -- tangling or aliasing -- which is important for generalization. Using context as the bonding element, we construct a pair of manifold untangling and tangling operators, known as tangling-untangling cycle (TUC). Untangling operator maps context-independent representations (CIR) in low-dimensional space to context-dependent representations (CDR) in high-dimensional space by inducing context as hidden variables. The tangling operator maps CDR back to CIR by a simple integral transformation for invariance and generalization. We also present the hierarchical extensions of TUC based on the Cartesian product and the fractal geometry. Despite the conceptual simplicity, TUC admits a biologically plausible and energy-efficient implementation based on the time-locking behavior of polychronization neural groups (PNG) and sleep-wake cycle (SWC). The TUC-based theory applies to the computational modeling of various cognitive functions by hippocampal-neocortical systems.
Authors: Tejas Kasetty, Divyat Mahajan, Gintare Karolina Dziugaite, Alexandre Drouin, Dhanya Sridhar
Abstract: Numerous decision-making tasks require estimating causal effects under interventions on different parts of a system. As practitioners consider using large language models (LLMs) to automate decisions, studying their causal reasoning capabilities becomes crucial. A recent line of work evaluates LLMs ability to retrieve commonsense causal facts, but these evaluations do not sufficiently assess how LLMs reason about interventions. Motivated by the role that interventions play in causal inference, in this paper, we conduct empirical analyses to evaluate whether LLMs can accurately update their knowledge of a data-generating process in response to an intervention. We create benchmarks that span diverse causal graphs (e.g., confounding, mediation) and variable types, and enable a study of intervention-based reasoning. These benchmarks allow us to isolate the ability of LLMs to accurately predict changes resulting from their ability to memorize facts or find other shortcuts. Our analysis on four LLMs highlights that while GPT- 4 models show promising accuracy at predicting the intervention effects, they remain sensitive to distracting factors in the prompts.
Authors: Seungyub Han, Yeongmo Kim, Taehyun Cho, Jungwoo Lee
Abstract: One of the objectives of continual learning is to prevent catastrophic forgetting in learning multiple tasks sequentially, and the existing solutions have been driven by the conceptualization of the plasticity-stability dilemma. However, the convergence of continual learning for each sequential task is less studied so far. In this paper, we provide a convergence analysis of memory-based continual learning with stochastic gradient descent and empirical evidence that training current tasks causes the cumulative degradation of previous tasks. We propose an adaptive method for nonconvex continual learning (NCCL), which adjusts step sizes of both previous and current tasks with the gradients. The proposed method can achieve the same convergence rate as the SGD method when the catastrophic forgetting term which we define in the paper is suppressed at each iteration. Further, we demonstrate that the proposed algorithm improves the performance of continual learning over existing methods for several image classification tasks.
Authors: Bowen Pan, Yikang Shen, Haokun Liu, Mayank Mishra, Gaoyuan Zhang, Aude Oliva, Colin Raffel, Rameswar Panda
Abstract: Mixture-of-Experts (MoE) language models can reduce computational costs by 2-4$\times$ compared to dense models without sacrificing performance, making them more efficient in computation-bounded scenarios. However, MoE models generally require 2-4$\times$ times more parameters to achieve comparable performance to a dense model, which incurs larger GPU memory requirements and makes MoE models less efficient in I/O-bounded scenarios like autoregressive generation. In this work, we propose a hybrid dense training and sparse inference framework for MoE models (DS-MoE) which achieves strong computation and parameter efficiency by employing dense computation across all experts during training and sparse computation during inference. Our experiments on training LLMs demonstrate that our DS-MoE models are more parameter-efficient than standard sparse MoEs and are on par with dense models in terms of total parameter size and performance while being computationally cheaper (activating 30-40% of the model's parameters). Performance tests using vLLM show that our DS-MoE-6B model runs up to $1.86\times$ faster than similar dense models like Mistral-7B, and between $1.50\times$ and $1.71\times$ faster than comparable MoEs, such as DeepSeekMoE-16B and Qwen1.5-MoE-A2.7B.
Authors: Shuai Guo, Jielei Chu, Lei Zhu, Tianrui Li
Abstract: Generative Flow Networks (GFlowNets) are probabilistic models predicated on Markov flows, employing specific amortization algorithms to learn stochastic policies that generate compositional substances including biomolecules, chemical materials, and more. Demonstrating formidable prowess in generating high-performance biochemical molecules, GFlowNets accelerate the discovery of scientific substances, effectively circumventing the time-consuming, labor-intensive, and costly shortcomings intrinsic to conventional material discovery. However, previous work often struggles to accumulate exploratory experience and is prone to becoming disoriented within expansive sampling spaces. Attempts to address this issue, such as LS-GFN, are limited to local greedy searches and lack broader global adjustments. This paper introduces a novel GFlowNet variant, the Dynamic Backtracking GFN (DB-GFN), which enhances the adaptability of decision-making steps through a reward-based dynamic backtracking mechanism. DB-GFN permits backtracking during the network construction process according to the current state's reward value, thus correcting disadvantageous decisions and exploring alternative pathways during the exploration process. Applied to generative tasks of biochemical molecules and genetic material sequences, DB-GFN surpasses existing GFlowNet models and traditional reinforcement learning methods in terms of sample quality, exploration sample quantity, and training convergence speed. Furthermore, the orthogonal nature of DB-GFN suggests its potential as a powerful tool for future improvements in GFN networks, with the promise of integrating with other strategies to achieve more efficient search performance.
Authors: Artem Vysogorets, Kartik Ahuja, Julia Kempe
Abstract: In the era of exceptionally data-hungry models, careful selection of the training data is essential to mitigate the extensive costs of deep learning. Data pruning offers a solution by removing redundant or uninformative samples from the dataset, which yields faster convergence and improved neural scaling laws. However, little is known about its impact on classification bias of the trained models. We conduct the first systematic study of this effect and reveal that existing data pruning algorithms can produce highly biased classifiers. At the same time, we argue that random data pruning with appropriate class ratios has potential to improve the worst-class performance. We propose a "fairness-aware" approach to pruning and empirically demonstrate its performance on standard computer vision benchmarks. In sharp contrast to existing algorithms, our proposed method continues improving robustness at a tolerable drop of average performance as we prune more from the datasets. We present theoretical analysis of the classification risk in a mixture of Gaussians to further motivate our algorithm and support our findings.
Authors: Zihan Pengmei, Zimu Li
Abstract: Graph Transformers have emerged as a powerful alternative to Message-Passing Graph Neural Networks (MP-GNNs) to address limitations such as over-squashing of information exchange. However, incorporating graph inductive bias into transformer architectures remains a significant challenge. In this report, we propose the Graph Spectral Token, a novel approach to directly encode graph spectral information, which captures the global structure of the graph, into the transformer architecture. By parameterizing the auxiliary [CLS] token and leaving other tokens representing graph nodes, our method seamlessly integrates spectral information into the learning process. We benchmark the effectiveness of our approach by enhancing two existing graph transformers, GraphTrans and SubFormer. The improved GraphTrans, dubbed GraphTrans-Spec, achieves over 10% improvements on large graph benchmark datasets while maintaining efficiency comparable to MP-GNNs. SubFormer-Spec demonstrates strong performance across various datasets.
Authors: Ao Zhou, Jianlei Yang, Tong Qiao, Yingjie Qi, Zhi Yang, Weisheng Zhao, Chunming Hu
Abstract: The key to device-edge co-inference paradigm is to partition models into computation-friendly and computation-intensive parts across the device and the edge, respectively. However, for Graph Neural Networks (GNNs), we find that simply partitioning without altering their structures can hardly achieve the full potential of the co-inference paradigm due to various computational-communication overheads of GNN operations over heterogeneous devices. We present GCoDE, the first automatic framework for GNN that innovatively Co-designs the architecture search and the mapping of each operation on Device-Edge hierarchies. GCoDE abstracts the device communication process into an explicit operation and fuses the search of architecture and the operations mapping in a unified space for joint-optimization. Also, the performance-awareness approach, utilized in the constraint-based search process of GCoDE, enables effective evaluation of architecture efficiency in diverse heterogeneous systems. We implement the co-inference engine and runtime dispatcher in GCoDE to enhance the deployment efficiency. Experimental results show that GCoDE can achieve up to $44.9\times$ speedup and $98.2\%$ energy reduction compared to existing approaches across various applications and system configurations.
Authors: Suiyao Chen, Xinyi Liu, Yulei Li, Jing Wu, Handong Yao
Abstract: As the aging population grows, particularly for the baby boomer generation, the United States is witnessing a significant increase in the elderly population experiencing multifunctional disabilities. These disabilities, stemming from a variety of chronic diseases, injuries, and impairments, present a complex challenge due to their multidimensional nature, encompassing both physical and cognitive aspects. Traditional methods often use univariate regression-based methods to model and predict single degradation conditions and assume population homogeneity, which is inadequate to address the complexity and diversity of aging-related degradation. This study introduces a novel framework for multi-functional degradation modeling that captures the multidimensional (e.g., physical and cognitive) and heterogeneous nature of elderly disabilities. Utilizing deep learning, our approach predicts health degradation scores and uncovers latent heterogeneity from elderly health histories, offering both efficient estimation and explainable insights into the diverse effects and causes of aging-related degradation. A real-case study demonstrates the effectiveness and marks a pivotal contribution to accurately modeling the intricate dynamics of elderly degradation, and addresses the healthcare challenges in the aging population.
Authors: Pietro Lesci, Andreas Vlachos
Abstract: Active learning for imbalanced classification tasks is challenging as the minority classes naturally occur rarely. Gathering a large pool of unlabelled data is thus essential to capture minority instances. Standard pool-based active learning is computationally expensive on large pools and often reaches low accuracy by overfitting the initial decision boundary, thus failing to explore the input space and find minority instances. To address these issues we propose AnchorAL. At each iteration, AnchorAL chooses class-specific instances from the labelled set, or anchors, and retrieves the most similar unlabelled instances from the pool. This resulting subpool is then used for active learning. Using a small, fixed-sized subpool AnchorAL allows scaling any active learning strategy to large pools. By dynamically selecting different anchors at each iteration it promotes class balance and prevents overfitting the initial decision boundary, thus promoting the discovery of new clusters of minority instances. Experiments across different classification tasks, active learning strategies, and model architectures AnchorAL is (i) faster, often reducing runtime from hours to minutes, (ii) trains more performant models, (iii) and returns more balanced datasets than competing methods.
Authors: Qun Li, Yuan Meng, Chen Tang, Jiacheng Jiang, Zhi Wang
Abstract: Quantization is a promising technique for reducing the bit-width of deep models to improve their runtime performance and storage efficiency, and thus becomes a fundamental step for deployment. In real-world scenarios, quantized models are often faced with adversarial attacks which cause the model to make incorrect inferences by introducing slight perturbations. However, recent studies have paid less attention to the impact of quantization on the model robustness. More surprisingly, existing studies on this topic even present inconsistent conclusions, which prompted our in-depth investigation. In this paper, we conduct a first-time analysis of the impact of the quantization pipeline components that can incorporate robust optimization under the settings of Post-Training Quantization and Quantization-Aware Training. Through our detailed analysis, we discovered that this inconsistency arises from the use of different pipelines in different studies, specifically regarding whether robust optimization is performed and at which quantization stage it occurs. Our research findings contribute insights into deploying more secure and robust quantized networks, assisting practitioners in reference for scenarios with high-security requirements and limited resources.
Authors: Miguel Costa, Sandro Pinto
Abstract: ML is shifting from the cloud to the edge. Edge computing reduces the surface exposing private data and enables reliable throughput guarantees in real-time applications. Of the panoply of devices deployed at the edge, resource-constrained MCUs, e.g., Arm Cortex-M, are more prevalent, orders of magnitude cheaper, and less power-hungry than application processors or GPUs. Thus, enabling intelligence at the deep edge is the zeitgeist, with researchers focusing on unveiling novel approaches to deploy ANNs on these constrained devices. Quantization is a well-established technique that has proved effective in enabling the deployment of neural networks on MCUs; however, it is still an open question to understand the robustness of QNNs in the face of adversarial examples. To fill this gap, we empirically evaluate the effectiveness of attacks and defenses from (full-precision) ANNs on (constrained) QNNs. Our evaluation includes three QNNs targeting TinyML applications, ten attacks, and six defenses. With this study, we draw a set of interesting findings. First, quantization increases the point distance to the decision boundary and leads the gradient estimated by some attacks to explode or vanish. Second, quantization can act as a noise attenuator or amplifier, depending on the noise magnitude, and causes gradient misalignment. Regarding adversarial defenses, we conclude that input pre-processing defenses show impressive results on small perturbations; however, they fall short as the perturbation increases. At the same time, train-based defenses increase the average point distance to the decision boundary, which holds after quantization. However, we argue that train-based defenses still need to smooth the quantization-shift and gradient misalignment phenomenons to counteract adversarial example transferability to QNNs. All artifacts are open-sourced to enable independent validation of results.
Authors: Sannyuya Liu, Qing Li, Xiaoxuan Shen, Jianwen Sun, Zongkai Yang
Abstract: Skill acquisition is a key area of research in cognitive psychology as it encompasses multiple psychological processes. The laws discovered under experimental paradigms are controversial and lack generalizability. This paper aims to unearth the laws of skill learning from large-scale training log data. A two-stage algorithm was developed to tackle the issues of unobservable cognitive states and algorithmic explosion in searching. Initially a deep learning model is employed to determine the learner's cognitive state and assess the feature importance. Subsequently, symbolic regression algorithms are utilized to parse the neural network model into algebraic equations. The experimental results of simulated data demonstrate that the proposed algorithm can accurately restore various preset laws within a certain range of noise, in continues feedback setting. Application of proposed method to Lumosity training data demonstrates superior performance compared to traditional and latest models in terms of fitness. The results indicate the discovery of two new forms of skill acquisition laws, while some previous findings have been reaffirmed.
Authors: Talha Hanif Butt, Prayag Tiwari, Fernando Alonso-Fernandez
Abstract: Safe overtakes in trucks are crucial to prevent accidents, reduce congestion, and ensure efficient traffic flow, making early prediction essential for timely and informed driving decisions. Accordingly, we investigate the detection of truck overtakes from CAN data. Three classifiers, Artificial Neural Networks (ANN), Random Forest, and Support Vector Machines (SVM), are employed for the task. Our analysis covers up to 10 seconds before the overtaking event, using an overlapping sliding window of 1 second to extract CAN features. We observe that the prediction scores of the overtake class tend to increase as we approach the overtake trigger, while the no-overtake class remain stable or oscillates depending on the classifier. Thus, the best accuracy is achieved when approaching the trigger, making early overtaking prediction challenging. The classifiers show good accuracy in classifying overtakes (Recall/TPR > 93%), but accuracy is suboptimal in classifying no-overtakes (TNR typically 80-90% and below 60% for one SVM variant). We further combine two classifiers (Random Forest and linear SVM) by averaging their output scores. The fusion is observed to improve no-overtake classification (TNR > 92%) at the expense of reducing overtake accuracy (TPR). However, the latter is kept above 91% near the overtake trigger. Therefore, the fusion balances TPR and TNR, providing more consistent performance than individual classifiers.
Authors: Lucas Lingle
Abstract: Large neural network models have become a mainstay of natural language processing and computer vision, yet their initialization and learning rates are set in a largely heuristic fashion, potentially varying from paper to paper and one model size to the next. The $\mu$-Parameterization ($\mu$P) offers a potential solution to these challenges, yielding scaling rules for model initialization and learning rates, and reportedly enabling zero-shot hyperparameter transfer from small to large models in a variety of cases. Despite the evident promise, the $\mu$P scaling rules are not yet widely adopted, perhaps due to higher implementation complexity, many variations, or complex theoretical background. This work investigates $\mu$P empirically, focusing on the ubiquitous transformer architecture, and aims to answer a simple question: does $\mu$-Transfer yield optimal learning rates in practice? From models with 2M to 10B parameters, we show that $\mu$-Transfer works as intended for the majority of important cases, but also identify some surprising cases where it may not.
Authors: Yining Wu, Shengyu Duan, Gaole Sai, Chenhong Cao, Guobing Zou
Abstract: Matrix factorization (MF) is a widely used collaborative filtering (CF) algorithm for recommendation systems (RSs), due to its high prediction accuracy, great flexibility and high efficiency in big data processing. However, with the dramatically increased number of users/items in current RSs, the computational complexity for training a MF model largely increases. Many existing works have accelerated MF, by either putting in additional computational resources or utilizing parallel systems, introducing a large cost. In this paper, we propose algorithmic methods to accelerate MF, without inducing any additional computational resources. In specific, we observe fine-grained structured sparsity in the decomposed feature matrices when considering a certain threshold. The fine-grained structured sparsity causes a large amount of unnecessary operations during both matrix multiplication and latent factor update, increasing the computational time of the MF training process. Based on the observation, we firstly propose to rearrange the feature matrices based on joint sparsity, which potentially makes a latent vector with a smaller index more dense than that with a larger index. The feature matrix rearrangement is given to limit the error caused by the later performed pruning process. We then propose to prune the insignificant latent factors by an early stopping process during both matrix multiplication and latent factor update. The pruning process is dynamically performed according to the sparsity of the latent factors for different users/items, to accelerate the process. The experiments show that our method can achieve 1.2-1.65 speedups, with up to 20.08% error increase, compared with the conventional MF training process. We also prove the proposed methods are applicable considering different hyperparameters including optimizer, optimization strategy and initialization method.
Authors: Joachim Baumann, Celestine Mendler-D\"unner
Abstract: We investigate algorithmic collective action in transformer-based recommender systems. Our use case is a collective of fans aiming to promote the visibility of an artist by strategically placing one of their songs in the existing playlists they control. The success of the collective is measured by the increase in test-time recommendations of the targeted song. We introduce two easily implementable strategies towards this goal and test their efficacy on a publicly available recommender system model released by a major music streaming platform. Our findings reveal that even small collectives (controlling less than 0.01% of the training data) can achieve up 25x amplification of recommendations by strategically choosing the position at which to insert the song. We then focus on investigating the externalities of the strategy. We find that the performance loss for the platform is negligible, and the recommendations of other songs are largely preserved, minimally impairing the user experience of participants. Moreover, the costs are evenly distributed among other artists. Taken together, our findings demonstrate how collective action strategies can be effective while not necessarily being adversarial, raising new questions around incentives, social dynamics, and equilibria in recommender systems.
Authors: Yassaman Ebrahimzadeh Maboud, Muhammad Adnan, Divya Mahajan, Prashant J. Nair
Abstract: Training recommendation models pose significant challenges regarding resource utilization and performance. Prior research has proposed an approach that categorizes embeddings into popular and non-popular classes to reduce the training time for recommendation models. We observe that, even among the popular embeddings, certain embeddings undergo rapid training and exhibit minimal subsequent variation, resulting in saturation. Consequently, updates to these embeddings lack any contribution to model quality. This paper presents Slipstream, a software framework that identifies stale embeddings on the fly and skips their updates to enhance performance. This capability enables Slipstream to achieve substantial speedup, optimize CPU-GPU bandwidth usage, and eliminate unnecessary memory access. SlipStream showcases training time reductions of 2x, 2.4x, 1.2x, and 1.175x across real-world datasets and configurations, compared to Baseline XDL, Intel-optimized DRLM, FAE, and Hotline, respectively.
Authors: Diego Vallarino
Abstract: By integrating survival analysis, machine learning algorithms, and economic interpretation, this research examines the temporal dynamics associated with attaining a 5 percent rise in purchasing power parity-adjusted GDP per capita over a period of 120 months (2013-2022). A comparative investigation reveals that DeepSurv is proficient at capturing non-linear interactions, although standard models exhibit comparable performance under certain circumstances. The weight matrix evaluates the economic ramifications of vulnerabilities, risks, and capacities. In order to meet the GDPpc objective, the findings emphasize the need of a balanced approach to risk-taking, strategic vulnerability reduction, and investment in governmental capacities and social cohesiveness. Policy guidelines promote individualized approaches that take into account the complex dynamics at play while making decisions.
Authors: Pratim Saha, Chengcui Zhang
Abstract: Translation-based Video Synthesis (TVS) has emerged as a vital research area in computer vision, aiming to facilitate the transformation of videos between distinct domains while preserving both temporal continuity and underlying content features. This technique has found wide-ranging applications, encompassing video super-resolution, colorization, segmentation, and more, by extending the capabilities of traditional image-to-image translation to the temporal domain. One of the principal challenges faced in TVS is the inherent risk of introducing flickering artifacts and inconsistencies between frames during the synthesis process. This is particularly challenging due to the necessity of ensuring smooth and coherent transitions between video frames. Efforts to tackle this challenge have induced the creation of diverse strategies and algorithms aimed at mitigating these unwanted consequences. This comprehensive review extensively examines the latest progress in the realm of TVS. It thoroughly investigates emerging methodologies, shedding light on the fundamental concepts and mechanisms utilized for proficient video synthesis. This survey also illuminates their inherent strengths, limitations, appropriate applications, and potential avenues for future development.
Authors: Yi Ren, Shangmin Guo, Linlu Qiu, Bailin Wang, Danica J. Sutherland
Abstract: With the widespread adoption of Large Language Models (LLMs), the prevalence of iterative interactions among these models is anticipated to increase. Notably, recent advancements in multi-round self-improving methods allow LLMs to generate new examples for training subsequent models. At the same time, multi-agent LLM systems, involving automated interactions among agents, are also increasing in prominence. Thus, in both short and long terms, LLMs may actively engage in an evolutionary process. We draw parallels between the behavior of LLMs and the evolution of human culture, as the latter has been extensively studied by cognitive scientists for decades. Our approach involves leveraging Iterated Learning (IL), a Bayesian framework that elucidates how subtle biases are magnified during human cultural evolution, to explain some behaviors of LLMs. This paper outlines key characteristics of agents' behavior in the Bayesian-IL framework, including predictions that are supported by experimental verification with various LLMs. This theoretical framework could help to more effectively predict and guide the evolution of LLMs in desired directions.
Authors: Nitesh Goyal, Minsuk Chang, Michael Terry
Abstract: Our ability to build autonomous agents that leverage Generative AI continues to increase by the day. As builders and users of such agents it is unclear what parameters we need to align on before the agents start performing tasks on our behalf. To discover these parameters, we ran a qualitative empirical research study about designing agents that can negotiate during a fictional yet relatable task of selling a camera online. We found that for an agent to perform the task successfully, humans/users and agents need to align over 6 dimensions: 1) Knowledge Schema Alignment 2) Autonomy and Agency Alignment 3) Operational Alignment and Training 4) Reputational Heuristics Alignment 5) Ethics Alignment and 6) Human Engagement Alignment. These empirical findings expand previous work related to process and specification alignment and the need for values and safety in Human-AI interactions. Subsequently we discuss three design directions for designers who are imagining a world filled with Human-Agent collaborations.
Authors: Hainan Xu, Zhehuai Chen, Fei Jia, Boris Ginsburg
Abstract: This paper proposes Transducers with Pronunciation-aware Embeddings (PET). Unlike conventional Transducers where the decoder embeddings for different tokens are trained independently, the PET model's decoder embedding incorporates shared components for text tokens with the same or similar pronunciations. With experiments conducted in multiple datasets in Mandarin Chinese and Korean, we show that PET models consistently improve speech recognition accuracy compared to conventional Transducers. Our investigation also uncovers a phenomenon that we call error chain reactions. Instead of recognition errors being evenly spread throughout an utterance, they tend to group together, with subsequent errors often following earlier ones. Our analysis shows that PET models effectively mitigate this issue by substantially reducing the likelihood of the model generating additional errors following a prior one. Our implementation will be open-sourced with the NeMo toolkit.
Authors: Dongwei Jiang, Jingyu Zhang, Orion Weller, Nathaniel Weir, Benjamin Van Durme, Daniel Khashabi
Abstract: Can LLMs continually improve their previous outputs for better results? An affirmative answer would require LLMs to be better at discriminating among previously-generated alternatives, than generating initial responses. We explore the validity of this hypothesis in practice. We first introduce a unified framework that allows us to compare the generative and discriminative capability of any model on any task. Then, in our resulting experimental analysis of several LLMs, we do not observe the performance of those models on discrimination to be reliably better than generation. We hope these findings inform the growing literature on self-improvement AI systems.
Authors: Wenguan Wang, Yi Yang, Yunhe Pan
Abstract: Visual knowledge is a new form of knowledge representation that can encapsulate visual concepts and their relations in a succinct, comprehensive, and interpretable manner, with a deep root in cognitive psychology. As the knowledge about the visual world has been identified as an indispensable component of human cognition and intelligence, visual knowledge is poised to have a pivotal role in establishing machine intelligence. With the recent advance of Artificial Intelligence (AI) techniques, large AI models (or foundation models) have emerged as a potent tool capable of extracting versatile patterns from broad data as implicit knowledge, and abstracting them into an outrageous amount of numeric parameters. To pave the way for creating visual knowledge empowered AI machines in this coming wave, we present a timely review that investigates the origins and development of visual knowledge in the pre-big model era, and accentuates the opportunities and unique role of visual knowledge in the big model era.
Authors: Nikolay Kalmykov, Rishat Zagidullin, Oleg Rogov, Sergey Rykovanov, Dmitry V. Dylov
Abstract: Modulation instability is a phenomenon of spontaneous pattern formation in nonlinear media, oftentimes leading to an unpredictable behaviour and a degradation of a signal of interest. We propose an approach based on reinforcement learning to suppress the unstable modes by optimizing the parameters for the time modulation of the potential in the nonlinear system. We test our approach in 1D and 2D cases and propose a new class of physically-meaningful reward functions to guarantee tamed instability.
Authors: Wenxuan Zuo, Zifan Zhu, Yuxuan Du, Yi-Chun Yeh, Jed A. Fuhrman, Jinchi Lv, Yingying Fan, Fengzhu Sun
Abstract: High-dimensional longitudinal time series data is prevalent across various real-world applications. Many such applications can be modeled as regression problems with high-dimensional time series covariates. Deep learning has been a popular and powerful tool for fitting these regression models. Yet, the development of interpretable and reproducible deep-learning models is challenging and remains underexplored. This study introduces a novel method, Deep Learning Inference using Knockoffs for Time series data (DeepLINK-T), focusing on the selection of significant time series variables in regression while controlling the false discovery rate (FDR) at a predetermined level. DeepLINK-T combines deep learning with knockoff inference to control FDR in feature selection for time series models, accommodating a wide variety of feature distributions. It addresses dependencies across time and features by leveraging a time-varying latent factor structure in time series covariates. Three key ingredients for DeepLINK-T are 1) a Long Short-Term Memory (LSTM) autoencoder for generating time series knockoff variables, 2) an LSTM prediction network using both original and knockoff variables, and 3) the application of the knockoffs framework for variable selection with FDR control. Extensive simulation studies have been conducted to evaluate DeepLINK-T's performance, showing its capability to control FDR effectively while demonstrating superior feature selection power for high-dimensional longitudinal time series data compared to its non-time series counterpart. DeepLINK-T is further applied to three metagenomic data sets, validating its practical utility and effectiveness, and underscoring its potential in real-world applications.
Authors: Yangqiaoyu Zhou, Haokun Liu, Tejes Srivastava, Hongyuan Mei, Chenhao Tan
Abstract: Effective generation of novel hypotheses is instrumental to scientific progress. So far, researchers have been the main powerhouse behind hypothesis generation by painstaking data analysis and thinking (also known as the Eureka moment). In this paper, we examine the potential of large language models (LLMs) to generate hypotheses. We focus on hypothesis generation based on data (i.e., labeled examples). To enable LLMs to handle arbitrarily long contexts, we generate initial hypotheses from a small number of examples and then update them iteratively to improve the quality of hypotheses. Inspired by multi-armed bandits, we design a reward function to inform the exploitation-exploration tradeoff in the update process. Our algorithm is able to generate hypotheses that enable much better predictive performance than few-shot prompting in classification tasks, improving accuracy by 31.7% on a synthetic dataset and by 13.9%, 3.3% and, 24.9% on three real-world datasets. We also outperform supervised learning by 12.8% and 11.2% on two challenging real-world datasets. Furthermore, we find that the generated hypotheses not only corroborate human-verified theories but also uncover new insights for the tasks.
Authors: Truman Yuen, Graham A. Watt, Yuri Lawryshyn
Abstract: Generative Large Language Models enable efficient analytics across knowledge domains, rivalling human experts in information comparisons. However, the applications of LLMs for information comparisons face scalability challenges due to the difficulties in maintaining information across large contexts and overcoming model token limitations. To address these challenges, we developed the novel Abstractive Summarization \& Criteria-driven Comparison Endpoint (ASC$^2$End) system to automate information comparison at scale. Our system employs Semantic Text Similarity comparisons for generating evidence-supported analyses. We utilize proven data-handling strategies such as abstractive summarization and retrieval augmented generation to overcome token limitations and retain relevant information during model inference. Prompts were designed using zero-shot strategies to contextualize information for improved model reasoning. We evaluated abstractive summarization using ROUGE scoring and assessed the generated comparison quality using survey responses. Models evaluated on the ASC$^2$End system show desirable results providing insights on the expected performance of the system. ASC$^2$End is a novel system and tool that enables accurate, automated information comparison at scale across knowledge domains, overcoming limitations in context length and retrieval.
Authors: GianCarlo Catalano, Alexander E. I. Brownlee, David Cairns, John McCall, Russell Ainslie
Abstract: Genetic Algorithms have established their capability for solving many complex optimization problems. Even as good solutions are produced, the user's understanding of a problem is not necessarily improved, which can lead to a lack of confidence in the results. To mitigate this issue, explainability aims to give insight to the user by presenting them with the knowledge obtained by the algorithm. In this paper we introduce Partial Solutions in order to improve the explainability of solutions to combinatorial optimization problems. Partial Solutions represent beneficial traits found by analyzing a population, and are presented to the user for explainability, but also provide an explicit model from which new solutions can be generated. We present an algorithm that assembles a collection of Partial Solutions chosen to strike a balance between high fitness, simplicity and atomicity. Experiments with standard benchmarks show that the proposed algorithm is able to find Partial Solutions which improve explainability at reasonable computational cost without affecting search performance.
Authors: Andy Yang, David Chiang
Abstract: Deriving formal bounds on the expressivity of transformers, as well as studying transformers that are constructed to implement known algorithms, are both effective methods for better understanding the computational power of transformers. Towards both ends, we introduce the temporal counting logic $\textbf{K}_\text{t}$[#] alongside the RASP variant $\textbf{C-RASP}$. We show they are equivalent to each other, and that together they are the best-known lower bound on the formal expressivity of future-masked soft attention transformers with unbounded input size. We prove this by showing all $\textbf{K}_\text{t}$[#] formulas can be compiled into these transformers. As a case study, we demonstrate on paper how to use $\textbf{C-RASP}$ to construct simple transformer language models that, using greedy decoding, can only generate sentences that have given properties formally specified in $\textbf{K}_\text{t}$[#].
Authors: Toru Shirakawa, Yi Li, Yulun Wu, Sky Qiu, Yuxuan Li, Mingduo Zhao, Hiroyasu Iso, Mark van der Laan
Abstract: We propose Deep Longitudinal Targeted Minimum Loss-based Estimation (Deep LTMLE), a novel approach to estimate the counterfactual mean of outcome under dynamic treatment policies in longitudinal problem settings. Our approach utilizes a transformer architecture with heterogeneous type embedding trained using temporal-difference learning. After obtaining an initial estimate using the transformer, following the targeted minimum loss-based likelihood estimation (TMLE) framework, we statistically corrected for the bias commonly associated with machine learning algorithms. Furthermore, our method also facilitates statistical inference by enabling the provision of 95% confidence intervals grounded in asymptotic statistical theory. Simulation results demonstrate our method's superior performance over existing approaches, particularly in complex, long time-horizon scenarios. It remains effective in small-sample, short-duration contexts, matching the performance of asymptotically efficient estimators. To demonstrate our method in practice, we applied our method to estimate counterfactual mean outcomes for standard versus intensive blood pressure management strategies in a real-world cardiovascular epidemiology cohort study.
Authors: Igor Cardoso, Rubens O. Moraes, Lucas N. Ferreira
Abstract: Neural models are one of the most popular approaches for music generation, yet there aren't standard large datasets tailored for learning music directly from game data. To address this research gap, we introduce a novel dataset named NES-VMDB, containing 98,940 gameplay videos from 389 NES games, each paired with its original soundtrack in symbolic format (MIDI). NES-VMDB is built upon the Nintendo Entertainment System Music Database (NES-MDB), encompassing 5,278 music pieces from 397 NES games. Our approach involves collecting long-play videos for 389 games of the original dataset, slicing them into 15-second-long clips, and extracting the audio from each clip. Subsequently, we apply an audio fingerprinting algorithm (similar to Shazam) to automatically identify the corresponding piece in the NES-MDB dataset. Additionally, we introduce a baseline method based on the Controllable Music Transformer to generate NES music conditioned on gameplay clips. We evaluated this approach with objective metrics, and the results showed that the conditional CMT improves musical structural quality when compared to its unconditional counterpart. Moreover, we used a neural classifier to predict the game genre of the generated pieces. Results showed that the CMT generator can learn correlations between gameplay videos and game genres, but further research has to be conducted to achieve human-level performance.
Authors: Danielle Van Boxel
Abstract: We apply Bayesian Additive Regression Tree (BART) principles to training an ensemble of small neural networks for regression tasks. Using Markov Chain Monte Carlo, we sample from the posterior distribution of neural networks that have a single hidden layer. To create an ensemble of these, we apply Gibbs sampling to update each network against the residual target value (i.e. subtracting the effect of the other networks). We demonstrate the effectiveness of this technique on several benchmark regression problems, comparing it to equivalent shallow neural networks, BART, and ordinary least squares. Our Bayesian Additive Regression Networks (BARN) provide more consistent and often more accurate results. On test data benchmarks, BARN averaged between 5 to 20 percent lower root mean square error. This error performance does come at the cost, however, of greater computation time. BARN sometimes takes on the order of a minute where competing methods take a second or less. But, BARN without cross-validated hyperparameter tuning takes about the same amount of computation time as tuned other methods. Yet BARN is still typically more accurate.
Authors: Selim Furkan Tekin, Fatih Ilhan, Tiansheng Huang, Sihao Hu, Ka-Ho Chow, Margaret L. Loper, Ling Liu
Abstract: This paper presents FusionShot, a focal diversity optimized few-shot ensemble learning approach for boosting the robustness and generalization performance of pre-trained few-shot models. The paper makes three original contributions. First, we explore the unique characteristics of few-shot learning to ensemble multiple few-shot (FS) models by creating three alternative fusion channels. Second, we introduce the concept of focal error diversity to learn the most efficient ensemble teaming strategy, rather than assuming that an ensemble of a larger number of base models will outperform those sub-ensembles of smaller size. We develop a focal-diversity ensemble pruning method to effectively prune out the candidate ensembles with low ensemble error diversity and recommend top-$K$ FS ensembles with the highest focal error diversity. Finally, we capture the complex non-linear patterns of ensemble few-shot predictions by designing the learn-to-combine algorithm, which can learn the diverse weight assignments for robust ensemble fusion over different member models. Extensive experiments on representative few-shot benchmarks show that the top-K ensembles recommended by FusionShot can outperform the representative SOTA few-shot models on novel tasks (different distributions and unknown at training), and can prevail over existing few-shot learners in both cross-domain settings and adversarial settings. For reproducibility purposes, FusionShot trained models, results, and code are made available at https://github.com/sftekin/fusionshot
Authors: Krishna Subramani, Paris Smaragdis, Takuya Higuchi, Mehrez Souden
Abstract: Non-negative Matrix Factorization (NMF) is a powerful technique for analyzing regularly-sampled data, i.e., data that can be stored in a matrix. For audio, this has led to numerous applications using time-frequency (TF) representations like the Short-Time Fourier Transform. However extending these applications to irregularly-spaced TF representations, like the Constant-Q transform, wavelets, or sinusoidal analysis models, has not been possible since these representations cannot be directly stored in matrix form. In this paper, we formulate NMF in terms of continuous functions (instead of fixed vectors) and show that NMF can be extended to a wider variety of signal classes that need not be regularly sampled.
Authors: Muhammad Asad, Ihsan Ullah, Ganesh Sistu, Michael G. Madden
Abstract: In novelty detection, the goal is to decide if a new data point should be categorized as an inlier or an outlier, given a training dataset that primarily captures the inlier distribution. Recent approaches typically use deep encoder and decoder network frameworks to derive a reconstruction error, and employ this error either to determine a novelty score, or as the basis for a one-class classifier. In this research, we use a similar framework but with a lightweight deep network, and we adopt a probabilistic score with reconstruction error. Our methodology calculates the probability of whether the sample comes from the inlier distribution or not. This work makes two key contributions. The first is that we compute the novelty probability by linearizing the manifold that holds the structure of the inlier distribution. This allows us to interpret how the probability is distributed and can be determined in relation to the local coordinates of the manifold tangent space. The second contribution is that we improve the training protocol for the network. Our results indicate that our approach is effective at learning the target class, and it outperforms recent state-of-the-art methods on several benchmark datasets.
Authors: Sentao Miao, Yining Wang
Abstract: This paper proposes a practically efficient algorithm with optimal theoretical regret which solves the classical network revenue management (NRM) problem with unknown, nonparametric demand. Over a time horizon of length $T$, in each time period the retailer needs to decide prices of $N$ types of products which are produced based on $M$ types of resources with unreplenishable initial inventory. When demand is nonparametric with some mild assumptions, Miao and Wang (2021) is the first paper which proposes an algorithm with $O(\text{poly}(N,M,\ln(T))\sqrt{T})$ type of regret (in particular, $\tilde O(N^{3.5}\sqrt{T})$ plus additional high-order terms that are $o(\sqrt{T})$ with sufficiently large $T\gg N$). In this paper, we improve the previous result by proposing a primal-dual optimization algorithm which is not only more practical, but also with an improved regret of $\tilde O(N^{3.25}\sqrt{T})$ free from additional high-order terms. A key technical contribution of the proposed algorithm is the so-called demand balancing, which pairs the primal solution (i.e., the price) in each time period with another price to offset the violation of complementary slackness on resource inventory constraints. Numerical experiments compared with several benchmark algorithms further illustrate the effectiveness of our algorithm.
Authors: Jing Du, Zesheng Ye, Bin Guo, Zhiwen Yu, Lina Yao
Abstract: Cross-Domain Recommendation (CDR) seeks to enable effective knowledge transfer across domains. Existing works rely on either representation alignment or transformation bridges, but they struggle on identifying domain-shared from domain-specific latent factors. Specifically, while CDR describes user representations as a joint distribution over two domains, these methods fail to account for its joint identifiability as they primarily fixate on the marginal distribution within a particular domain. Such a failure may overlook the conditionality between two domains and how it contributes to latent factor disentanglement, leading to negative transfer when domains are weakly correlated. In this study, we explore what should and should not be transferred in cross-domain user representations from a causality perspective. We propose a Hierarchical subspace disentanglement approach to explore the Joint IDentifiability of cross-domain joint distribution, termed HJID, to preserve domain-specific behaviors from domain-shared factors. HJID organizes user representations into layers: generic shallow subspaces and domain-oriented deep subspaces. We first encode the generic pattern in the shallow subspace by minimizing the Maximum Mean Discrepancy of initial layer activation. Then, to dissect how domain-oriented latent factors are encoded in deeper layers activation, we construct a cross-domain causality-based data generation graph, which identifies cross-domain consistent and domain-specific components, adhering to the Minimal Change principle. This allows HJID to maintain stability whilst discovering unique factors for different domains, all within a generative framework of invertible transformations that guarantee the joint identifiability. With experiments on real-world datasets, we show that HJID outperforms SOTA methods on a range of strongly and weakly correlated CDR tasks.
Authors: Suk Yee Yong, K. E. Harborne, Caroline Foster, Robert Bassett, Gregory B. Poole, Mitchell Cavanagh
Abstract: Since the turn of the century, astronomers have been exploiting the rich information afforded by combining stellar kinematic maps and imaging in an attempt to recover the intrinsic, three-dimensional (3D) shape of a galaxy. A common intrinsic shape recovery method relies on an expected monotonic relationship between the intrinsic misalignment of the kinematic and morphological axes and the triaxiality parameter. Recent studies have, however, cast doubt about underlying assumptions relating shape and intrinsic kinematic misalignment. In this work, we aim to recover the 3D shape of individual galaxies using their projected stellar kinematic and flux distributions using a supervised machine learning approach with mixture density network (MDN). Using a mock dataset of the EAGLE hydrodynamical cosmological simulation, we train the MDN model for a carefully selected set of common kinematic and photometric parameters. Compared to previous methods, we demonstrate potential improvements achieved with the MDN model to retrieve the 3D galaxy shape along with the uncertainties, especially for prolate and triaxial systems. We make specific recommendations for recovering galaxy intrinsic shapes relevant for current and future integral field spectroscopic galaxy surveys.
Authors: Tomoya Wakayama
Abstract: The remarkable generalization performance of overparameterized models has challenged the conventional wisdom of statistical learning theory. While recent theoretical studies have shed light on this behavior in linear models or nonlinear classifiers, a comprehensive understanding of overparameterization in nonlinear regression remains lacking. This paper explores the predictive properties of overparameterized nonlinear regression within the Bayesian framework, extending the methodology of adaptive prior based on the intrinsic spectral structure of the data. We establish posterior contraction for single-neuron models with Lipschitz continuous activation functions and for generalized linear models, demonstrating that our approach achieves consistent predictions in the overparameterized regime. Moreover, our Bayesian framework allows for uncertainty estimation of the predictions. The proposed method is validated through numerical simulations and a real data application, showcasing its ability to achieve accurate predictions and reliable uncertainty estimates. Our work advances the theoretical understanding of the blessing of overparameterization and offers a principled Bayesian approach for prediction in large nonlinear models.
Authors: Suppakit Waiwitlikhit, Ion Stoica, Yi Sun, Tatsunori Hashimoto, Daniel Kang
Abstract: There is an increasing conflict between business incentives to hide models and data as trade secrets, and the societal need for algorithmic transparency. For example, a rightsholder wishing to know whether their copyrighted works have been used during training must convince the model provider to allow a third party to audit the model and data. Finding a mutually agreeable third party is difficult, and the associated costs often make this approach impractical. In this work, we show that it is possible to simultaneously allow model providers to keep their model weights (but not architecture) and data secret while allowing other parties to trustlessly audit model and data properties. We do this by designing a protocol called ZkAudit in which model providers publish cryptographic commitments of datasets and model weights, alongside a zero-knowledge proof (ZKP) certifying that published commitments are derived from training the model. Model providers can then respond to audit requests by privately computing any function F of the dataset (or model) and releasing the output of F alongside another ZKP certifying the correct execution of F. To enable ZkAudit, we develop new methods of computing ZKPs for SGD on modern neural nets for simple recommender systems and image classification models capable of high accuracies on ImageNet. Empirically, we show it is possible to provide trustless audits of DNNs, including copyright, censorship, and counterfactual audits with little to no loss in accuracy.
Authors: Shreyasi Mandal, Ashutosh Modi
Abstract: Large Language models (LLMs) have demonstrated state-of-the-art performance in various natural language processing (NLP) tasks across multiple domains, yet they are prone to shortcut learning and factual inconsistencies. This research investigates LLMs' robustness, consistency, and faithful reasoning when performing Natural Language Inference (NLI) on breast cancer Clinical Trial Reports (CTRs) in the context of SemEval 2024 Task 2: Safe Biomedical Natural Language Inference for Clinical Trials. We examine the reasoning capabilities of LLMs and their adeptness at logical problem-solving. A comparative analysis is conducted on pre-trained language models (PLMs), GPT-3.5, and Gemini Pro under zero-shot settings using Retrieval-Augmented Generation (RAG) framework, integrating various reasoning chains. The evaluation yields an F1 score of 0.69, consistency of 0.71, and a faithfulness score of 0.90 on the test dataset.
Authors: Udvas Basak, Rajarshi Dutta, Shivam Pandey, Ashutosh Modi
Abstract: This paper describes our system developed for the SemEval-2024 Task 1: Semantic Textual Relatedness. The challenge is focused on automatically detecting the degree of relatedness between pairs of sentences for 14 languages including both high and low-resource Asian and African languages. Our team participated in two subtasks consisting of Track A: supervised and Track B: unsupervised. This paper focuses on a BERT-based contrastive learning and similarity metric based approach primarily for the supervised track while exploring autoencoders for the unsupervised track. It also aims on the creation of a bigram relatedness corpus using negative sampling strategy, thereby producing refined word embeddings.
Authors: Shreenaga Chikoti, Shrey Mehta, Ashutosh Modi
Abstract: Memes are one of the most popular types of content used in an online disinformation campaign. They are primarily effective on social media platforms since they can easily reach many users. Memes in a disinformation campaign achieve their goal of influencing the users through several rhetorical and psychological techniques, such as causal oversimplification, name-calling, and smear. The SemEval 2024 Task 4 \textit{Multilingual Detection of Persuasion Technique in Memes} on identifying such techniques in the memes is divided across three sub-tasks: ($\mathbf{1}$) Hierarchical multi-label classification using only textual content of the meme, ($\mathbf{2}$) Hierarchical multi-label classification using both, textual and visual content of the meme and ($\mathbf{3}$) Binary classification of whether the meme contains a persuasion technique or not using it's textual and visual content. This paper proposes an ensemble of Class Definition Prediction (CDP) and hyperbolic embeddings-based approaches for this task. We enhance meme classification accuracy and comprehensiveness by integrating HypEmo's hierarchical label embeddings (Chen et al., 2023) and a multi-task learning framework for emotion prediction. We achieve a hierarchical F1-score of 0.60, 0.67, and 0.48 on the respective sub-tasks.
Authors: Zhiyuan Peng, Xuyang Wu, Qifan Wang, Sravanthi Rajanala, Yi Fang
Abstract: Parameter Efficient Fine-Tuning (PEFT) methods have been extensively utilized in Large Language Models (LLMs) to improve the down-streaming tasks without the cost of fine-tuing the whole LLMs. Recent studies have shown how to effectively use PEFT for fine-tuning LLMs in ranking tasks with convincing performance; there are some limitations, including the learned prompt being fixed for different documents, overfitting to specific tasks, and low adaptation ability. In this paper, we introduce a query-dependent parameter efficient fine-tuning (Q-PEFT) approach for text reranking to leak the information of the true queries to LLMs and then make the generation of true queries from input documents much easier. Specifically, we utilize the query to extract the top-$k$ tokens from concatenated documents, serving as contextual clues. We further augment Q-PEFT by substituting the retrieval mechanism with a multi-head attention layer to achieve end-to-end training and cover all the tokens in the documents, guiding the LLMs to generate more document-specific synthetic queries, thereby further improving the reranking performance. Extensive experiments are conducted on four public datasets, demonstrating the effectiveness of our proposed approach.
Authors: Shubham Patel, Divyaksh Shukla, Ashutosh Modi
Abstract: This paper presents our approach for the SemEval-2024 Task 10: Emotion Discovery and Reasoning its Flip in Conversations. For the Emotion Recognition in Conversations (ERC) task, we utilize a masked-memory network along with speaker participation. We propose a transformer-based speaker-centric model for the Emotion Flip Reasoning (EFR) task. We also introduce Probable Trigger Zone, a region of the conversation that is more likely to contain the utterances causing the emotion to flip. For sub-task 3, the proposed approach achieves a 5.9 (F1 score) improvement over the task baseline. The ablation study results highlight the significance of various design choices in the proposed method.
Authors: Xubin Wang, Yunhe Wang, Zhiqing Ma, Ka-Chun Wong, Xiangtao Li
Abstract: Accurate screening of cancer types is crucial for effective cancer detection and precise treatment selection. However, the association between gene expression profiles and tumors is often limited to a small number of biomarker genes. While computational methods using nature-inspired algorithms have shown promise in selecting predictive genes, existing techniques are limited by inefficient search and poor generalization across diverse datasets. This study presents a framework termed Evolutionary Optimized Diverse Ensemble Learning (EODE) to improve ensemble learning for cancer classification from gene expression data. The EODE methodology combines an intelligent grey wolf optimization algorithm for selective feature space reduction, guided random injection modeling for ensemble diversity enhancement, and subset model optimization for synergistic classifier combinations. Extensive experiments were conducted across 35 gene expression benchmark datasets encompassing varied cancer types. Results demonstrated that EODE obtained significantly improved screening accuracy over individual and conventionally aggregated models. The integrated optimization of advanced feature selection, directed specialized modeling, and cooperative classifier ensembles helps address key challenges in current nature-inspired approaches. This provides an effective framework for robust and generalized ensemble learning with gene expression biomarkers. Specifically, we have opened EODE source code on Github at https://github.com/wangxb96/EODE.
Authors: A. Martina Neuman, Philipp Christian Petersen
Abstract: We study the learning problem associated with spiking neural networks. Specifically, we consider hypothesis sets of spiking neural networks with affine temporal encoders and decoders and simple spiking neurons having only positive synaptic weights. We demonstrate that the positivity of the weights continues to enable a wide range of expressivity results, including rate-optimal approximation of smooth functions or approximation without the curse of dimensionality. Moreover, positive-weight spiking neural networks are shown to depend continuously on their parameters which facilitates classical covering number-based generalization statements. Finally, we observe that from a generalization perspective, contrary to feedforward neural networks or previous results for general spiking neural networks, the depth has little to no adverse effect on the generalization capabilities.
Authors: Felicia Lo, Shin-Ming Cheng, Rafael Kaliski
Abstract: Malware intrusion is problematic for Internet of Things (IoT) and Artificial Intelligence of Things (AIoT) devices as they often reside in an ecosystem of connected devices, such as a smart home. If any devices are infected, the whole ecosystem can be compromised. Although various Machine Learning (ML) models are deployed to detect malware and network intrusion, generally speaking, robust high-accuracy models tend to require resources not found in all IoT devices, compared to less robust models defined by weak learners. In order to combat this issue, Fadhilla proposed a meta-learner ensemble model comprised of less robust prediction results inherent with weak learner ML models to produce a highly robust meta-learning ensemble model. The main problem with the prior research is that it cannot be deployed in low-end AIoT devices due to the limited resources comprising processing power, storage, and memory (the required libraries quickly exhaust low-end AIoT devices' resources.) Hence, this research aims to optimize the proposed super learner meta-learning ensemble model to make it viable for low-end AIoT devices. We show the library and ML model memory requirements associated with each optimization stage and emphasize that optimization of current ML models is necessitated for low-end AIoT devices. Our results demonstrate that we can obtain similar accuracy and False Positive Rate (FPR) metrics from high-end AIoT devices running the derived ML model, with a lower inference duration and smaller memory footprint.
Authors: Rajat Sarkar, Krishna Sai Sudhir Aripirala, Vishal Sudam Jadhav, Sagar Srinivas Sakhinana, Venkataramana Runkana
Abstract: Computational Fluid Dynamics (CFD) serves as a powerful tool for simulating fluid flow across diverse industries. High-resolution CFD simulations offer valuable insights into fluid behavior and flow patterns, aiding in optimizing design features or enhancing system performance. However, as resolution increases, computational data requirements and time increase proportionately. This presents a persistent challenge in CFD. Recently, efforts have been directed towards accurately predicting fine-mesh simulations using coarse-mesh simulations, with geometry and boundary conditions as input. Drawing inspiration from models designed for super-resolution, deep learning techniques like UNets have been applied to address this challenge. However, these existing methods are limited to structured data and fail if the mesh is unstructured due to its inability to convolute. Additionally, incorporating geometry/mesh information in the training process introduces drawbacks such as increased data requirements, challenges in generalizing to unseen geometries for the same physical phenomena, and issues with robustness to mesh distortions. To address these concerns, we propose a novel framework, PointSAGE a mesh-independent network that leverages the unordered, mesh-less nature of Pointcloud to learn the complex fluid flow and directly predict fine simulations, completely neglecting mesh information. Utilizing an adaptable framework, the model accurately predicts the fine data across diverse point cloud sizes, regardless of the training dataset's dimension. We have evaluated the effectiveness of PointSAGE on diverse datasets in different scenarios, demonstrating notable results and a significant acceleration in computational time in generating fine simulations compared to standard CFD techniques.
Authors: Ashok Mondal (Vellore Institute of Technology, Chennai), Satyam Singh (Vellore Institute of Technology, Chennai)
Abstract: In recent years, large-scale adoption of cloud storage solutions has revolutionized the way we think about digital data storage. However, the exponential increase in data volume, especially images, has raised environmental concerns regarding power and resource consumption, as well as the rising digital carbon footprint emissions. The aim of this research is to propose a methodology for cloud-based image storage by integrating image compression technology with SuperResolution Generative Adversarial Networks (SRGAN). Rather than storing images in their original format directly on the cloud, our approach involves initially reducing the image size through compression and downsizing techniques before storage. Upon request, these compressed images will be retrieved and processed by SRGAN to generate images. The efficacy of the proposed method is evaluated in terms of PSNR and SSIM metrics. Additionally, a mathematical analysis is given to calculate power consumption and carbon footprint assesment. The proposed data compression technique provides a significant solution to achieve a reasonable trade off between environmental sustainability and industrial efficiency.
Authors: Yingting Li, Rishabh Bhardwaj, Ambuj Mehrish, Bo Cheng, Soujanya Poria
Abstract: Neural speech synthesis, or text-to-speech (TTS), aims to transform a signal from the text domain to the speech domain. While developing TTS architectures that train and test on the same set of speakers has seen significant improvements, out-of-domain speaker performance still faces enormous limitations. Domain adaptation on a new set of speakers can be achieved by fine-tuning the whole model for each new domain, thus making it parameter-inefficient. This problem can be solved by Adapters that provide a parameter-efficient alternative to domain adaptation. Although famous in NLP, speech synthesis has not seen much improvement from Adapters. In this work, we present HyperTTS, which comprises a small learnable network, "hypernetwork", that generates parameters of the Adapter blocks, allowing us to condition Adapters on speaker representations and making them dynamic. Extensive evaluations of two domain adaptation settings demonstrate its effectiveness in achieving state-of-the-art performance in the parameter-efficient regime. We also compare different variants of HyperTTS, comparing them with baselines in different studies. Promising results on the dynamic adaptation of adapter parameters using hypernetworks open up new avenues for domain-generic multi-speaker TTS systems. The audio samples and code are available at https://github.com/declare-lab/HyperTTS.
Authors: Nicolas Yax, Pierre-Yves Oudeyer, Stefano Palminteri
Abstract: This paper introduces PhyloLM, a method applying phylogenetic algorithms to Large Language Models to explore their finetuning relationships, and predict their performance characteristics. By leveraging the phylogenetic distance metric, we construct dendrograms, which satisfactorily capture distinct LLM families (across a set of 77 open-source and 22 closed models). Furthermore, phylogenetic distance predicts performances in benchmarks (we test MMLU and ARC), thus enabling a time and cost-effective estimation of LLM capabilities. The approach translates genetic concepts to machine learning, offering tools to infer LLM development, relationships, and capabilities, even in the absence of transparent training information.
Authors: Ziyuan Qu, Omkar Vengurlekar, Mohamad Qadri, Kevin Zhang, Michael Kaess, Christopher Metzler, Suren Jayasuriya, Adithya Pediredla
Abstract: Differentiable 3D-Gaussian splatting (GS) is emerging as a prominent technique in computer vision and graphics for reconstructing 3D scenes. GS represents a scene as a set of 3D Gaussians with varying opacities and employs a computationally efficient splatting operation along with analytical derivatives to compute the 3D Gaussian parameters given scene images captured from various viewpoints. Unfortunately, capturing surround view ($360^{\circ}$ viewpoint) images is impossible or impractical in many real-world imaging scenarios, including underwater imaging, rooms inside a building, and autonomous navigation. In these restricted baseline imaging scenarios, the GS algorithm suffers from a well-known 'missing cone' problem, which results in poor reconstruction along the depth axis. In this manuscript, we demonstrate that using transient data (from sonars) allows us to address the missing cone problem by sampling high-frequency data along the depth axis. We extend the Gaussian splatting algorithms for two commonly used sonars and propose fusion algorithms that simultaneously utilize RGB camera data and sonar data. Through simulations, emulations, and hardware experiments across various imaging scenarios, we show that the proposed fusion algorithms lead to significantly better novel view synthesis (5 dB improvement in PSNR) and 3D geometry reconstruction (60% lower Chamfer distance).
Authors: Gianluca Detommaso, Martin Bertran, Riccardo Fogliato, Aaron Roth
Abstract: This paper proposes the use of "multicalibration" to yield interpretable and reliable confidence scores for outputs generated by large language models (LLMs). Multicalibration asks for calibration not just marginally, but simultaneously across various intersecting groupings of the data. We show how to form groupings for prompt/completion pairs that are correlated with the probability of correctness via two techniques: clustering within an embedding space, and "self-annotation" - querying the LLM by asking it various yes-or-no questions about the prompt. We also develop novel variants of multicalibration algorithms that offer performance improvements by reducing their tendency to overfit. Through systematic benchmarking across various question answering datasets and LLMs, we show how our techniques can yield confidence scores that provide substantial improvements in fine-grained measures of both calibration and accuracy compared to existing methods.
Authors: Poorvi Joshi (National University of Singapore), Alakesh Kalita (Singapore University of Technology and Design), Mohan Gurusamy (National University of Singapore)
Abstract: Unmanned Aerial Vehicles (UAVs) are integral in various sectors like agriculture, surveillance, and logistics, driven by advancements in 5G. However, existing research lacks a comprehensive approach addressing both data freshness and security concerns. In this paper, we address the intricate challenges of data freshness, and security, especially in the context of eavesdropping and jamming in modern UAV networks. Our framework incorporates exponential AoI metrics and emphasizes secrecy rate to tackle eavesdropping and jamming threats. We introduce a transformer-enhanced Deep Reinforcement Learning (DRL) approach to optimize task offloading processes. Comparative analysis with existing algorithms showcases the superiority of our scheme, indicating its promising advancements in UAV network management.
Authors: Pawanesh Yadav, Charu Sharma, Niteesh Sahni
Abstract: This paper presents an analysis of the Indian stock market using a method based on embedding the network in a hyperbolic space using Machine learning techniques. We claim novelty on four counts. First, it is demonstrated that the hyperbolic clusters resemble the topological network communities more closely than the Euclidean clusters. Second, we are able to clearly distinguish between periods of market stability and volatility through a statistical analysis of hyperbolic distance and hyperbolic shortest path distance corresponding to the embedded network. Third, we demonstrate that using the modularity of the embedded network significant market changes can be spotted early. Lastly, the coalescent embedding is able to segregate the certain market sectors thereby underscoring its natural clustering ability.
Authors: Iury B. de A. Santos, Andr\'e C. P. L. F. de Carvalho
Abstract: The adoption of Deep Learning algorithms in the medical imaging field is a prominent area of research, with high potential for advancing AI-based Computer-aided diagnosis (AI-CAD) solutions. However, current solutions face challenges due to a lack of interpretability features and high data demands, prompting recent efforts to address these issues. In this study, we propose the ProtoAL method, where we integrate an interpretable DL model into the Deep Active Learning (DAL) framework. This approach aims to address both challenges by focusing on the medical imaging context and utilizing an inherently interpretable model based on prototypes. We evaluated ProtoAL on the Messidor dataset, achieving an area under the precision-recall curve of 0.79 while utilizing only 76.54\% of the available labeled data. These capabilities can enhances the practical usability of a DL model in the medical field, providing a means of trust calibration in domain experts and a suitable solution for learning in the data scarcity context often found.
Authors: Busayo Awobade, Mardiyyah Oduwole, Steven Kolawole
Abstract: Compression techniques have been crucial in advancing machine learning by enabling efficient training and deployment of large-scale language models. However, these techniques have received limited attention in the context of low-resource language models, which are trained on even smaller amounts of data and under computational constraints, a scenario known as the "low-resource double-bind." This paper investigates the effectiveness of pruning, knowledge distillation, and quantization on an exclusively low-resourced, small-data language model, AfriBERTa. Through a battery of experiments, we assess the effects of compression on performance across several metrics beyond accuracy. Our study provides evidence that compression techniques significantly improve the efficiency and effectiveness of small-data language models, confirming that the prevailing beliefs regarding the effects of compression on large, heavily parameterized models hold true for less-parameterized, small-data models.
Authors: Yuqi Song, Rongzhi Dong, Lai Wei, Qin Li, Jianjun Hu
Abstract: Computational prediction of stable crystal structures has a profound impact on the large-scale discovery of novel functional materials. However, predicting the crystal structure solely from a material's composition or formula is a promising yet challenging task, as traditional ab initio crystal structure prediction (CSP) methods rely on time-consuming global searches and first-principles free energy calculations. Inspired by the recent success of deep learning approaches in protein structure prediction, which utilize pairwise amino acid interactions to describe 3D structures, we present AlphaCrystal-II, a novel knowledge-based solution that exploits the abundant inter-atomic interaction patterns found in existing known crystal structures. AlphaCrystal-II predicts the atomic distance matrix of a target crystal material and employs this matrix to reconstruct its 3D crystal structure. By leveraging the wealth of inter-atomic relationships of known crystal structures, our approach demonstrates remarkable effectiveness and reliability in structure prediction through comprehensive experiments. This work highlights the potential of data-driven methods in accelerating the discovery and design of new materials with tailored properties.
Authors: Hongzheng Chen, Niansong Zhang, Shaojie Xiang, Zhichen Zeng, Mengjia Dai, Zhiru Zhang
Abstract: Special-purpose hardware accelerators are increasingly pivotal for sustaining performance improvements in emerging applications, especially as the benefits of technology scaling continue to diminish. However, designers currently lack effective tools and methodologies to construct complex, high-performance accelerator architectures in a productive manner. Existing high-level synthesis (HLS) tools often require intrusive source-level changes to attain satisfactory quality of results. Despite the introduction of several new accelerator design languages (ADLs) aiming to enhance or replace HLS, their advantages are more evident in relatively simple applications with a single kernel. Existing ADLs prove less effective for realistic hierarchical designs with multiple kernels, even if the design hierarchy is flattened. In this paper, we introduce Allo, a composable programming model for efficient spatial accelerator design. Allo decouples hardware customizations, including compute, memory, communication, and data type from algorithm specification, and encapsulates them as a set of customization primitives. Allo preserves the hierarchical structure of an input program by combining customizations from different functions in a bottom-up, type-safe manner. This approach facilitates holistic optimizations that span across function boundaries. We conduct comprehensive experiments on commonly-used HLS benchmarks and several realistic deep learning models. Our evaluation shows that Allo can outperform state-of-the-art HLS tools and ADLs on all test cases in the PolyBench. For the GPT2 model, the inference latency of the Allo generated accelerator is 1.7x faster than the NVIDIA A100 GPU with 5.4x higher energy efficiency, demonstrating the capability of Allo to handle large-scale designs.
Authors: Fatemeh Asghari, Mohammad Reza Soheili, Faezeh Gholamrezaie
Abstract: The global fashion industry plays a pivotal role in the global economy, and addressing fundamental issues within the industry is crucial for developing innovative solutions. One of the most pressing challenges in the fashion industry is the mismatch between body shapes and the garments of individuals they purchase. This issue is particularly prevalent among individuals with non-ideal body shapes, exacerbating the challenges faced. Considering inter-individual variability in body shapes is essential for designing and producing garments that are widely accepted by consumers. Traditional methods for determining human body shape are limited due to their low accuracy, high costs, and time-consuming nature. New approaches, utilizing digital imaging and deep neural networks (DNN), have been introduced to identify human body shape. In this study, the Style4BodyShape dataset is used for classifying body shapes into five categories: Rectangle, Triangle, Inverted Triangle, Hourglass, and Apple. In this paper, the body shape segmentation of a person is extracted from the image, disregarding the surroundings and background. Then, Various pre-trained models, such as ResNet18, ResNet34, ResNet50, VGG16, VGG19, and Inception v3, are used to classify the segmentation results. Among these pre-trained models, the Inception V3 model demonstrates superior performance regarding f1-score evaluation metric and accuracy compared to the other models.
Authors: Youran Zhou, Sunil Aryal, Mohamed Reda Bouadjenek
Abstract: Missing data poses a significant challenge in data science, affecting decision-making processes and outcomes. Understanding what missing data is, how it occurs, and why it is crucial to handle it appropriately is paramount when working with real-world data, especially in tabular data, one of the most commonly used data types in the real world. Three missing mechanisms are defined in the literature: Missing Completely At Random (MCAR), Missing At Random (MAR), and Missing Not At Random (MNAR), each presenting unique challenges in imputation. Most existing work are focused on MCAR that is relatively easy to handle. The special missing mechanisms of MNAR and MAR are less explored and understood. This article reviews existing literature on handling missing values. It compares and contrasts existing methods in terms of their ability to handle different missing mechanisms and data types. It identifies research gap in the existing literature and lays out potential directions for future research in the field. The information in this review will help data analysts and researchers to adopt and promote good practices for handling missing data in real-world problems.
Authors: Yiyang Ma, Wenhan Yang, Jiaying Liu
Abstract: The images produced by diffusion models can attain excellent perceptual quality. However, it is challenging for diffusion models to guarantee distortion, hence the integration of diffusion models and image compression models still needs more comprehensive explorations. This paper presents a diffusion-based image compression method that employs a privileged end-to-end decoder model as correction, which achieves better perceptual quality while guaranteeing the distortion to an extent. We build a diffusion model and design a novel paradigm that combines the diffusion model and an end-to-end decoder, and the latter is responsible for transmitting the privileged information extracted at the encoder side. Specifically, we theoretically analyze the reconstruction process of the diffusion models at the encoder side with the original images being visible. Based on the analysis, we introduce an end-to-end convolutional decoder to provide a better approximation of the score function $\nabla_{\mathbf{x}_t}\log p(\mathbf{x}_t)$ at the encoder side and effectively transmit the combination. Experiments demonstrate the superiority of our method in both distortion and perception compared with previous perceptual compression methods.
Authors: Yi Luo, Jianwei Yu, Hangting Chen, Rongzhi Gu, Chao Weng
Abstract: We introduce Gull, a generative multifunctional audio codec. Gull is a general purpose neural audio compression and decompression model which can be applied to a wide range of tasks and applications such as real-time communication, audio super-resolution, and codec language models. The key components of Gull include (1) universal-sample-rate modeling via subband modeling schemes motivated by recent progress in audio source separation, (2) gain-shape representations motivated by traditional audio codecs, (3) improved residual vector quantization modules for simpler training, (4) elastic decoder network that enables user-defined model size and complexity during inference time, (5) built-in ability for audio super-resolution without the increase of bitrate. We compare Gull with existing traditional and neural audio codecs and show that Gull is able to achieve on par or better performance across various sample rates, bitrates and model complexities in both subjective and objective evaluation metrics.
Authors: Anirban Mukherjee, Hannah Hanwen Chang
Abstract: Social science research often hinges on the relationship between categorical variables and outcomes. We introduce CAVIAR, a novel method for embedding categorical variables that assume values in a high-dimensional ambient space but are sampled from an underlying manifold. Our theoretical and numerical analyses outline challenges posed by such categorical variables in causal inference. Specifically, dynamically varying and sparse levels can lead to violations of the Donsker conditions and a failure of the estimation functionals to converge to a tight Gaussian process. Traditional approaches, including the exclusion of rare categorical levels and principled variable selection models like LASSO, fall short. CAVIAR embeds the data into a lower-dimensional global coordinate system. The mapping can be derived from both structured and unstructured data, and ensures stable and robust estimates through dimensionality reduction. In a dataset of direct-to-consumer apparel sales, we illustrate how high-dimensional categorical variables, such as zip codes, can be succinctly represented, facilitating inference and analysis.
Authors: Valentin Koch, Sophia J. Wagner, Salome Kazeminia, Ece Sancar, Matthias Hehr, Julia Schnabel, Tingying Peng, Carsten Marr
Abstract: In hematology, computational models offer significant potential to improve diagnostic accuracy, streamline workflows, and reduce the tedious work of analyzing single cells in peripheral blood or bone marrow smears. However, clinical adoption of computational models has been hampered by the lack of generalization due to large batch effects, small dataset sizes, and poor performance in transfer learning from natural images. To address these challenges, we introduce DinoBloom, the first foundation model for single cell images in hematology, utilizing a tailored DINOv2 pipeline. Our model is built upon an extensive collection of 13 diverse, publicly available datasets of peripheral blood and bone marrow smears, the most substantial open-source cohort in hematology so far, comprising over 380,000 white blood cell images. To assess its generalization capability, we evaluate it on an external dataset with a challenging domain shift. We show that our model outperforms existing medical and non-medical vision models in (i) linear probing and k-nearest neighbor evaluations for cell-type classification on blood and bone marrow smears and (ii) weakly supervised multiple instance learning for acute myeloid leukemia subtyping by a large margin. A family of four DinoBloom models (small, base, large, and giant) can be adapted for a wide range of downstream applications, be a strong baseline for classification problems, and facilitate the assessment of batch effects in new datasets. All models are available at github.com/marrlab/DinoBloom.
Authors: Abhranil Das
Abstract: We present several exact and approximate mathematical methods and open-source software to compute the cdf, pdf and inverse cdf of the generalized chi-square distribution, which appears in Bayesian classification problems. Some methods are geared for speed, while others are designed to be accurate far into the tails, using which we can also measure large values of the discriminability index $d'$ between multinormals. We compare the accuracy and speed of these methods against the best existing methods.
Authors: Yimu Wang, Shuai Yuan, Xiangru Jian, Wei Pang, Mushi Wang, Ning Yu
Abstract: While recent progress in video-text retrieval has been driven by the exploration of powerful model architectures and training strategies, the representation learning ability of video-text retrieval models is still limited due to low-quality and scarce training data annotations. To address this issue, we present a novel video-text learning paradigm, HaVTR, which augments video and text data to learn more generalized features. Specifically, we first adopt a simple augmentation method, which generates self-similar data by randomly duplicating or dropping subwords and frames. In addition, inspired by the recent advancement in visual and language generative models, we propose a more powerful augmentation method through textual paraphrasing and video stylization using large language models (LLMs) and visual generative models (VGMs). Further, to bring richer information into video and text, we propose a hallucination-based augmentation method, where we use LLMs and VGMs to generate and add new relevant information to the original data. Benefiting from the enriched data, extensive experiments on several video-text retrieval benchmarks demonstrate the superiority of HaVTR over existing methods.
Authors: Alexandre Muzio, Alex Sun, Churan He
Abstract: The advancement of deep learning has led to the emergence of Mixture-of-Experts (MoEs) models, known for their dynamic allocation of computational resources based on input. Despite their promise, MoEs face challenges, particularly in terms of memory requirements. To address this, our work introduces SEER-MoE, a novel two-stage framework for reducing both the memory footprint and compute requirements of pre-trained MoE models. The first stage involves pruning the total number of experts using a heavy-hitters counting guidance, while the second stage employs a regularization-based fine-tuning strategy to recover accuracy loss and reduce the number of activated experts during inference. Our empirical studies demonstrate the effectiveness of our method, resulting in a sparse MoEs model optimized for inference efficiency with minimal accuracy trade-offs.
Authors: Yousef Sadegheih, Afshin Bozorgpour, Pratibha Kumari, Reza Azad, Dorit Merhof
Abstract: As a result of the rise of Transformer architectures in medical image analysis, specifically in the domain of medical image segmentation, a multitude of hybrid models have been created that merge the advantages of Convolutional Neural Networks (CNNs) and Transformers. These hybrid models have achieved notable success by significantly improving segmentation accuracy. Yet, this progress often comes at the cost of increased model complexity, both in terms of parameters and computational demand. Moreover, many of these models fail to consider the crucial interplay between spatial and channel features, which could further refine and improve segmentation outcomes. To address this, we introduce LHU-Net, a Light Hybrid U-Net architecture optimized for volumetric medical image segmentation. LHU-Net is meticulously designed to prioritize spatial feature analysis in its initial layers before shifting focus to channel-based features in its deeper layers, ensuring a comprehensive feature extraction process. Rigorous evaluation across five benchmark datasets - Synapse, LA, Pancreas, ACDC, and BRaTS 2018 - underscores LHU-Net's superior performance, showcasing its dual capacity for efficiency and accuracy. Notably, LHU-Net sets new performance benchmarks, such as attaining a Dice score of 92.66 on the ACDC dataset, while simultaneously reducing parameters by 85% and quartering the computational load compared to existing state-of-the-art models. Achieved without any reliance on pre-training, additional data, or model ensemble, LHU-Net's effectiveness is further evidenced by its state-of-the-art performance across all evaluated datasets, utilizing fewer than 11 million parameters. This achievement highlights that balancing computational efficiency with high accuracy in medical image segmentation is feasible. Our implementation of LHU-Net is freely accessible to the research community on GitHub.
Authors: Mohsen Heidari, Masih Mozakka, Wojciech Szpankowski
Abstract: Hybrid quantum-classical optimization and learning strategies are among the most promising approaches to harnessing quantum information or gaining a quantum advantage over classical methods. However, efficient estimation of the gradient of the objective function in such models remains a challenge due to several factors including the exponential dimensionality of the Hilbert spaces, and information loss of quantum measurements. In this work, we study generic parameterized circuits in the context of variational methods. We develop a framework for gradient estimation that exploits the algebraic symmetries of Hamiltonian characterized through Lie algebra or group theory. Particularly, we prove that when the dimension of the dynamical Lie algebra is polynomial in the number of qubits, one can estimate the gradient with polynomial classical and quantum resources. This is done by a series of Hadamard tests applied to the output of the ansatz with no change to its circuit. We show that this approach can be equipped with classical shadow tomography to further reduce the measurement shot complexity to scale logarithmically with the number of parameters.
Authors: Nazifa Azam Khan, Mikolaj Cieslak, Ian McQuillan
Abstract: Artificial neural networks are often used to identify features of crop plants. However, training their models requires many annotated images, which can be expensive and time-consuming to acquire. Procedural models of plants, such as those developed with Lindenmayer-systems (L-systems) can be created to produce visually realistic simulations, and hence images of plant simulations, where annotations are implicitly known. These synthetic images can either augment or completely replace real images in training neural networks for phenotyping tasks. In this paper, we systematically vary amounts of real and synthetic images used for training in both maize and canola to better understand situations where synthetic images generated from L-systems can help prediction on real images. This work also explores the degree to which realism in the synthetic images improves prediction. Furthermore, we see how neural network predictions can be used to help calibrate L-systems themselves, creating a feedback loop.
Authors: Rohan Deepak Ajwani, Zining Zhu, Jonathan Rose, Frank Rudzicz
Abstract: Transformer-based Large Language Models (LLMs) have shown exceptional language generation capabilities in response to text-based prompts. However, controlling the direction of generation via textual prompts has been challenging, especially with smaller models. In this work, we explore the use of Prompt Tuning to achieve controlled language generation. Generated text is steered using prompt embeddings, which are trained using a small language model, used as a discriminator. Moreover, we demonstrate that these prompt embeddings can be trained with a very small dataset, with as low as a few hundred training examples. Our method thus offers a data and parameter efficient solution towards controlling language model outputs. We carry out extensive evaluation on four datasets: SST-5 and Yelp (sentiment analysis), GYAFC (formality) and JIGSAW (toxic language). Finally, we demonstrate the efficacy of our method towards mitigating harmful, toxic, and biased text generated by language models.
Authors: HyoJe Jung, Yunha Kim, Heejung Choi, Hyeram Seo, Minkyoung Kim, JiYe Han, Gaeun Kee, Seohyun Park, Soyoung Ko, Byeolhee Kim, Suyeon Kim, Tae Joon Jun, Young-Hak Kim
Abstract: Medical documentation, including discharge notes, is crucial for ensuring patient care quality, continuity, and effective medical communication. However, the manual creation of these documents is not only time-consuming but also prone to inconsistencies and potential errors. The automation of this documentation process using artificial intelligence (AI) represents a promising area of innovation in healthcare. This study directly addresses the inefficiencies and inaccuracies in creating discharge notes manually, particularly for cardiac patients, by employing AI techniques, specifically large language model (LLM). Utilizing a substantial dataset from a cardiology center, encompassing wide-ranging medical records and physician assessments, our research evaluates the capability of LLM to enhance the documentation process. Among the various models assessed, Mistral-7B distinguished itself by accurately generating discharge notes that significantly improve both documentation efficiency and the continuity of care for patients. These notes underwent rigorous qualitative evaluation by medical expert, receiving high marks for their clinical relevance, completeness, readability, and contribution to informed decision-making and care planning. Coupled with quantitative analyses, these results confirm Mistral-7B's efficacy in distilling complex medical information into concise, coherent summaries. Overall, our findings illuminate the considerable promise of specialized LLM, such as Mistral-7B, in refining healthcare documentation workflows and advancing patient care. This study lays the groundwork for further integrating advanced AI technologies in healthcare, demonstrating their potential to revolutionize patient documentation and support better care outcomes.
Authors: Roopkatha Dey, Aivy Debnath, Sayak Kumar Dutta, Kaustav Ghosh, Arijit Mitra, Arghya Roy Chowdhury, Jaydip Sen
Abstract: In various real-world applications such as machine translation, sentiment analysis, and question answering, a pivotal role is played by NLP models, facilitating efficient communication and decision-making processes in domains ranging from healthcare to finance. However, a significant challenge is posed to the robustness of these natural language processing models by text adversarial attacks. These attacks involve the deliberate manipulation of input text to mislead the predictions of the model while maintaining human interpretability. Despite the remarkable performance achieved by state-of-the-art models like BERT in various natural language processing tasks, they are found to remain vulnerable to adversarial perturbations in the input text. In addressing the vulnerability of text classifiers to adversarial attacks, three distinct attack mechanisms are explored in this paper using the victim model BERT: BERT-on-BERT attack, PWWS attack, and Fraud Bargain's Attack (FBA). Leveraging the IMDB, AG News, and SST2 datasets, a thorough comparative analysis is conducted to assess the effectiveness of these attacks on the BERT classifier model. It is revealed by the analysis that PWWS emerges as the most potent adversary, consistently outperforming other methods across multiple evaluation scenarios, thereby emphasizing its efficacy in generating adversarial examples for text classification. Through comprehensive experimentation, the performance of these attacks is assessed and the findings indicate that the PWWS attack outperforms others, demonstrating lower runtime, higher accuracy, and favorable semantic similarity scores. The key insight of this paper lies in the assessment of the relative performances of three prevalent state-of-the-art attack mechanisms.
Authors: Chih-Chung Hsu, Chia-Ming Lee, Chun-Hung Sun, Kuang-Ming Wu
Abstract: Traditional defect classification approaches are facing with two barriers. (1) Insufficient training data and unstable data quality. Collecting sufficient defective sample is expensive and time-costing, consequently leading to dataset variance. It introduces the difficulty on recognition and learning. (2) Over-dependence on visual modality. When the image pattern and texture is monotonic for all defect classes in a given dataset, the performance of conventional AOI system cannot be guaranteed. In scenarios where image quality is compromised due to mechanical failures or when defect information is inherently difficult to discern, the performance of deep models cannot be guaranteed. A main question is, "how to solve those two problems when they occur at the same time?" The feasible strategy is to explore another feature within dataset and combine an eminent vision-language model (VLM) and Large-Language model (LLM) with their astonishing zero-shot capability. In this work, we propose the special ASE dataset, including rich data description recorded on image, for defect classification, but the defect feature is uneasy to learn directly. Secondly, We present the prompting for VLM-LLM against defect classification with the proposed ASE dataset to activate extra-modality feature from images to enhance performance. Then, We design the novel progressive feature alignment (PFA) block to refine image-text feature to alleviate the difficulty of alignment under few-shot scenario. Finally, the proposed Cross-modality attention fusion (CMAF) module can effectively fuse different modality feature. Experiment results have demonstrated our method's effectiveness over several defect classification methods for the ASE dataset.
Authors: Juan C. Mej {\i}a-Fragoso, Manuel A. Florez, Roc{\i}o Bernal-Olaya
Abstract: Accurate determination of the geothermal gradient is critical for assessing the geothermal energy potential of a given region. Of particular interest is the case of Colombia, a country with abundant geothermal resources. A history of active oil and gas exploration and production has left drilled boreholes in different geological settings, providing direct measurements of the geothermal gradient. Unfortunately, large regions of the country where geothermal resources might exist lack such measurements. Indirect geophysical measurements are costly and difficult to perform at regional scales. Computational thermal models could be constructed, but they require very detailed knowledge of the underlying geology and uniform sampling of subsurface temperatures to be well-constrained. We present an alternative approach that leverages recent advances in supervised machine learning and available direct measurements to predict the geothermal gradient in regions where only global-scale geophysical datasets and course geological knowledge are available. We find that a Gradient Boosted Regression Tree algorithm yields optimal predictions and extensively validate the trained model. We show that predictions of our model are within 12\% accuracy and that independent measurements performed by other authors agree well with our model. Finnally, we present a geothermal gradient map for Colombia that highlights regions where futher exploration and data collection should be performed.
Authors: Huafu Liao, Alp\'ar R. M\'esz\'aros, Chenchen Mou, Chao Zhou
Abstract: This paper deals with a class of neural SDEs and studies the limiting behavior of the associated sampled optimal control problems as the sample size grows to infinity. The neural SDEs with N samples can be linked to the N-particle systems with centralized control. We analyze the Hamilton--Jacobi--Bellman equation corresponding to the N-particle system and establish regularity results which are uniform in N. The uniform regularity estimates are obtained by the stochastic maximum principle and the analysis of a backward stochastic Riccati equation. Using these uniform regularity results, we show the convergence of the minima of objective functionals and optimal parameters of the neural SDEs as the sample size N tends to infinity. The limiting objects can be identified with suitable functions defined on the Wasserstein space of Borel probability measures. Furthermore, quantitative algebraic convergence rates are also obtained.
Authors: George Leotescu, Daniel Voinea, Alin-Ionut Popa
Abstract: The transformer is a powerful data modelling framework responsible for remarkable performance on a wide range of tasks. However, they are limited in terms of scalability as it is suboptimal and inefficient to process long-sequence data. To this purpose we introduce BLRP (Bidirectional Long-Range Parser), a novel and versatile attention mechanism designed to increase performance and efficiency on long-sequence tasks. It leverages short and long range heuristics in the form of a local sliding window approach combined with a global bidirectional latent space synthesis technique. We show the benefits and versatility of our approach on vision and language domains by demonstrating competitive results against state-of-the-art methods on the Long-Range-Arena and CIFAR benchmarks together with ablations demonstrating the computational efficiency.
Authors: Yusuf Umut Ciftci, Zeyuan Feng, Somil Bansal
Abstract: Behavior Cloning is a popular approach to Imitation Learning, in which a robot observes an expert supervisor and learns a control policy. However, behavior cloning suffers from the "compounding error" problem - the policy errors compound as it deviates from the expert demonstrations and might lead to catastrophic system failures, limiting its use in safety-critical applications. On-policy data aggregation methods are able to address this issue at the cost of rolling out and repeated training of the imitation policy, which can be tedious and computationally prohibitive. We propose SAFE-GIL, an off-policy behavior cloning method that guides the expert via adversarial disturbance during data collection. The algorithm abstracts the imitation error as an adversarial disturbance in the system dynamics, injects it during data collection to expose the expert to safety critical states, and collects corrective actions. Our method biases training to more closely replicate expert behavior in safety-critical states and allows more variance in less critical states. We compare our method with several behavior cloning techniques and DAgger on autonomous navigation and autonomous taxiing tasks and show higher task success and safety, especially in low data regimes where the likelihood of error is higher, at a slight drop in the performance.
Authors: Seyedehdelaram Esfahani, Giovanni De Toni, Bruno Lepri, Andrea Passerini, Katya Tentori, Massimo Zancanaro
Abstract: Algorithmic Recourse aims to provide actionable explanations, or recourse plans, to overturn potentially unfavourable decisions taken by automated machine learning models. In this paper, we propose an interaction paradigm based on a guided interaction pattern aimed at both eliciting the users' preferences and heading them toward effective recourse interventions. In a fictional task of money lending, we compare this approach with an exploratory interaction pattern based on a combination of alternative plans and the possibility of freely changing the configurations by the users themselves. Our results suggest that users may recognize that the guided interaction paradigm improves efficiency. However, they also feel less freedom to experiment with "what-if" scenarios. Nevertheless, the time spent on the purely exploratory interface tends to be perceived as a lack of efficiency, which reduces attractiveness, perspicuity, and dependability. Conversely, for the guided interface, more time on the interface seems to increase its attractiveness, perspicuity, and dependability while not impacting the perceived efficiency. That might suggest that this type of interfaces should combine these two approaches by trying to support exploratory behavior while gently pushing toward a guided effective solution.
Authors: Jihanne El Haouari, Jean-Michel Gaucel, Christelle Pittet, Jean-Yves Tourneret, Herwig Wendt
Abstract: Accurate estimates of Instrument Spectral Response Functions (ISRFs) are crucial in order to have a good characterization of high resolution spectrometers. Spectrometers are composed of different optical elements that can induce errors in the measurements and therefore need to be modeled as accurately as possible. Parametric models are currently used to estimate these response functions. However, these models cannot always take into account the diversity of ISRF shapes that are encountered in practical applications. This paper studies a new ISRF estimation method based on a sparse representation of atoms belonging to a dictionary. This method is applied to different high-resolution spectrometers in order to assess its reproducibility for multiple remote sensing missions. The proposed method is shown to be very competitive when compared to the more commonly used parametric models, and yields normalized ISRF estimation errors less than 1%.
Authors: Omran Ayoub, Davide Andreoletti, Aleksandra Knapi\'nska, R\'o\.za Go\'scie\'n, Piotr Lechowicz, Tiziano Leidi, Silvia Giordano, Cristina Rottondi, Krzysztof Walkowiak
Abstract: Adapting to concept drift is a challenging task in machine learning, which is usually tackled using incremental learning techniques that periodically re-fit a learning model leveraging newly available data. A primary limitation of these techniques is their reliance on substantial amounts of data for retraining. The necessity of acquiring fresh data introduces temporal delays prior to retraining, potentially rendering the models inaccurate if a sudden concept drift occurs in-between two consecutive retrainings. In communication networks, such issue emerges when performing traffic forecasting following a~failure event: post-failure re-routing may induce a drastic shift in distribution and pattern of traffic data, thus requiring a timely model adaptation. In this work, we address this challenge for the problem of traffic forecasting and propose an approach that exploits adaptive learning algorithms, namely, liquid neural networks, which are capable of self-adaptation to abrupt changes in data patterns without requiring any retraining. Through extensive simulations of failure scenarios, we compare the predictive performance of our proposed approach to that of a reference method based on incremental learning. Experimental results show that our proposed approach outperforms incremental learning-based methods in situations where the shifts in traffic patterns are drastic.
Authors: Yu-Hsi Chen
Abstract: Automated data labeling techniques are crucial for accelerating the development of deep learning models, particularly in complex medical imaging applications. However, ensuring accuracy and efficiency remains challenging. This paper presents iterative refinement strategies for automated data labeling in facial landmark diagnosis to enhance accuracy and efficiency for deep learning models in medical applications, including dermatology, plastic surgery, and ophthalmology. Leveraging feedback mechanisms and advanced algorithms, our approach iteratively refines initial labels, reducing reliance on manual intervention while improving label quality. Through empirical evaluation and case studies, we demonstrate the effectiveness of our proposed strategies in deep learning tasks across medical imaging domains. Our results highlight the importance of iterative refinement in automated data labeling to enhance the capabilities of deep learning systems in medical imaging applications.
Authors: Quentin Renau, Emma Hart
Abstract: The choice of input-data used to train algorithm-selection models is recognised as being a critical part of the model success. Recently, feature-free methods for algorithm-selection that use short trajectories obtained from running a solver as input have shown promise. However, it is unclear to what extent these trajectories reliably discriminate between solvers. We propose a meta approach to generating discriminatory trajectories with respect to a portfolio of solvers. The algorithm-configuration tool irace is used to tune the parameters of a simple Simulated Annealing algorithm (SA) to produce trajectories that maximise the performance metrics of ML models trained on this data. We show that when the trajectories obtained from the tuned SA algorithm are used in ML models for algorithm-selection and performance prediction, we obtain significantly improved performance metrics compared to models trained both on raw trajectory data and on exploratory landscape features.
Authors: Jan Klhufek, Miroslav Safar, Vojtech Mrazek, Zdenek Vasicek, Lukas Sekanina
Abstract: Energy efficiency and memory footprint of a convolutional neural network (CNN) implemented on a CNN inference accelerator depend on many factors, including a weight quantization strategy (i.e., data types and bit-widths) and mapping (i.e., placement and scheduling of DNN elementary operations on hardware units of the accelerator). We show that enabling rich mixed quantization schemes during the implementation can open a previously hidden space of mappings that utilize the hardware resources more effectively. CNNs utilizing quantized weights and activations and suitable mappings can significantly improve trade-offs among the accuracy, energy, and memory requirements compared to less carefully optimized CNN implementations. To find, analyze, and exploit these mappings, we: (i) extend a general-purpose state-of-the-art mapping tool (Timeloop) to support mixed quantization, which is not currently available; (ii) propose an efficient multi-objective optimization algorithm to find the most suitable bit-widths and mapping for each DNN layer executed on the accelerator; and (iii) conduct a detailed experimental evaluation to validate the proposed method. On two CNNs (MobileNetV1 and MobileNetV2) and two accelerators (Eyeriss and Simba) we show that for a given quality metric (such as the accuracy on ImageNet), energy savings are up to 37% without any accuracy drop.
Authors: Zeyuan Allen-Zhu, Yuanzhi Li
Abstract: Scaling laws describe the relationship between the size of language models and their capabilities. Unlike prior studies that evaluate a model's capability via loss or benchmarks, we estimate the number of knowledge bits a model stores. We focus on factual knowledge represented as tuples, such as (USA, capital, Washington D.C.) from a Wikipedia page. Through multiple controlled datasets, we establish that language models can and only can store 2 bits of knowledge per parameter, even when quantized to int8, and such knowledge can be flexibly extracted for downstream applications. Consequently, a 7B model can store 14B bits of knowledge, surpassing the English Wikipedia and textbooks combined based on our estimation. More broadly, we present 12 results on how (1) training duration, (2) model architecture, (3) quantization, (4) sparsity constraints such as MoE, and (5) data signal-to-noise ratio affect a model's knowledge storage capacity. Notable insights include: * The GPT-2 architecture, with rotary embedding, matches or even surpasses LLaMA/Mistral architectures in knowledge storage, particularly over shorter training durations. This arises because LLaMA/Mistral uses GatedMLP, which is less stable and harder to train. * Prepending training data with domain names (e.g., wikipedia.org) significantly increases a model's knowledge capacity. Language models can autonomously identify and prioritize domains rich in knowledge, optimizing their storage capacity.
Authors: Tobias Meggendorfer, Maximilian Weininger, Patrick Wienh\"oft
Abstract: Markov decision processes (MDPs) are a fundamental model for decision making under uncertainty. They exhibit non-deterministic choice as well as probabilistic uncertainty. Traditionally, verification algorithms assume exact knowledge of the probabilities that govern the behaviour of an MDP. As this assumption is often unrealistic in practice, statistical model checking (SMC) was developed in the past two decades. It allows to analyse MDPs with unknown transition probabilities and provide probably approximately correct (PAC) guarantees on the result. Model-based SMC algorithms sample the MDP and build a model of it by estimating all transition probabilities, essentially for every transition answering the question: ``What are the odds?'' However, so far the statistical methods employed by the state of the art SMC algorithms are quite naive. Our contribution are several fundamental improvements to those methods: On the one hand, we survey statistics literature for better concentration inequalities; on the other hand, we propose specialised approaches that exploit our knowledge of the MDP. Our improvements are generally applicable to many kinds of problem statements because they are largely independent of the setting. Moreover, our experimental evaluation shows that they lead to significant gains, reducing the number of samples that the SMC algorithm has to collect by up to two orders of magnitude.
Authors: Michael Wagner, Carmen Carlan
Abstract: A system safety case is a compelling, comprehensible, and valid argument about the satisfaction of the safety goals of a given system operating in a given environment supported by convincing evidence. Since the publication of UL 4600 in 2020, safety cases have become a best practice for measuring, managing, and communicating the safety of autonomous vehicles (AVs). Although UL 4600 provides guidance on how to build the safety case for an AV, the complexity of AVs and their operating environments, the novelty of the used technology, the need for complying with various regulations and technical standards, and for addressing cybersecurity concerns and ethical considerations make the development of safety cases for AVs challenging. To this end, safety case frameworks have been proposed that bring strategies, argument templates, and other guidance together to support the development of a safety case. This paper introduces the Open Autonomy Safety Case Framework, developed over years of work with the autonomous vehicle industry, as a roadmap for how AVs can be deployed safely and responsibly.
Authors: Hong Ye Tan, Ziruo Cai, Marcelo Pereyra, Subhadip Mukherjee, Junqi Tang, Carola-Bibiane Sch\"onlieb
Abstract: Unsupervised learning is a training approach in the situation where ground truth data is unavailable, such as inverse imaging problems. We present an unsupervised Bayesian training approach to learning convex neural network regularizers using a fixed noisy dataset, based on a dual Markov chain estimation method. Compared to classical supervised adversarial regularization methods, where there is access to both clean images as well as unlimited to noisy copies, we demonstrate close performance on natural image Gaussian deconvolution and Poisson denoising tasks.
Authors: Lluis Castrejon, Thomas Mensink, Howard Zhou, Vittorio Ferrari, Andre Araujo, Jasper Uijlings
Abstract: Combining Large Language Models (LLMs) with external specialized tools (LLMs+tools) is a recent paradigm to solve multimodal tasks such as Visual Question Answering (VQA). While this approach was demonstrated to work well when optimized and evaluated for each individual benchmark, in practice it is crucial for the next generation of real-world AI systems to handle a broad range of multimodal problems. Therefore we pose the VQA problem from a unified perspective and evaluate a single system on a varied suite of VQA tasks including counting, spatial reasoning, OCR-based reasoning, visual pointing, external knowledge, and more. In this setting, we demonstrate that naively applying the LLM+tools approach using the combined set of all tools leads to poor results. This motivates us to introduce HAMMR: HierArchical MultiModal React. We start from a multimodal ReAct-based system and make it hierarchical by enabling our HAMMR agents to call upon other specialized agents. This enhances the compositionality of the LLM+tools approach, which we show to be critical for obtaining high accuracy on generic VQA. Concretely, on our generic VQA suite, HAMMR outperforms the naive LLM+tools approach by 19.5%. Additionally, HAMMR achieves state-of-the-art results on this task, outperforming the generic standalone PaLI-X VQA model by 5.0%.
Authors: Hugo Caselles-Dupr\'e, Charles Mellerio, Paul H\'erent, Aliz\'ee Lopez-Persem, Benoit B\'eranger, Mathieu Soularue, Pierre Fautrel, Gauthier Vernier, Matthieu Cord
Abstract: The reconstruction of images observed by subjects from fMRI data collected during visual stimuli has made significant strides in the past decade, thanks to the availability of extensive fMRI datasets and advancements in generative models for image generation. However, the application of visual reconstruction has remained limited. Reconstructing visual imagination presents a greater challenge, with potentially revolutionary applications ranging from aiding individuals with disabilities to verifying witness accounts in court. The primary hurdles in this field are the absence of data collection protocols for visual imagery and the lack of datasets on the subject. Traditionally, fMRI-to-image relies on data collected from subjects exposed to visual stimuli, which poses issues for generating visual imagery based on the difference of brain activity between visual stimulation and visual imagery. For the first time, we have compiled a substantial dataset (around 6h of scans) on visual imagery along with a proposed data collection protocol. We then train a modified version of an fMRI-to-image model and demonstrate the feasibility of reconstructing images from two modes of imagination: from memory and from pure imagination. This marks an important step towards creating a technology that allow direct reconstruction of visual imagery.
Authors: Vivek Agarwal, Joshua Harvey, Dmitry Rinberg, Vasant Dhar
Abstract: Advances in neural sensing technology are making it possible to observe the olfactory process in great detail. In this paper, we conceptualize smell from a Data Science and AI perspective, that relates the properties of odorants to how they are sensed and analyzed in the olfactory system from the nose to the brain. Drawing distinctions to color vision, we argue that smell presents unique measurement challenges, including the complexity of stimuli, the high dimensionality of the sensory apparatus, as well as what constitutes ground truth. In the face of these challenges, we argue for the centrality of odorant-receptor interactions in developing a theory of olfaction. Such a theory is likely to find widespread industrial applications, and enhance our understanding of smell, and in the longer-term, how it relates to other senses and language. As an initial use case of the data, we present results using machine learning-based classification of neural responses to odors as they are recorded in the mouse olfactory bulb with calcium imaging.
Authors: Hamed Haghighi, Amir Samadi, Mehrdad Dianati, Valentina Donzella, Kurt Debattista
Abstract: Diffusion Models (DMs) have achieved State-Of-The-Art (SOTA) results in the Lidar point cloud generation task, benefiting from their stable training and iterative refinement during sampling. However, DMs often fail to realistically model Lidar raydrop noise due to their inherent denoising process. To retain the strength of iterative sampling while enhancing the generation of raydrop noise, we introduce LidarGRIT, a generative model that uses auto-regressive transformers to iteratively sample the range images in the latent space rather than image space. Furthermore, LidarGRIT utilises VQ-VAE to separately decode range images and raydrop masks. Our results show that LidarGRIT achieves superior performance compared to SOTA models on KITTI-360 and KITTI odometry datasets. Code available at:https://github.com/hamedhaghighi/LidarGRIT.
Authors: Saman Motamed, Wouter Van Gansbeke, Luc Van Gool
Abstract: With recent advances in image and video diffusion models for content creation, a plethora of techniques have been proposed for customizing their generated content. In particular, manipulating the cross-attention layers of Text-to-Image (T2I) diffusion models has shown great promise in controlling the shape and location of objects in the scene. Transferring image-editing techniques to the video domain, however, is extremely challenging as object motion and temporal consistency are difficult to capture accurately. In this work, we take a first look at the role of cross-attention in Text-to-Video (T2V) diffusion models for zero-shot video editing. While one-shot models have shown potential in controlling motion and camera movement, we demonstrate zero-shot control over object shape, position and movement in T2V models. We show that despite the limitations of current T2V models, cross-attention guidance can be a promising approach for editing videos.
Authors: Tim Baumg\"artner, Yang Gao, Dana Alon, Donald Metzler
Abstract: Reinforcement Learning from Human Feedback (RLHF) is a popular method for aligning Language Models (LM) with human values and preferences. RLHF requires a large number of preference pairs as training data, which are often used in both the Supervised Fine-Tuning and Reward Model training, and therefore publicly available datasets are commonly used. In this work, we study to what extent a malicious actor can manipulate the LMs generations by poisoning the preferences, i.e., injecting poisonous preference pairs into these datasets and the RLHF training process. We propose strategies to build poisonous preference pairs and test their performance by poisoning two widely used preference datasets. Our results show that preference poisoning is highly effective: by injecting a small amount of poisonous data (1-5% of the original dataset), we can effectively manipulate the LM to generate a target entity in a target sentiment (positive or negative). The findings from our experiments also shed light on strategies to defend against the preference poisoning attack.
Authors: Matteo Zecchin, Kai Zu, Osvaldo Simeone
Abstract: Large pre-trained sequence models, such as transformers, excel as few-shot learners capable of in-context learning (ICL). In ICL, a model is trained to adapt its operation to a new task based on limited contextual information, typically in the form of a few training examples for the given task. Previous work has explored the use of ICL for channel equalization in single-user multi-input and multiple-output (MIMO) systems. In this work, we demonstrate that ICL can be also used to tackle the problem of multi-user equalization in cell-free MIMO systems with limited fronthaul capacity. In this scenario, a task is defined by channel statistics, signal-to-noise ratio, and modulation schemes. The context encompasses the users' pilot sequences, the corresponding quantized received signals, and the current received data signal. Different prompt design strategies are proposed and evaluated that encompass also large-scale fading and modulation information. Experiments demonstrate that ICL-based equalization provides estimates with lower mean squared error as compared to the linear minimum mean squared error equalizer, especially in the presence of limited fronthaul capacity and pilot contamination.
Authors: Mehran Safayani, Amir Sartipi, Amir Hossein Ahmadi, Parniyan Jalali, Amir Hossein Mansouri, Mohammad Bisheh-Niasar, Zahra Pourbahman
Abstract: The proliferation of hate speech and offensive comments on social media has become increasingly prevalent due to user activities. Such comments can have detrimental effects on individuals' psychological well-being and social behavior. While numerous datasets in the English language exist in this domain, few equivalent resources are available for Persian language. To address this gap, this paper introduces two offensive datasets. The first dataset comprises annotations provided by domain experts, while the second consists of a large collection of unlabeled data obtained through web crawling for unsupervised learning purposes. To ensure the quality of the former dataset, a meticulous three-stage labeling process was conducted, and kappa measures were computed to assess inter-annotator agreement. Furthermore, experiments were performed on the dataset using state-of-the-art language models, both with and without employing masked language modeling techniques, as well as machine learning algorithms, in order to establish the baselines for the dataset using contemporary cutting-edge approaches. The obtained F1-scores for the three-class and two-class versions of the dataset were 76.9% and 89.9% for XLM-RoBERTa, respectively.
Authors: Olivier Binette, Youngsoo Baek, Siddharth Engineer, Christina Jones, Abel Dasylva, Jerome P. Reiter
Abstract: Entity resolution (record linkage, microclustering) systems are notoriously difficult to evaluate. Looking for a needle in a haystack, traditional evaluation methods use sophisticated, application-specific sampling schemes to find matching pairs of records among an immense number of non-matches. We propose an alternative that facilitates the creation of representative, reusable benchmark data sets without necessitating complex sampling schemes. These benchmark data sets can then be used for model training and a variety of evaluation tasks. Specifically, we propose an entity-centric data labeling methodology that integrates with a unified framework for monitoring summary statistics, estimating key performance metrics such as cluster and pairwise precision and recall, and analyzing root causes for errors. We validate the framework in an application to inventor name disambiguation and through simulation studies. Software: https://github.com/OlivierBinette/er-evaluation/
Authors: Sohag Rahman, Sai Zhang, Min Xian, Shoukun Sun, Fei Xu, Zhegang Ma
Abstract: Industry-wide nuclear power plant operating experience is a critical source of raw data for performing parameter estimations in reliability and risk models. Much operating experience information pertains to failure events and is stored as reports containing unstructured data, such as narratives. Event reports are essential for understanding how failures are initiated and propagated, including the numerous causal relations involved. Causal relation extraction using deep learning represents a significant frontier in the field of natural language processing (NLP), and is crucial since it enables the interpretation of intricate narratives and connections contained within vast amounts of written information. This paper proposed a hybrid framework for causality detection and extraction from nuclear licensee event reports. The main contributions include: (1) we compiled an LER corpus with 20,129 text samples for causality analysis, (2) developed an interactive tool for labeling cause effect pairs, (3) built a deep-learning-based approach for causal relation detection, and (4) developed a knowledge based cause-effect extraction approach.
Authors: Yuheng Lai, Leying Guan
Abstract: Equalized odds, as a popular notion of algorithmic fairness, aims to ensure that sensitive variables, such as race and gender, do not unfairly influence the algorithm prediction when conditioning on the true outcome. Despite rapid advancements, most of the current research focuses on the violation of equalized odds caused by one sensitive attribute, leaving the challenge of simultaneously accounting for multiple attributes under-addressed. We address this gap by introducing a fairness learning approach that integrates adversarial learning with a novel inverse conditional permutation. This approach effectively and flexibly handles multiple sensitive attributes, potentially of mixed data types. The efficacy and flexibility of our method are demonstrated through both simulation studies and empirical analysis of real-world datasets.
Authors: Ionut M. Motoi, Leonardo Saraceni, Daniele Nardi, Thomas A. Ciarfuglia
Abstract: Satellite imagery is crucial for tasks like environmental monitoring and urban planning. Typically, it relies on semantic segmentation or Land Use Land Cover (LULC) classification to categorize each pixel. Despite the advancements brought about by Deep Neural Networks (DNNs), their performance in segmentation tasks is hindered by challenges such as limited availability of labeled data, class imbalance and the inherent variability and complexity of satellite images. In order to mitigate those issues, our study explores the effectiveness of a Cut-and-Paste augmentation technique for semantic segmentation in satellite images. We adapt this augmentation, which usually requires labeled instances, to the case of semantic segmentation. By leveraging the connected components in the semantic segmentation labels, we extract instances that are then randomly pasted during training. Using the DynamicEarthNet dataset and a U-Net model for evaluation, we found that this augmentation significantly enhances the mIoU score on the test set from 37.9 to 44.1. This finding highlights the potential of the Cut-and-Paste augmentation to improve the generalization capabilities of semantic segmentation models in satellite imagery.
Authors: Ahmad Idrissi-Yaghir, Amin Dada, Henning Sch\"afer, Kamyar Arzideh, Giulia Baldini, Jan Trienes, Max Hasin, Jeanette Bewersdorff, Cynthia S. Schmidt, Marie Bauer, Kaleb E. Smith, Jiang Bian, Yonghui Wu, J\"org Schl\"otterer, Torsten Zesch, Peter A. Horn, Christin Seifert, Felix Nensa, Jens Kleesiek, Christoph M. Friedrich
Abstract: Recent advances in natural language processing (NLP) can be largely attributed to the advent of pre-trained language models such as BERT and RoBERTa. While these models demonstrate remarkable performance on general datasets, they can struggle in specialized domains such as medicine, where unique domain-specific terminologies, domain-specific abbreviations, and varying document structures are common. This paper explores strategies for adapting these models to domain-specific requirements, primarily through continuous pre-training on domain-specific data. We pre-trained several German medical language models on 2.4B tokens derived from translated public English medical data and 3B tokens of German clinical data. The resulting models were evaluated on various German downstream tasks, including named entity recognition (NER), multi-label classification, and extractive question answering. Our results suggest that models augmented by clinical and translation-based pre-training typically outperform general domain models in medical contexts. We conclude that continuous pre-training has demonstrated the ability to match or even exceed the performance of clinical models trained from scratch. Furthermore, pre-training on clinical data or leveraging translated texts have proven to be reliable methods for domain adaptation in medical NLP tasks.
Authors: Xinyang Gu, Yen-Jen Wang, Jianyu Chen
Abstract: Humanoid-Gym is an easy-to-use reinforcement learning (RL) framework based on Nvidia Isaac Gym, designed to train locomotion skills for humanoid robots, emphasizing zero-shot transfer from simulation to the real-world environment. Humanoid-Gym also integrates a sim-to-sim framework from Isaac Gym to Mujoco that allows users to verify the trained policies in different physical simulations to ensure the robustness and generalization of the policies. This framework is verified by RobotEra's XBot-S (1.2-meter tall humanoid robot) and XBot-L (1.65-meter tall humanoid robot) in a real-world environment with zero-shot sim-to-real transfer. The project website and source code can be found at: https://sites.google.com/view/humanoid-gym/.
Authors: Peng Zhao, Yu-Jie Zhang, Lijun Zhang, Zhi-Hua Zhou
Abstract: We investigate online convex optimization in non-stationary environments and choose dynamic regret as the performance measure, defined as the difference between cumulative loss incurred by the online algorithm and that of any feasible comparator sequence. Let $T$ be the time horizon and $P_T$ be the path length that essentially reflects the non-stationarity of environments, the state-of-the-art dynamic regret is $\mathcal{O}(\sqrt{T(1+P_T)})$. Although this bound is proved to be minimax optimal for convex functions, in this paper, we demonstrate that it is possible to further enhance the guarantee for some easy problem instances, particularly when online functions are smooth. Specifically, we introduce novel online algorithms that can exploit smoothness and replace the dependence on $T$ in dynamic regret with problem-dependent quantities: the variation in gradients of loss functions, the cumulative loss of the comparator sequence, and the minimum of these two terms. These quantities are at most $\mathcal{O}(T)$ while could be much smaller in benign environments. Therefore, our results are adaptive to the intrinsic difficulty of the problem, since the bounds are tighter than existing results for easy problems and meanwhile safeguard the same rate in the worst case. Notably, our proposed algorithms can achieve favorable dynamic regret with only one gradient per iteration, sharing the same gradient query complexity as the static regret minimization methods. To accomplish this, we introduce the collaborative online ensemble framework. The proposed framework employs a two-layer online ensemble to handle non-stationarity, and uses optimistic online learning and further introduces crucial correction terms to enable effective collaboration within the meta-base two layers, thereby attaining adaptivity. We believe the framework can be useful for broader problems.
Authors: Sivan Sabato, Eran Treister, Elad Yom-Tov
Abstract: We propose a new interpretable measure of unfairness, that allows providing a quantitative analysis of classifier fairness, beyond a dichotomous fair/unfair distinction. We show how this measure can be calculated when the classifier's conditional confusion matrices are known. We further propose methods for auditing classifiers for their fairness when the confusion matrices cannot be obtained or even estimated. Our approach lower-bounds the unfairness of a classifier based only on aggregate statistics, which may be provided by the owner of the classifier or collected from freely available data. We use the equalized odds criterion, which we generalize to the multiclass case. We report experiments on data sets representing diverse applications, which demonstrate the effectiveness and the wide range of possible uses of the proposed methodology. An implementation of the procedures proposed in this paper and as the code for running the experiments are provided in https://github.com/sivansabato/unfairness.
Authors: Chulin Xie, Pin-Yu Chen, Qinbin Li, Arash Nourian, Ce Zhang, Bo Li
Abstract: Federated learning (FL) enables distributed resource-constrained devices to jointly train shared models while keeping the training data local for privacy purposes. Vertical FL (VFL), which allows each client to collect partial features, has attracted intensive research efforts recently. We identified the main challenges that existing VFL frameworks are facing: the server needs to communicate gradients with the clients for each training step, incurring high communication cost that leads to rapid consumption of privacy budgets. To address these challenges, in this paper, we introduce a VFL framework with multiple heads (VIM), which takes the separate contribution of each client into account, and enables an efficient decomposition of the VFL optimization objective to sub-objectives that can be iteratively tackled by the server and the clients on their own. In particular, we propose an Alternating Direction Method of Multipliers (ADMM)-based method to solve our optimization problem, which allows clients to conduct multiple local updates before communication, and thus reduces the communication cost and leads to better performance under differential privacy (DP). We provide the user-level DP mechanism for our framework to protect user privacy. Moreover, we show that a byproduct of VIM is that the weights of learned heads reflect the importance of local clients. We conduct extensive evaluations and show that on four vertical FL datasets, VIM achieves significantly higher performance and faster convergence compared with the state-of-the-art. We also explicitly evaluate the importance of local clients and show that VIM enables functionalities such as client-level explanation and client denoising. We hope this work will shed light on a new way of effective VFL training and understanding.
Authors: Yujun Shi, Jian Liang, Wenqing Zhang, Vincent Y. F. Tan, Song Bai
Abstract: Federated learning aims to train models collaboratively across different clients without the sharing of data for privacy considerations. However, one major challenge for this learning paradigm is the {\em data heterogeneity} problem, which refers to the discrepancies between the local data distributions among various clients. To tackle this problem, we first study how data heterogeneity affects the representations of the globally aggregated models. Interestingly, we find that heterogeneous data results in the global model suffering from severe {\em dimensional collapse}, in which representations tend to reside in a lower-dimensional space instead of the ambient space. Moreover, we observe a similar phenomenon on models locally trained on each client and deduce that the dimensional collapse on the global model is inherited from local models. In addition, we theoretically analyze the gradient flow dynamics to shed light on how data heterogeneity result in dimensional collapse for local models. To remedy this problem caused by the data heterogeneity, we propose {\sc FedDecorr}, a novel method that can effectively mitigate dimensional collapse in federated learning. Specifically, {\sc FedDecorr} applies a regularization term during local training that encourages different dimensions of representations to be uncorrelated. {\sc FedDecorr}, which is implementation-friendly and computationally-efficient, yields consistent improvements over baselines on standard benchmark datasets. Code: https://github.com/bytedance/FedDecorr.
Authors: Anshuman Suri, Yifu Lu, Yanjin Chen, David Evans
Abstract: A distribution inference attack aims to infer statistical properties of data used to train machine learning models. These attacks are sometimes surprisingly potent, but the factors that impact distribution inference risk are not well understood and demonstrated attacks often rely on strong and unrealistic assumptions such as full knowledge of training environments even in supposedly black-box threat scenarios. To improve understanding of distribution inference risks, we develop a new black-box attack that even outperforms the best known white-box attack in most settings. Using this new attack, we evaluate distribution inference risk while relaxing a variety of assumptions about the adversary's knowledge under black-box access, like known model architectures and label-only access. Finally, we evaluate the effectiveness of previously proposed defenses and introduce new defenses. We find that although noise-based defenses appear to be ineffective, a simple re-sampling defense can be highly effective. Code is available at https://github.com/iamgroot42/dissecting_distribution_inference
URLs: https://github.com/iamgroot42/dissecting_distribution_inference
Authors: Bojian Hou, Hongming Li, Zhicheng Jiao, Zhen Zhou, Hao Zheng, Yong Fan
Abstract: Conventional survival analysis methods are typically ineffective to characterize heterogeneity in the population while such information can be used to assist predictive modeling. In this study, we propose a hybrid survival analysis method, referred to as deep clustering survival machines, that combines the discriminative and generative mechanisms. Similar to the mixture models, we assume that the timing information of survival data is generatively described by a mixture of certain numbers of parametric distributions, i.e., expert distributions. We learn weights of the expert distributions for individual instances according to their features discriminatively such that each instance's survival information can be characterized by a weighted combination of the learned constant expert distributions. This method also facilitates interpretable subgrouping/clustering of all instances according to their associated expert distributions. Extensive experiments on both real and synthetic datasets have demonstrated that the method is capable of obtaining promising clustering results and competitive time-to-event predicting performance.
Authors: Chulin Xie, De-An Huang, Wenda Chu, Daguang Xu, Chaowei Xiao, Bo Li, Anima Anandkumar
Abstract: Personalized Federated Learning (pFL) has emerged as a promising solution to tackle data heterogeneity across clients in FL. However, existing pFL methods either (1) introduce high communication and computation costs or (2) overfit to local data, which can be limited in scope, and are vulnerable to evolved test samples with natural shifts. In this paper, we propose PerAda, a parameter-efficient pFL framework that reduces communication and computational costs and exhibits superior generalization performance, especially under test-time distribution shifts. PerAda reduces the costs by leveraging the power of pretrained models and only updates and communicates a small number of additional parameters from adapters. PerAda has good generalization since it regularizes each client's personalized adapter with a global adapter, while the global adapter uses knowledge distillation to aggregate generalized information from all clients. Theoretically, we provide generalization bounds to explain why PerAda improves generalization, and we prove its convergence to stationary points under non-convex settings. Empirically, PerAda demonstrates competitive personalized performance (+4.85% on CheXpert) and enables better out-of-distribution generalization (+5.23% on CIFAR-10-C) on different datasets across natural and medical domains compared with baselines, while only updating 12.6% of parameters per model based on the adapter. Our code is available at https://github.com/NVlabs/PerAda.
Authors: Yue Kang, Cho-Jui Hsieh, Thomas C. M. Lee
Abstract: In stochastic contextual bandits, an agent sequentially makes actions from a time-dependent action set based on past experience to minimize the cumulative regret. Like many other machine learning algorithms, the performance of bandits heavily depends on the values of hyperparameters, and theoretically derived parameter values may lead to unsatisfactory results in practice. Moreover, it is infeasible to use offline tuning methods like cross-validation to choose hyperparameters under the bandit environment, as the decisions should be made in real-time. To address this challenge, we propose the first online continuous hyperparameter tuning framework for contextual bandits to learn the optimal parameter configuration in practice within a search space on the fly. Specifically, we use a double-layer bandit framework named CDT (Continuous Dynamic Tuning) and formulate the hyperparameter optimization as a non-stationary continuum-armed bandit, where each arm represents a combination of hyperparameters, and the corresponding reward is the algorithmic result. For the top layer, we propose the Zooming TS algorithm that utilizes Thompson Sampling (TS) for exploration and a restart technique to get around the \textit{switching} environment. The proposed CDT framework can be easily utilized to tune contextual bandit algorithms without any pre-specified candidate set for multiple hyperparameters. We further show that it could achieve a sublinear regret in theory and performs consistently better than all existing methods on both synthetic and real datasets.
Authors: Zayd Hammoudeh, Daniel Lowd
Abstract: Sparse or $\ell_0$ adversarial attacks arbitrarily perturb an unknown subset of the features. $\ell_0$ robustness analysis is particularly well-suited for heterogeneous (tabular) data where features have different types or scales. State-of-the-art $\ell_0$ certified defenses are based on randomized smoothing and apply to evasion attacks only. This paper proposes feature partition aggregation (FPA) -- a certified defense against the union of $\ell_0$ evasion, backdoor, and poisoning attacks. FPA generates its stronger robustness guarantees via an ensemble whose submodels are trained on disjoint feature sets. Compared to state-of-the-art $\ell_0$ defenses, FPA is up to 3,000${\times}$ faster and provides larger median robustness guarantees (e.g., median certificates of 13 pixels over 10 for CIFAR10, 12 pixels over 10 for MNIST, 4 features over 1 for Weather, and 3 features over 1 for Ames), meaning FPA provides the additional dimensions of robustness essentially for free.
Authors: Prasun C. Tripathi, Mohammod N. I. Suvon, Lawrence Schobs, Shuo Zhou, Samer Alabed, Andrew J. Swift, Haiping Lu
Abstract: Heart failure is a serious and life-threatening condition that can lead to elevated pressure in the left ventricle. Pulmonary Arterial Wedge Pressure (PAWP) is an important surrogate marker indicating high pressure in the left ventricle. PAWP is determined by Right Heart Catheterization (RHC) but it is an invasive procedure. A non-invasive method is useful in quickly identifying high-risk patients from a large population. In this work, we develop a tensor learning-based pipeline for identifying PAWP from multimodal cardiac Magnetic Resonance Imaging (MRI). This pipeline extracts spatial and temporal features from high-dimensional scans. For quality control, we incorporate an epistemic uncertainty-based binning strategy to identify poor-quality training samples. To improve the performance, we learn complementary information by integrating features from multimodal data: cardiac MRI with short-axis and four-chamber views, and Electronic Health Records. The experimental analysis on a large cohort of $1346$ subjects who underwent the RHC procedure for PAWP estimation indicates that the proposed pipeline has a diagnostic value and can produce promising performance with significant improvement over the baseline in clinical practice (i.e., $\Delta$AUC $=0.10$, $\Delta$Accuracy $=0.06$, and $\Delta$MCC $=0.39$). The decision curve analysis further confirms the clinical utility of our method.
Authors: Kaiwen Zheng, Cheng Lu, Jianfei Chen, Jun Zhu
Abstract: Diffusion models have exhibited excellent performance in various domains. The probability flow ordinary differential equation (ODE) of diffusion models (i.e., diffusion ODEs) is a particular case of continuous normalizing flows (CNFs), which enables deterministic inference and exact likelihood evaluation. However, the likelihood estimation results by diffusion ODEs are still far from those of the state-of-the-art likelihood-based generative models. In this work, we propose several improved techniques for maximum likelihood estimation for diffusion ODEs, including both training and evaluation perspectives. For training, we propose velocity parameterization and explore variance reduction techniques for faster convergence. We also derive an error-bounded high-order flow matching objective for finetuning, which improves the ODE likelihood and smooths its trajectory. For evaluation, we propose a novel training-free truncated-normal dequantization to fill the training-evaluation gap commonly existing in diffusion ODEs. Building upon these techniques, we achieve state-of-the-art likelihood estimation results on image datasets (2.56 on CIFAR-10, 3.43/3.69 on ImageNet-32) without variational dequantization or data augmentation, and 2.42 on CIFAR-10 with data augmentation. Code is available at \url{https://github.com/thu-ml/i-DODE}.
Authors: Hang Lou, Siran Li, Hao Ni
Abstract: Generating high-fidelity time series data using generative adversarial networks (GANs) remains a challenging task, as it is difficult to capture the temporal dependence of joint probability distributions induced by time-series data. Towards this goal, a key step is the development of an effective discriminator to distinguish between time series distributions. We propose the so-called PCF-GAN, a novel GAN that incorporates the path characteristic function (PCF) as the principled representation of time series distribution into the discriminator to enhance its generative performance. On the one hand, we establish theoretical foundations of the PCF distance by proving its characteristicity, boundedness, differentiability with respect to generator parameters, and weak continuity, which ensure the stability and feasibility of training the PCF-GAN. On the other hand, we design efficient initialisation and optimisation schemes for PCFs to strengthen the discriminative power and accelerate training efficiency. To further boost the capabilities of complex time series generation, we integrate the auto-encoder structure via sequential embedding into the PCF-GAN, which provides additional reconstruction functionality. Extensive numerical experiments on various datasets demonstrate the consistently superior performance of PCF-GAN over state-of-the-art baselines, in both generation and reconstruction quality. Code is available at https://github.com/DeepIntoStreams/PCF-GAN.
Authors: Guanghui Wang, Zihao Hu, Claudio Gentile, Vidya Muthukumar, Jacob Abernethy
Abstract: First-order optimization methods tend to inherently favor certain solutions over others when minimizing an underdetermined training objective that has multiple global optima. This phenomenon, known as implicit bias, plays a critical role in understanding the generalization capabilities of optimization algorithms. Recent research has revealed that in separable binary classification tasks gradient-descent-based methods exhibit an implicit bias for the $\ell_2$-maximal margin classifier. Similarly, generic optimization methods, such as mirror descent and steepest descent, have been shown to converge to maximal margin classifiers defined by alternative geometries. While gradient-descent-based algorithms provably achieve fast implicit bias rates, corresponding rates in the literature for generic optimization methods are relatively slow. To address this limitation, we present a series of state-of-the-art implicit bias rates for mirror descent and steepest descent algorithms. Our primary technique involves transforming a generic optimization algorithm into an online optimization dynamic that solves a regularized bilinear game, providing a unified framework for analyzing the implicit bias of various optimization methods. Our accelerated rates are derived by leveraging the regret bounds of online learning algorithms within this game framework. We then show the flexibility of this framework by analyzing the implicit bias in adversarial training, and again obtain significantly improved convergence rates.
Authors: Lei Guan, Dongsheng Li, Yanqi Shi, Jian Meng
Abstract: In this paper, we propose a general deep learning training framework XGrad which introduces weight prediction into the popular gradient-based optimizers to boost their convergence and generalization when training the deep neural network (DNN) models. In particular, ahead of each mini-batch training, the future weights are predicted according to the update rule of the used optimizer and are then applied to both the forward pass and backward propagation. In this way, during the whole training period, the optimizer always utilizes the gradients w.r.t. the future weights to update the DNN parameters, making the gradient-based optimizer achieve better convergence and generalization compared to the original optimizer without weight prediction. XGrad is rather straightforward to implement yet pretty effective in boosting the convergence of gradient-based optimizers and the accuracy of DNN models. Empirical results concerning five popular optimizers including SGD with momentum, Adam, AdamW, AdaBelief, and AdaM3 demonstrate the effectiveness of our proposal. The experimental results validate that XGrad can attain higher model accuracy than the baseline optimizers when training the DNN models. The code of XGrad will be available at: https://github.com/guanleics/XGrad.
Authors: Enrique Tom\'as Mart\'inez Beltr\'an, \'Angel Luis Perales G\'omez, Chao Feng, Pedro Miguel S\'anchez S\'anchez, Sergio L\'opez Bernal, G\'er\^ome Bovet, Manuel Gil P\'erez, Gregorio Mart\'inez P\'erez, Alberto Huertas Celdr\'an
Abstract: In 2016, Google proposed Federated Learning (FL) as a novel paradigm to train Machine Learning (ML) models across the participants of a federation while preserving data privacy. Since its birth, Centralized FL (CFL) has been the most used approach, where a central entity aggregates participants' models to create a global one. However, CFL presents limitations such as communication bottlenecks, single point of failure, and reliance on a central server. Decentralized Federated Learning (DFL) addresses these issues by enabling decentralized model aggregation and minimizing dependency on a central entity. Despite these advances, current platforms training DFL models struggle with key issues such as managing heterogeneous federation network topologies. To overcome these challenges, this paper presents Fedstellar, a platform extended from p2pfl library and designed to train FL models in a decentralized, semi-decentralized, and centralized fashion across diverse federations of physical or virtualized devices. The Fedstellar implementation encompasses a web application with an interactive graphical interface, a controller for deploying federations of nodes using physical or virtual devices, and a core deployed on each device which provides the logic needed to train, aggregate, and communicate in the network. The effectiveness of the platform has been demonstrated in two scenarios: a physical deployment involving single-board devices such as Raspberry Pis for detecting cyberattacks, and a virtualized deployment comparing various FL approaches in a controlled environment using MNIST and CIFAR-10 datasets. In both scenarios, Fedstellar demonstrated consistent performance and adaptability, achieving F1 scores of 91%, 98%, and 91.2% using DFL for detecting cyberattacks and classifying MNIST and CIFAR-10, respectively, reducing training time by 32% compared to centralized approaches.
Authors: Kexin Zhang, Qingsong Wen, Chaoli Zhang, Rongyao Cai, Ming Jin, Yong Liu, James Zhang, Yuxuan Liang, Guansong Pang, Dongjin Song, Shirui Pan
Abstract: Self-supervised learning (SSL) has recently achieved impressive performance on various time series tasks. The most prominent advantage of SSL is that it reduces the dependence on labeled data. Based on the pre-training and fine-tuning strategy, even a small amount of labeled data can achieve high performance. Compared with many published self-supervised surveys on computer vision and natural language processing, a comprehensive survey for time series SSL is still missing. To fill this gap, we review current state-of-the-art SSL methods for time series data in this article. To this end, we first comprehensively review existing surveys related to SSL and time series, and then provide a new taxonomy of existing time series SSL methods by summarizing them from three perspectives: generative-based, contrastive-based, and adversarial-based. These methods are further divided into ten subcategories with detailed reviews and discussions about their key intuitions, main frameworks, advantages and disadvantages. To facilitate the experiments and validation of time series SSL methods, we also summarize datasets commonly used in time series forecasting, classification, anomaly detection, and clustering tasks. Finally, we present the future directions of SSL for time series analysis.
Authors: Yuchang Sun, Yuyi Mao, Jun Zhang
Abstract: Federated learning (FL) is a promising framework for privacy-preserving collaborative learning, where model training tasks are distributed to clients and only the model updates need to be collected at a server. However, when being deployed at mobile edge networks, clients may have unpredictable availability and drop out of the training process, which hinders the convergence of FL. This paper tackles such a critical challenge. Specifically, we first investigate the convergence of the classical FedAvg algorithm with arbitrary client dropouts. We find that with the common choice of a decaying learning rate, FedAvg oscillates around a stationary point of the global loss function, which is caused by the divergence between the aggregated and desired central update. Motivated by this new observation, we then design a novel training algorithm named MimiC, where the server modifies each received model update based on the previous ones. The proposed modification of the received model updates mimics the imaginary central update irrespective of dropout clients. The theoretical analysis of MimiC shows that divergence between the aggregated and central update diminishes with proper learning rates, leading to its convergence. Simulation results further demonstrate that MimiC maintains stable convergence performance and learns better models than the baseline methods.
Authors: Bo Li, Yasin Esfandiari, Mikkel N. Schmidt, Tommy S. Alstr{\o}m, Sebastian U. Stich
Abstract: In federated learning, data heterogeneity is a critical challenge. A straightforward solution is to shuffle the clients' data to homogenize the distribution. However, this may violate data access rights, and how and when shuffling can accelerate the convergence of a federated optimization algorithm is not theoretically well understood. In this paper, we establish a precise and quantifiable correspondence between data heterogeneity and parameters in the convergence rate when a fraction of data is shuffled across clients. We prove that shuffling can quadratically reduce the gradient dissimilarity with respect to the shuffling percentage, accelerating convergence. Inspired by the theory, we propose a practical approach that addresses the data access rights issue by shuffling locally generated synthetic data. The experimental results show that shuffling synthetic data improves the performance of multiple existing federated learning algorithms by a large margin.
Authors: Mahyar Jahaninasab, Mohamad Ali Bijarchi
Abstract: This study introduces an accelerated training method for Vanilla Physics-Informed-Neural-Networks (PINN) addressing three factors that imbalance the loss function: initial weight state of a neural network, domain to boundary points ratio, and loss weighting factor. We propose a novel two-stage training method. During the initial stage, we create a unique loss function using a subset of boundary conditions and partial differential equation terms. Furthermore, we introduce preprocessing procedures that aim to decrease the variance during initialization and choose domain points according to the initial weight state of various neural networks. The second phase resembles Vanilla-PINN training, but a portion of the random weights are substituted with weights from the first phase. This implies that the neural network's structure is designed to prioritize the boundary conditions, subsequently affecting the overall convergence. Three benchmarks are utilized: two-dimensional flow over a cylinder, an inverse problem of inlet velocity determination, and the Burger equation. It is found that incorporating weights generated in the first training phase into the structure of a neural network neutralizes the effects of imbalance factors. For instance, in the first benchmark, as a result of our process, the second phase of training is balanced across a wide range of ratios and is not affected by the initial state of weights, while the Vanilla-PINN failed to converge in most cases. Lastly, the initial training process not only eliminates the need for hyperparameter tuning to balance the loss function, but it also outperforms the Vanilla-PINN in terms of speed.
Authors: Zahra Zamanzadeh Darban, Geoffrey I. Webb, Shirui Pan, Charu C. Aggarwal, Mahsa Salehi
Abstract: One main challenge in time series anomaly detection (TSAD) is the lack of labelled data in many real-life scenarios. Most of the existing anomaly detection methods focus on learning the normal behaviour of unlabelled time series in an unsupervised manner. The normal boundary is often defined tightly, resulting in slight deviations being classified as anomalies, consequently leading to a high false positive rate and a limited ability to generalise normal patterns. To address this, we introduce a novel end-to-end self-supervised ContrAstive Representation Learning approach for time series Anomaly detection (CARLA). While existing contrastive learning methods assume that augmented time series windows are positive samples and temporally distant windows are negative samples, we argue that these assumptions are limited as augmentation of time series can transform them to negative samples, and a temporally distant window can represent a positive sample. Our contrastive approach leverages existing generic knowledge about time series anomalies and injects various types of anomalies as negative samples. Therefore, CARLA not only learns normal behaviour but also learns deviations indicating anomalies. It creates similar representations for temporally closed windows and distinct ones for anomalies. Additionally, it leverages the information about representations' neighbours through a self-supervised approach to classify windows based on their nearest/furthest neighbours to further enhance the performance of anomaly detection. In extensive tests on seven major real-world time series anomaly detection datasets, CARLA shows superior performance over state-of-the-art self-supervised and unsupervised TSAD methods. Our research shows the potential of contrastive representation learning to advance time series anomaly detection.
Authors: Axel Laborieux, Friedemann Zenke
Abstract: Equilibrium propagation (EP) is a compelling alternative to the backpropagation of error algorithm (BP) for computing gradients of neural networks on biological or analog neuromorphic substrates. Still, the algorithm requires weight symmetry and infinitesimal equilibrium perturbations, i.e., nudges, to estimate unbiased gradients efficiently. Both requirements are challenging to implement in physical systems. Yet, whether and how weight asymmetry affects its applicability is unknown because, in practice, it may be masked by biases introduced through the finite nudge. To address this question, we study generalized EP, which can be formulated without weight symmetry, and analytically isolate the two sources of bias. For complex-differentiable non-symmetric networks, we show that the finite nudge does not pose a problem, as exact derivatives can still be estimated via a Cauchy integral. In contrast, weight asymmetry introduces bias resulting in low task performance due to poor alignment of EP's neuronal error vectors compared to BP. To mitigate this issue, we present a new homeostatic objective that directly penalizes functional asymmetries of the Jacobian at the network's fixed point. This homeostatic objective dramatically improves the network's ability to solve complex tasks such as ImageNet 32x32. Our results lay the theoretical groundwork for studying and mitigating the adverse effects of imperfections of physical networks on learning algorithms that rely on the substrate's relaxation dynamics.
Authors: Ran Wei, Siliang Zeng, Chenliang Li, Alfredo Garcia, Anthony McDonald, Mingyi Hong
Abstract: We consider a Bayesian approach to offline model-based inverse reinforcement learning (IRL). The proposed framework differs from existing offline model-based IRL approaches by performing simultaneous estimation of the expert's reward function and subjective model of environment dynamics. We make use of a class of prior distributions which parameterizes how accurate the expert's model of the environment is to develop efficient algorithms to estimate the expert's reward and subjective dynamics in high-dimensional settings. Our analysis reveals a novel insight that the estimated policy exhibits robust performance when the expert is believed (a priori) to have a highly accurate model of the environment. We verify this observation in the MuJoCo environments and show that our algorithms outperform state-of-the-art offline IRL algorithms.
Authors: Derek Yadgaroff, Alessandro Sestini, Konrad Tollmar, Ayca Ozcelikkale, Linus Gissl\'en
Abstract: Imitation learning is an effective approach for training game-playing agents and, consequently, for efficient game production. However, generalization - the ability to perform well in related but unseen scenarios - is an essential requirement that remains an unsolved challenge for game AI. Generalization is difficult for imitation learning agents because it requires the algorithm to take meaningful actions outside of the training distribution. In this paper we propose a solution to this challenge. Inspired by the success of data augmentation in supervised learning, we augment the training data so the distribution of states and actions in the dataset better represents the real state-action distribution. This study evaluates methods for combining and applying data augmentations to observations, to improve generalization of imitation learning agents. It also provides a performance benchmark of these augmentations across several 3D environments. These results demonstrate that data augmentation is a promising framework for improving generalization in imitation learning agents.
Authors: Haoran Wang, Zeshen Tang, Leya Yang, Yaoru Sun, Fang Wang, Siyu Zhang, Yeming Chen
Abstract: Goal-conditioned hierarchical reinforcement learning (HRL) presents a promising approach for enabling effective exploration in complex, long-horizon reinforcement learning (RL) tasks through temporal abstraction. Empirically, heightened inter-level communication and coordination can induce more stable and robust policy improvement in hierarchical systems. Yet, most existing goal-conditioned HRL algorithms have primarily focused on the subgoal discovery, neglecting inter-level cooperation. Here, we propose a goal-conditioned HRL framework named Guided Cooperation via Model-based Rollout (GCMR), aiming to bridge inter-layer information synchronization and cooperation by exploiting forward dynamics. Firstly, the GCMR mitigates the state-transition error within off-policy correction via model-based rollout, thereby enhancing sample efficiency. Secondly, to prevent disruption by the unseen subgoals and states, lower-level Q-function gradients are constrained using a gradient penalty with a model-inferred upper bound, leading to a more stable behavioral policy conducive to effective exploration. Thirdly, we propose a one-step rollout-based planning, using higher-level critics to guide the lower-level policy. Specifically, we estimate the value of future states of the lower-level policy using the higher-level critic function, thereby transmitting global task information downwards to avoid local pitfalls. These three critical components in GCMR are expected to facilitate inter-level cooperation significantly. Experimental results demonstrate that incorporating the proposed GCMR framework with a disentangled variant of HIGL, namely ACLG, yields more stable and robust policy improvement compared to various baselines and significantly outperforms previous state-of-the-art algorithms.
Authors: Yuezhu Xu, S. Sivaranjani
Abstract: Consider an unknown nonlinear dynamical system that is known to be dissipative. The objective of this paper is to learn a neural dynamical model that approximates this system, while preserving the dissipativity property in the model. In general, imposing dissipativity constraints during neural network training is a hard problem for which no known techniques exist. In this work, we address the problem of learning a dissipative neural dynamical system model in two stages. First, we learn an unconstrained neural dynamical model that closely approximates the system dynamics. Next, we derive sufficient conditions to perturb the weights of the neural dynamical model to ensure dissipativity, followed by perturbation of the biases to retain the fit of the model to the trajectories of the nonlinear system. We show that these two perturbation problems can be solved independently to obtain a neural dynamical model that is guaranteed to be dissipative while closely approximating the nonlinear system.
Authors: Tony Shen, Seonghwan Seo, Grayson Lee, Mohit Pandey, Jason R Smith, Artem Cherkasov, Woo Youn Kim, Martin Ester
Abstract: Searching the vast chemical space for drug-like and synthesizable molecules with high binding affinity to a protein pocket is a challenging task in drug discovery. Recently, molecular deep generative models have been introduced which promise to be more efficient than exhaustive virtual screening, by directly generating molecules based on the protein structure. However, since they learn the distribution of a limited protein-ligand complex dataset, the existing methods struggle with generating novel molecules with significant property improvements. In this paper, we frame the generation task as a Reinforcement Learning task, where the goal is to search the wider chemical space for molecules with desirable properties as opposed to fitting a training data distribution. More specifically, we propose TacoGFN, a Generative Flow Network conditioned on protein pocket structure, using binding affinity, drug-likeliness and synthesizability measures as our reward. Empirically, our method outperforms state-of-art methods on the CrossDocked2020 benchmark for every molecular property (Vina score, QED, SA), while significantly improving the generation time. TacoGFN achieves $-8.82$ in median docking score and $52.63\%$ in Novel Hit Rate.
Authors: Ran Wei, Nathan Lambert, Anthony McDonald, Alfredo Garcia, Roberto Calandra
Abstract: Model-based Reinforcement Learning (MBRL) aims to make agents more sample-efficient, adaptive, and explainable by learning an explicit model of the environment. While the capabilities of MBRL agents have significantly improved in recent years, how to best learn the model is still an unresolved question. The majority of MBRL algorithms aim at training the model to make accurate predictions about the environment and subsequently using the model to determine the most rewarding actions. However, recent research has shown that model predictive accuracy is often not correlated with action quality, tracing the root cause to the objective mismatch between accurate dynamics model learning and policy optimization of rewards. A number of interrelated solution categories to the objective mismatch problem have emerged as MBRL continues to mature as a research area. In this work, we provide an in-depth survey of these solution categories and propose a taxonomy to foster future research.
Authors: Mirco Mutti, Riccardo De Santi, Marcello Restelli, Alexander Marx, Giorgia Ramponi
Abstract: Posterior sampling allows exploitation of prior knowledge on the environment's transition dynamics to improve the sample efficiency of reinforcement learning. The prior is typically specified as a class of parametric distributions, the design of which can be cumbersome in practice, often resulting in the choice of uninformative priors. In this work, we propose a novel posterior sampling approach in which the prior is given as a (partial) causal graph over the environment's variables. The latter is often more natural to design, such as listing known causal dependencies between biometric features in a medical treatment study. Specifically, we propose a hierarchical Bayesian procedure, called C-PSRL, simultaneously learning the full causal graph at the higher level and the parameters of the resulting factored dynamics at the lower level. We provide an analysis of the Bayesian regret of C-PSRL that explicitly connects the regret rate with the degree of prior knowledge. Our numerical evaluation conducted in illustrative domains confirms that C-PSRL strongly improves the efficiency of posterior sampling with an uninformative prior while performing close to posterior sampling with the full causal graph.
Authors: Juliette Achddou, Nicol\`o Cesa-Bianchi, Pierre Laforgue
Abstract: We study multitask online learning in a setting where agents can only exchange information with their neighbors on an arbitrary communication network. We introduce $\texttt{MT-CO}_2\texttt{OL}$, a decentralized algorithm for this setting whose regret depends on the interplay between the task similarities and the network structure. Our analysis shows that the regret of $\texttt{MT-CO}_2\texttt{OL}$ is never worse (up to constants) than the bound obtained when agents do not share information. On the other hand, our bounds significantly improve when neighboring agents operate on similar tasks. In addition, we prove that our algorithm can be made differentially private with a negligible impact on the regret. Finally, we provide experimental support for our theory.
Authors: Valentin Duruisseaux, Amit Chakraborty
Abstract: In numerous contexts, high-resolution solutions to partial differential equations are required to capture faithfully essential dynamics which occur at small spatiotemporal scales, but these solutions can be very difficult and slow to obtain using traditional methods due to limited computational resources. A recent direction to circumvent these computational limitations is to use machine learning techniques for super-resolution, to reconstruct high-resolution numerical solutions from low-resolution simulations which can be obtained more efficiently. The proposed approach, the Super Resolution Operator Network (SROpNet), frames super-resolution as an operator learning problem and draws inspiration from existing architectures to learn continuous representations of solutions to parametric differential equations from low-resolution approximations, which can then be evaluated at any desired location. In addition, no restrictions are imposed on the locations of (the fixed number of) spatiotemporal sensors at which the low-resolution approximations are provided, thereby enabling the consideration of a broader spectrum of problems arising in practice, for which many existing super-resolution approaches are not well-suited.
Authors: Matt Gorbett, Hossein Shirazi, Indrakshi Ray
Abstract: Learning from the collective knowledge of data dispersed across private sources can provide neural networks with enhanced generalization capabilities. Federated learning, a method for collaboratively training a machine learning model across remote clients, achieves this by combining client models via the orchestration of a central server. However, current approaches face two critical limitations: i) they struggle to converge when client domains are sufficiently different, and ii) current aggregation techniques produce an identical global model for each client. In this work, we address these issues by reformulating the typical federated learning setup: rather than learning a single global model, we learn N models each optimized for a common objective. To achieve this, we apply a weighted distance minimization to model parameters shared in a peer-to-peer topology. The resulting framework, Iterative Parameter Alignment, applies naturally to the cross-silo setting, and has the following properties: (i) a unique solution for each participant, with the option to globally converge each model in the federation, and (ii) an optional early-stopping mechanism to elicit fairness among peers in collaborative learning settings. These characteristics jointly provide a flexible new framework for iteratively learning from peer models trained on disparate datasets. We find that the technique achieves competitive results on a variety of data partitions compared to state-of-the-art approaches. Further, we show that the method is robust to divergent domains (i.e. disjoint classes across peers) where existing approaches struggle.
Authors: Boumediene Hamzi, Marcus Hutter, Houman Owhadi
Abstract: Machine Learning (ML) and Algorithmic Information Theory (AIT) look at Complexity from different points of view. We explore the interface between AIT and Kernel Methods (that are prevalent in ML) by adopting an AIT perspective on the problem of learning kernels from data, in kernel ridge regression, through the method of Sparse Kernel Flows. In particular, by looking at the differences and commonalities between Minimal Description Length (MDL) and Regularization in Machine Learning (RML), we prove that the method of Sparse Kernel Flows is the natural approach to adopt to learn kernels from data. This approach aligns naturally with the MDL principle, offering a more robust theoretical basis than the existing reliance on cross-validation. The study reveals that deriving Sparse Kernel Flows does not require a statistical approach; instead, one can directly engage with code-lengths and complexities, concepts central to AIT. Thereby, this approach opens the door to reformulating algorithms in machine learning using tools from AIT, with the aim of providing them a more solid theoretical foundation.
Authors: Dengke Yan, Ming Hu, Zeke Xia, Yanxin Yang, Jun Xia, Xiaofei Xie, Mingsong Chen
Abstract: Due to its advantages in resource constraint scenarios, Split Federated Learning (SFL) is promising in AIoT systems. However, due to data heterogeneity and stragglers, SFL suffers from the challenges of low inference accuracy and low efficiency. To address these issues, this paper presents a novel SFL approach, named Sliding Split Federated Learning (S$^2$FL), which adopts an adaptive sliding model split strategy and a data balance-based training mechanism. By dynamically dispatching different model portions to AIoT devices according to their computing capability, S$^2$FL can alleviate the low training efficiency caused by stragglers. By combining features uploaded by devices with different data distributions to generate multiple larger batches with a uniform distribution for back-propagation, S$^2$FL can alleviate the performance degradation caused by data heterogeneity. Experimental results demonstrate that, compared to conventional SFL, S$^2$FL can achieve up to 16.5\% inference accuracy improvement and 3.54X training acceleration.
Authors: Haoyu Jiang, Haiyang Yu, Nan Li, Ping Yi
Abstract: Deep neural networks (DNNs) have been found vulnerable to backdoor attacks, raising security concerns about their deployment in mission-critical applications. There are various approaches to detect backdoor attacks, however they all make certain assumptions about the target attack to be detected and require equal and huge numbers of clean and backdoor samples for training, which renders these detection methods quite limiting in real-world circumstances. This study proposes a novel one-class classification framework called One-class Graph Embedding Classification (OCGEC) that uses GNNs for model-level backdoor detection with only a little amount of clean data. First, we train thousands of tiny models as raw datasets from a small number of clean datasets. Following that, we design a ingenious model-to-graph method for converting the model's structural details and weight features into graph data. We then pre-train a generative self-supervised graph autoencoder (GAE) to better learn the features of benign models in order to detect backdoor models without knowing the attack strategy. After that, we dynamically combine the GAE and one-class classifier optimization goals to form classification boundaries that distinguish backdoor models from benign models. Our OCGEC combines the powerful representation capabilities of graph neural networks with the utility of one-class classification techniques in the field of anomaly detection. In comparison to other baselines, it achieves AUC scores of more than 98% on a number of tasks, which far exceeds existing methods for detection even when they rely on a huge number of positive and negative samples. Our pioneering application of graphic scenarios for generic backdoor detection can provide new insights that can be used to improve other backdoor defense tasks. Code is available at https://github.com/jhy549/OCGEC.
Authors: Sina Baharlouei, Shivam Patel, Meisam Razaviyayn
Abstract: Training and deploying machine learning models that meet fairness criteria for protected groups are fundamental in modern artificial intelligence. While numerous constraints and regularization terms have been proposed in the literature to promote fairness in machine learning tasks, most of these methods are not amenable to stochastic optimization due to the complex and nonlinear structure of constraints and regularizers. Here, the term "stochastic" refers to the ability of the algorithm to work with small mini-batches of data. Motivated by the limitation of existing literature, this paper presents a unified stochastic optimization framework for fair empirical risk minimization based on f-divergence measures (f-FERM). The proposed stochastic algorithm enjoys theoretical convergence guarantees. In addition, our experiments demonstrate the superiority of fairness-accuracy tradeoffs offered by f-FERM for almost all batch sizes (ranging from full-batch to batch size of one). Moreover, we show that our framework can be extended to the case where there is a distribution shift from training to the test data. Our extension is based on a distributionally robust optimization reformulation of f-FERM objective under $L_p$ norms as uncertainty sets. Again, in this distributionally robust setting, f-FERM not only enjoys theoretical convergence guarantees but also outperforms other baselines in the literature in the tasks involving distribution shifts. An efficient stochastic implementation of $f$-FERM is publicly available.
Authors: Zhijun Zeng, Pipi Hu, Chenglong Bao, Yi Zhu, Zuoqiang Shi
Abstract: In this paper, we study the method to reconstruct dynamical systems from data without time labels. Data without time labels appear in many applications, such as molecular dynamics, single-cell RNA sequencing etc. Reconstruction of dynamical system from time sequence data has been studied extensively. However, these methods do not apply if time labels are unknown. Without time labels, sequence data becomes distribution data. Based on this observation, we propose to treat the data as samples from a probability distribution and try to reconstruct the underlying dynamical system by minimizing the distribution loss, sliced Wasserstein distance more specifically. Extensive experiment results demonstrate the effectiveness of the proposed method.
Authors: Tom Tirer, Raja Giryes, Se Young Chun, Yonina C. Eldar
Abstract: Deep learning, in general, focuses on training a neural network from large labeled datasets. Yet, in many cases there is value in training a network just from the input at hand. This is particularly relevant in many signal and image processing problems where training data is scarce and diversity is large on the one hand, and on the other, there is a lot of structure in the data that can be exploited. Using this information is the key to deep internal-learning strategies, which may involve training a network from scratch using a single input or adapting an already trained network to a provided input example at inference time. This survey paper aims at covering deep internal-learning techniques that have been proposed in the past few years for these two important directions. While our main focus will be on image processing problems, most of the approaches that we survey are derived for general signals (vectors with recurring patterns that can be distinguished from noise) and are therefore applicable to other modalities.
Authors: Qirui Ji, Jiangmeng Li, Jie Hu, Rui Wang, Changwen Zheng, Fanjiang Xu
Abstract: Graph contrastive learning is a general learning paradigm excelling at capturing invariant information from diverse perturbations in graphs. Recent works focus on exploring the structural rationale from graphs, thereby increasing the discriminability of the invariant information. However, such methods may incur in the mis-learning of graph models towards the interpretability of graphs, and thus the learned noisy and task-agnostic information interferes with the prediction of graphs. To this end, with the purpose of exploring the intrinsic rationale of graphs, we accordingly propose to capture the dimensional rationale from graphs, which has not received sufficient attention in the literature. The conducted exploratory experiments attest to the feasibility of the aforementioned roadmap. To elucidate the innate mechanism behind the performance improvement arising from the dimensional rationale, we rethink the dimensional rationale in graph contrastive learning from a causal perspective and further formalize the causality among the variables in the pre-training stage to build the corresponding structural causal model. On the basis of the understanding of the structural causal model, we propose the dimensional rationale-aware graph contrastive learning approach, which introduces a learnable dimensional rationale acquiring network and a redundancy reduction constraint. The learnable dimensional rationale acquiring network is updated by leveraging a bi-level meta-learning technique, and the redundancy reduction constraint disentangles the redundant features through a decorrelation process during learning. Empirically, compared with state-of-the-art methods, our method can yield significant performance boosts on various benchmarks with respect to discriminability and transferability. The code implementation of our method is available at https://github.com/ByronJi/DRGCL.
Authors: Hongzheng Chen, Jiahao Zhang, Yixiao Du, Shaojie Xiang, Zichao Yue, Niansong Zhang, Yaohui Cai, Zhiru Zhang
Abstract: Recent advancements in large language models (LLMs) boasting billions of parameters have generated a significant demand for efficient deployment in inference workloads. The majority of existing approaches rely on temporal architectures that reuse hardware units for different network layers and operators. However, these methods often encounter challenges in achieving low latency due to considerable memory access overhead. This paper investigates the feasibility and potential of model-specific spatial acceleration for LLM inference on FPGAs. Our approach involves the specialization of distinct hardware units for specific operators or layers, facilitating direct communication between them through a dataflow architecture while minimizing off-chip memory accesses. We introduce a comprehensive analytical model for estimating the performance of a spatial LLM accelerator, taking into account the on-chip compute and memory resources available on an FPGA. Through our analysis, we can determine the scenarios in which FPGA-based spatial acceleration can outperform its GPU-based counterpart. To enable more productive implementations of an LLM model on FPGAs, we further provide a library of high-level synthesis (HLS) kernels that are composable and reusable. This library will be made available as open-source. To validate the effectiveness of both our analytical model and HLS library, we have implemented BERT and GPT2 on an AMD Alveo U280 FPGA device. Experimental results demonstrate our approach can achieve up to 13.4x speedup when compared to previous FPGA-based accelerators for the BERT model. For GPT generative inference, we attain a 2.2x speedup compared to DFX, an FPGA overlay, in the prefill stage, while achieving a 1.9x speedup and a 5.7x improvement in energy efficiency compared to the NVIDIA A100 GPU in the decode stage.
Authors: Hideaki Takahashi
Abstract: This paper introduces AIJack, an open-source library designed to assess security and privacy risks associated with the training and deployment of machine learning models. Amid the growing interest in big data and AI, advancements in machine learning research and business are accelerating. However, recent studies reveal potential threats, such as the theft of training data and the manipulation of models by malicious attackers. Therefore, a comprehensive understanding of machine learning's security and privacy vulnerabilities is crucial for the safe integration of machine learning into real-world products. AIJack aims to address this need by providing a library with various attack and defense methods through a unified API. The library is publicly available on GitHub (https://github.com/Koukyosyumei/AIJack).
Authors: Vijay Ekambaram, Arindam Jati, Nam H. Nguyen, Pankaj Dayama, Chandra Reddy, Wesley M. Gifford, Jayant Kalagnanam
Abstract: Large pre-trained models for zero/few-shot learning excel in language and vision domains but encounter challenges in multivariate time series (TS) due to the diverse nature and scarcity of publicly available pre-training data. Consequently, there has been a recent surge in utilizing pre-trained large language models (LLMs) with token adaptations for TS forecasting. These approaches employ cross-domain transfer learning and surprisingly yield impressive results. However, these models are typically very slow and large (~billion parameters) and do not consider cross-channel correlations. To address this, we present Tiny Time Mixers (TTM), a significantly small model based on the lightweight TSMixer architecture. TTM marks the first success in developing fast and tiny general pre-trained models (<1M parameters), exclusively trained on public TS datasets, with effective transfer learning capabilities for forecasting. To tackle the complexity of pre-training on multiple datasets with varied temporal resolutions, we introduce several novel enhancements such as adaptive patching, dataset augmentation via downsampling, and resolution prefix tuning. Moreover, we employ a multi-level modeling strategy to effectively model channel correlations and infuse exogenous signals during fine-tuning, a crucial capability lacking in existing benchmarks. TTM shows significant accuracy gains (12-38\%) over popular benchmarks in few/zero-shot forecasting. It also drastically reduces the compute needs as compared to LLM-TS methods, with a 14X cut in learnable parameters, 106X less total parameters, and substantial reductions in fine-tuning (65X) and inference time (54X). In fact, TTM's zero-shot often surpasses the few-shot results in many popular benchmarks, highlighting the efficacy of our approach. Code and pre-trained models will be open-sourced.
Authors: Kaixin Huang, Li Shen, Chen Zhao, Chun Yuan, Dacheng Tao
Abstract: Continuous offline reinforcement learning (CORL) combines continuous and offline reinforcement learning, enabling agents to learn multiple tasks from static datasets without forgetting prior tasks. However, CORL faces challenges in balancing stability and plasticity. Existing methods, employing Actor-Critic structures and experience replay (ER), suffer from distribution shifts, low efficiency, and weak knowledge-sharing. We aim to investigate whether Decision Transformer (DT), another offline RL paradigm, can serve as a more suitable offline continuous learner to address these issues. We first compare AC-based offline algorithms with DT in the CORL framework. DT offers advantages in learning efficiency, distribution shift mitigation, and zero-shot generalization but exacerbates the forgetting problem during supervised parameter updates. We introduce multi-head DT (MH-DT) and low-rank adaptation DT (LoRA-DT) to mitigate DT's forgetting problem. MH-DT stores task-specific knowledge using multiple heads, facilitating knowledge sharing with common components. It employs distillation and selective rehearsal to enhance current task learning when a replay buffer is available. In buffer-unavailable scenarios, LoRA-DT merges less influential weights and fine-tunes DT's decisive MLP layer to adapt to the current task. Extensive experiments on MoJuCo and Meta-World benchmarks demonstrate that our methods outperform SOTA CORL baselines and showcase enhanced learning capabilities and superior memory efficiency.
Authors: Yihan Du, R. Srikant, Wei Chen
Abstract: Cascading bandits have gained popularity in recent years due to their applicability to recommendation systems and online advertising. In the cascading bandit model, at each timestep, an agent recommends an ordered subset of items (called an item list) from a pool of items, each associated with an unknown attraction probability. Then, the user examines the list, and clicks the first attractive item (if any), and after that, the agent receives a reward. The goal of the agent is to maximize the expected cumulative reward. However, the prior literature on cascading bandits ignores the influences of user states (e.g., historical behaviors) on recommendations and the change of states as the session proceeds. Motivated by this fact, we propose a generalized cascading RL framework, which considers the impact of user states and state transition into decisions. In cascading RL, we need to select items not only with large attraction probabilities but also leading to good successor states. This imposes a huge computational challenge due to the combinatorial action space. To tackle this challenge, we delve into the properties of value functions, and design an oracle BestPerm to efficiently find the optimal item list. Equipped with BestPerm, we develop two algorithms CascadingVI and CascadingBPI, which are both computationally-efficient and sample-efficient, and provide near-optimal regret and sample complexity guarantees. Furthermore, we present experiments to show the improved computational and sample efficiencies of our algorithms compared to straightforward adaptations of existing RL algorithms in practice.
Authors: Zhiyuan Wang, Xiaoyang Qu, Jing Xiao, Bokui Chen, Jianzong Wang
Abstract: This paper introduces INCPrompt, an innovative continual learning solution that effectively addresses catastrophic forgetting. INCPrompt's key innovation lies in its use of adaptive key-learner and task-aware prompts that capture task-relevant information. This unique combination encapsulates general knowledge across tasks and encodes task-specific knowledge. Our comprehensive evaluation across multiple continual learning benchmarks demonstrates INCPrompt's superiority over existing algorithms, showing its effectiveness in mitigating catastrophic forgetting while maintaining high performance. These results highlight the significant impact of task-aware incremental prompting on continual learning performance.
Authors: Zichen Liu, Chao Du, Wee Sun Lee, Min Lin
Abstract: Acquiring an accurate world model online for model-based reinforcement learning (MBRL) is challenging due to data nonstationarity, which typically causes catastrophic forgetting for neural networks (NNs). From the online learning perspective, a Follow-The-Leader (FTL) world model is desirable, which optimally fits all previous experiences at each round. Unfortunately, NN-based models need re-training on all accumulated data at every interaction step to achieve FTL, which is computationally expensive for lifelong agents. In this paper, we revisit models that can achieve FTL with incremental updates. Specifically, our world model is a linear regression model supported by nonlinear random features. The linear part ensures efficient FTL update while the nonlinear random feature empowers the fitting of complex environments. To best trade off model capacity and computation efficiency, we introduce a locality sensitive sparse encoding, which allows us to conduct efficient sparse updates even with very high dimensional nonlinear features. We validate the representation power of our encoding and verify that it allows efficient online learning under data covariate shift. We also show, in the Dyna MBRL setting, that our world models learned online using a single pass of trajectory data either surpass or match the performance of deep world models trained with replay and other continual learning methods.
Authors: Zhengpeng Xie
Abstract: PPO (Proximal Policy Optimization) algorithm has demonstrated excellent performance in many fields, and it is considered as a simple version of TRPO (Trust Region Policy Optimization) algorithm. However, the ratio clipping operation in PPO may not always effectively enforce the trust region constraints, this can be a potential factor affecting the stability of the algorithm. In this paper, we propose Simple Policy Optimization (SPO) algorithm, which introduces a novel clipping method for KL divergence between the old and current policies. Extensive experimental results in Atari 2600 environments indicate that, compared to the mainstream variants of PPO, SPO achieves better sample efficiency, extremely low KL divergence, and higher policy entropy, and is robust to the increase in network depth or complexity. More importantly, SPO maintains the simplicity of an unconstrained first-order algorithm. Code is available at https://github.com/MyRepositories-hub/Simple-Policy-Optimization.
URLs: https://github.com/MyRepositories-hub/Simple-Policy-Optimization.
Authors: Lifan Zhao, Yanyan Shen
Abstract: Recently, channel-independent methods have achieved state-of-the-art performance in multivariate time series (MTS) forecasting. Despite reducing overfitting risks, these methods miss potential opportunities in utilizing channel dependence for accurate predictions. We argue that there exist locally stationary lead-lag relationships between variates, i.e., some lagged variates may follow the leading indicators within a short time period. Exploiting such channel dependence is beneficial since leading indicators offer advance information that can be used to reduce the forecasting difficulty of the lagged variates. In this paper, we propose a new method named LIFT that first efficiently estimates leading indicators and their leading steps at each time step and then judiciously allows the lagged variates to utilize the advance information from leading indicators. LIFT plays as a plugin that can be seamlessly collaborated with arbitrary time series forecasting methods. Extensive experiments on six real-world datasets demonstrate that LIFT improves the state-of-the-art methods by 5.5% in average forecasting performance. Our code is available at https://github.com/SJTU-Quant/LIFT.
Authors: Yiming Sun, Yuhe Gao, Runxue Bao, Gregory F. Cooper, Jessi Espino, Harry Hochheiser, Marian G. Michaels, John M. Aronis, Chenxi Song, Ye Ye
Abstract: Transfer learning has become a pivotal technique in machine learning and has proven to be effective in various real-world applications. However, utilizing this technique for classification tasks with sequential data often faces challenges, primarily attributed to the scarcity of class labels. To address this challenge, we introduce Multi-Source Adaptive Weighting (MSAW), an online multi-source transfer learning method. MSAW integrates a dynamic weighting mechanism into an ensemble framework, enabling automatic adjustment of weights based on the relevance and contribution of each source (representing historical knowledge) and target model (learning from newly acquired data). We demonstrate the effectiveness of MSAW by applying it to detect Respiratory Syncytial Virus cases within Emergency Department visits, utilizing multiple years of electronic health records from the University of Pittsburgh Medical Center. Our method demonstrates performance improvements over many baselines, including refining pre-trained models with online learning as well as three static weighting approaches, showing MSAW's capacity to integrate historical knowledge with progressively accumulated new data. This study indicates the potential of online transfer learning in healthcare, particularly for developing machine learning models that dynamically adapt to evolving situations where new data is incrementally accumulated.
Authors: Shaogang Ren, Xiaoning Qian
Abstract: Maximizing a target variable as an operational objective in a structured causal model is an important problem. Existing Causal Bayesian Optimization (CBO) methods either rely on hard interventions that alter the causal structure to maximize the reward; or introduce action nodes to endogenous variables so that the data generation mechanisms are adjusted to achieve the objective. In this paper, a novel method is introduced to learn the distribution of exogenous variables, which is typically ignored or marginalized through expectation by existing methods. Exogenous distribution learning improves the approximation accuracy of structured causal models in a surrogate model that is usually trained with limited observational data. Moreover, the learned exogenous distribution extends existing CBO to general causal schemes beyond Additive Noise Models (ANM). The recovery of exogenous variables allows us to use a more flexible prior for noise or unobserved hidden variables. A new CBO method is developed by leveraging the learned exogenous distribution. Experiments on different datasets and applications show the benefits of our proposed method.
Authors: Rui Jiao, Wenbing Huang, Yu Liu, Deli Zhao, Yang Liu
Abstract: Crystals are the foundation of numerous scientific and industrial applications. While various learning-based approaches have been proposed for crystal generation, existing methods seldom consider the space group constraint which is crucial in describing the geometry of crystals and closely relevant to many desirable properties. However, considering space group constraint is challenging owing to its diverse and nontrivial forms. In this paper, we reduce the space group constraint into an equivalent formulation that is more tractable to be handcrafted into the generation process. In particular, we translate the space group constraint into two parts: the basis constraint of the invariant logarithmic space of the lattice matrix and the Wyckoff position constraint of the fractional coordinates. Upon the derived constraints, we then propose DiffCSP++, a novel diffusion model that has enhanced a previous work DiffCSP by further taking space group constraint into account. Experiments on several popular datasets verify the benefit of the involvement of the space group constraint, and show that our DiffCSP++ achieves promising performance on crystal structure prediction, ab initio crystal generation and controllable generation with customized space groups.
Authors: Pranav Kulkarni, Andrew Chan, Nithya Navarathna, Skylar Chan, Paul H. Yi, Vishwa S. Parekh
Abstract: The proliferation of artificial intelligence (AI) in radiology has shed light on the risk of deep learning (DL) models exacerbating clinical biases towards vulnerable patient populations. While prior literature has focused on quantifying biases exhibited by trained DL models, demographically targeted adversarial bias attacks on DL models and its implication in the clinical environment remains an underexplored field of research in medical imaging. In this work, we demonstrate that demographically targeted label poisoning attacks can introduce undetectable underdiagnosis bias in DL models. Our results across multiple performance metrics and demographic groups like sex, age, and their intersectional subgroups show that adversarial bias attacks demonstrate high-selectivity for bias in the targeted group by degrading group model performance without impacting overall model performance. Furthermore, our results indicate that adversarial bias attacks result in biased DL models that propagate prediction bias even when evaluated with external datasets.
Authors: Yehui Tang, Yunhe Wang, Jianyuan Guo, Zhijun Tu, Kai Han, Hailin Hu, Dacheng Tao
Abstract: Transformer plays a vital role in the realms of natural language processing (NLP) and computer vision (CV), specially for constructing large language models (LLM) and large vision models (LVM). Model compression methods reduce the memory and computational cost of Transformer, which is a necessary step to implement large language/vision models on practical devices. Given the unique architecture of Transformer, featuring alternative attention and feedforward neural network (FFN) modules, specific compression techniques are usually required. The efficiency of these compression methods is also paramount, as retraining large models on the entire training dataset is usually impractical. This survey provides a comprehensive review of recent compression methods, with a specific focus on their application to Transformer-based models. The compression methods are primarily categorized into pruning, quantization, knowledge distillation, and efficient architecture design (Mamba, RetNet, RWKV, etc.). In each category, we discuss compression methods for both language and vision tasks, highlighting common underlying principles. Finally, we delve into the relation between various compression methods, and discuss further directions in this domain.
Authors: Haonan Yuan, Qingyun Sun, Xingcheng Fu, Cheng Ji, Jianxin Li
Abstract: Dynamic Graphs widely exist in the real world, which carry complicated spatial and temporal feature patterns, challenging their representation learning. Dynamic Graph Neural Networks (DGNNs) have shown impressive predictive abilities by exploiting the intrinsic dynamics. However, DGNNs exhibit limited robustness, prone to adversarial attacks. This paper presents the novel Dynamic Graph Information Bottleneck (DGIB) framework to learn robust and discriminative representations. Leveraged by the Information Bottleneck (IB) principle, we first propose the expected optimal representations should satisfy the Minimal-Sufficient-Consensual (MSC) Condition. To compress redundant as well as conserve meritorious information into latent representation, DGIB iteratively directs and refines the structural and feature information flow passing through graph snapshots. To meet the MSC Condition, we decompose the overall IB objectives into DGIB$_{MS}$ and DGIB$_C$, in which the DGIB$_{MS}$ channel aims to learn the minimal and sufficient representations, with the DGIB$_{MS}$ channel guarantees the predictive consensus. Extensive experiments on real-world and synthetic dynamic graph datasets demonstrate the superior robustness of DGIB against adversarial attacks compared with state-of-the-art baselines in the link prediction task. To the best of our knowledge, DGIB is the first work to learn robust representations of dynamic graphs grounded in the information-theoretic IB principle.
Authors: Chester Luo, Kevin Lai
Abstract: In recent years, there has been a significant increase in the utilization of deep learning methods, particularly convolutional neural networks (CNNs), which have emerged as the dominant approach in various domains that involve structured grid data, such as picture analysis and processing. Nevertheless, the exponential growth in the utilization of LiDAR and 3D sensors across many domains has resulted in an increased need for the analysis of 3D point clouds. The utilization of 3D point clouds is crucial in various applications, including object recognition and segmentation, as they offer a spatial depiction of things within a three-dimensional environment. In contrast to photos, point clouds exhibit sparsity and lack a regular grid, hence posing distinct processing and computational issues.
Authors: Ayush K Tarun, Vikram S Chundawat, Murari Mandal, Hong Ming Tan, Bowei Chen, Mohan Kankanhalli
Abstract: Quantifying the value of data within a machine learning workflow can play a pivotal role in making more strategic decisions in machine learning initiatives. The existing Shapley value based frameworks for data valuation in machine learning are computationally expensive as they require considerable amount of repeated training of the model to obtain the Shapley value. In this paper, we introduce an efficient data valuation framework EcoVal, to estimate the value of data for machine learning models in a fast and practical manner. Instead of directly working with individual data sample, we determine the value of a cluster of similar data points. This value is further propagated amongst all the member cluster points. We show that the overall data value can be determined by estimating the intrinsic and extrinsic value of each data. This is enabled by formulating the performance of a model as a \textit{production function}, a concept which is popularly used to estimate the amount of output based on factors like labor and capital in a traditional free economic market. We provide a formal proof of our valuation technique and elucidate the principles and mechanisms that enable its accelerated performance. We demonstrate the real-world applicability of our method by showcasing its effectiveness for both in-distribution and out-of-sample data. This work addresses one of the core challenges of efficient data valuation at scale in machine learning models.
Authors: Tianyi Qiu, Fanzhi Zeng, Jiaming Ji, Dong Yan, Kaile Wang, Jiayi Zhou, Yang Han, Josef Dai, Xuehai Pan, Yaodong Yang
Abstract: There is a trilemma in reinforcement learning from human feedback (RLHF): the incompatibility between highly diverse contexts, low labeling cost, and reliable alignment performance. We mitigate such incompatibility through the design of dataset information structures during reward modeling, and introduce the Induced Bayesian Network (IBN), the first theory of reward generalization capable of generating substantial verified predictions on large language models (LLMs). Specifically, we first reexamine the RLHF process and propose a theoretical framework portraying it as an autoencoding process over text distributions. Our framework formalizes the RLHF objective of ensuring distributional consistency between human preference and LLM behavior. Then, based on this framework, we introduce the IBN to analyze generalization in the reward modeling stage of RLHF. Drawing from random graph theory and causal analysis, it enables empirically grounded derivation of generalization error bounds, a key improvement over classical theories of generalization. Finally, an insight from our analysis is the superiority of the tree-based information structure in reward modeling, compared to chain-based baselines in conventional RLHF methods. With IBN, we derive that in complex contexts with limited data, the tree-based reward model (RM), trained on a tree-structured preference dataset, induces up to $\Theta(\log n/\log\log n)$ times less variance than the baseline, where $n$ is the dataset size. As validation, we demonstrate that on three NLP tasks, the tree-based RM achieves 65% win rate on average against chain-based baselines. It shows that alignment performance can be gained for free via the design of dataset information structure, without the need for any other changes.
Authors: Haoming Li, Yusen Huo, Shuai Dou, Zhenzhe Zheng, Zhilin Zhang, Chuan Yu, Jian Xu, Fan Wu
Abstract: In online advertising, advertisers participate in ad auctions to acquire ad opportunities, often by utilizing auto-bidding tools provided by demand-side platforms (DSPs). The current auto-bidding algorithms typically employ reinforcement learning (RL). However, due to safety concerns, most RL-based auto-bidding policies are trained in simulation, leading to a performance degradation when deployed in online environments. To narrow this gap, we can deploy multiple auto-bidding agents in parallel to collect a large interaction dataset. Offline RL algorithms can then be utilized to train a new policy. The trained policy can subsequently be deployed for further data collection, resulting in an iterative training framework, which we refer to as iterative offline RL. In this work, we identify the performance bottleneck of this iterative offline RL framework, which originates from the ineffective exploration and exploitation caused by the inherent conservatism of offline RL algorithms. To overcome this bottleneck, we propose Trajectory-wise Exploration and Exploitation (TEE), which introduces a novel data collecting and data utilization method for iterative offline RL from a trajectory perspective. Furthermore, to ensure the safety of online exploration while preserving the dataset quality for TEE, we propose Safe Exploration by Adaptive Action Selection (SEAS). Both offline experiments and real-world experiments on Alibaba display advertising platform demonstrate the effectiveness of our proposed method.
Authors: Siyuan Guo, Cheng Deng, Ying Wen, Hechang Chen, Yi Chang, Jun Wang
Abstract: In this work, we investigate the potential of large language models (LLMs) based agents to automate data science tasks, with the goal of comprehending task requirements, then building and training the best-fit machine learning models. Despite their widespread success, existing LLM agents are hindered by generating unreasonable experiment plans within this scenario. To this end, we present DS-Agent, a novel automatic framework that harnesses LLM agent and case-based reasoning (CBR). In the development stage, DS-Agent follows the CBR framework to structure an automatic iteration pipeline, which can flexibly capitalize on the expert knowledge from Kaggle, and facilitate consistent performance improvement through the feedback mechanism. Moreover, DS-Agent implements a low-resource deployment stage with a simplified CBR paradigm to adapt past successful solutions from the development stage for direct code generation, significantly reducing the demand on foundational capabilities of LLMs. Empirically, DS-Agent with GPT-4 achieves an unprecedented 100% success rate in the development stage, while attaining 36% improvement on average one pass rate across alternative LLMs in the deployment stage. In both stages, DS-Agent achieves the best rank in performance, costing \$1.60 and \$0.13 per run with GPT-4, respectively. Our code is open-sourced at https://github.com/guosyjlu/DS-Agent.
Authors: Chenhui Deng, Zichao Yue, Zhiru Zhang
Abstract: Graph transformers (GTs) have emerged as a promising architecture that is theoretically more expressive than message-passing graph neural networks (GNNs). However, typical GT models have at least quadratic complexity and thus cannot scale to large graphs. While there are several linear GTs recently proposed, they still lag behind GNN counterparts on several popular graph datasets, which poses a critical concern on their practical expressivity. To balance the trade-off between expressivity and scalability of GTs, we propose Polynormer, a polynomial-expressive GT model with linear complexity. Polynormer is built upon a novel base model that learns a high-degree polynomial on input features. To enable the base model permutation equivariant, we integrate it with graph topology and node features separately, resulting in local and global equivariant attention models. Consequently, Polynormer adopts a linear local-to-global attention scheme to learn high-degree equivariant polynomials whose coefficients are controlled by attention scores. Polynormer has been evaluated on $13$ homophilic and heterophilic datasets, including large graphs with millions of nodes. Our extensive experiment results show that Polynormer outperforms state-of-the-art GNN and GT baselines on most datasets, even without the use of nonlinear activation functions.
Authors: Chenhui Deng, Zichao Yue, Cunxi Yu, Gokce Sarar, Ryan Carey, Rajeev Jain, Zhiru Zhang
Abstract: While graph neural networks (GNNs) have gained popularity for learning circuit representations in various electronic design automation (EDA) tasks, they face challenges in scalability when applied to large graphs and exhibit limited generalizability to new designs. These limitations make them less practical for addressing large-scale, complex circuit problems. In this work we propose HOGA, a novel attention-based model for learning circuit representations in a scalable and generalizable manner. HOGA first computes hop-wise features per node prior to model training. Subsequently, the hop-wise features are solely used to produce node representations through a gated self-attention module, which adaptively learns important features among different hops without involving the graph topology. As a result, HOGA is adaptive to various structures across different circuits and can be efficiently trained in a distributed manner. To demonstrate the efficacy of HOGA, we consider two representative EDA tasks: quality of results (QoR) prediction and functional reasoning. Our experimental results indicate that (1) HOGA reduces estimation error over conventional GNNs by 46.76% for predicting QoR after logic synthesis; (2) HOGA improves 10.0% reasoning accuracy over GNNs for identifying functional blocks on unseen gate-level netlists after complex technology mapping; (3) The training time for HOGA almost linearly decreases with an increase in computing resources.
Authors: Sourav Ganguly, Saprativa Bhattacharjee
Abstract: Modern deep learning tools are remarkably effective in addressing intricate problems. However, their operation as black-box models introduces increased uncertainty in predictions. Additionally, they contend with various challenges, including the need for substantial storage space in large networks, issues of overfitting, underfitting, vanishing gradients, and more. This study explores the concept of Bayesian Neural Networks, presenting a novel architecture designed to significantly alleviate the storage space complexity of a network. Furthermore, we introduce an algorithm adept at efficiently handling uncertainties, ensuring robust convergence values without becoming trapped in local optima, particularly when the objective function lacks perfect convexity.
Authors: Muhammad Adnan, Akhil Arunkumar, Gaurav Jain, Prashant J. Nair, Ilya Soloveychik, Purushotham Kamath
Abstract: Transformers have emerged as the underpinning architecture for Large Language Models (LLMs). In generative language models, the inference process involves two primary phases: prompt processing and token generation. Token generation, which constitutes the majority of the computational workload, primarily entails vector-matrix multiplications and interactions with the Key-Value (KV) Cache. This phase is constrained by memory bandwidth due to the overhead of transferring weights and KV cache values from the memory system to the computing units. This memory bottleneck becomes particularly pronounced in applications that require long-context and extensive text generation, both of which are increasingly crucial for LLMs. This paper introduces "Keyformer", an innovative inference-time approach, to mitigate the challenges associated with KV cache size and memory bandwidth utilization. Keyformer leverages the observation that approximately 90% of the attention weight in generative inference focuses on a specific subset of tokens, referred to as "key" tokens. Keyformer retains only the key tokens in the KV cache by identifying these crucial tokens using a novel score function. This approach effectively reduces both the KV cache size and memory bandwidth usage without compromising model accuracy. We evaluate Keyformer's performance across three foundational models: GPT-J, Cerebras-GPT, and MPT, which employ various positional embedding algorithms. Our assessment encompasses a variety of tasks, with a particular emphasis on summarization and conversation tasks involving extended contexts. Keyformer's reduction of KV cache reduces inference latency by 2.1x and improves token generation throughput by 2.4x, while preserving the model's accuracy.
Authors: Shokichi Takakura, Taiji Suzuki
Abstract: In this paper, we study the feature learning ability of two-layer neural networks in the mean-field regime through the lens of kernel methods. To focus on the dynamics of the kernel induced by the first layer, we utilize a two-timescale limit, where the second layer moves much faster than the first layer. In this limit, the learning problem is reduced to the minimization problem over the intrinsic kernel. Then, we show the global convergence of the mean-field Langevin dynamics and derive time and particle discretization error. We also demonstrate that two-layer neural networks can learn a union of multiple reproducing kernel Hilbert spaces more efficiently than any kernel methods, and neural networks acquire data-dependent kernel which aligns with the target function. In addition, we develop a label noise procedure, which converges to the global optimum and show that the degrees of freedom appears as an implicit regularization.
Authors: Carlos Gomes, Thomas Brunschwiler
Abstract: As repositories of large scale data in earth observation (EO) have grown, so have transfer and storage costs for model training and inference, expending significant resources. We introduce Neural Embedding Compression (NEC), based on the transfer of compressed embeddings to data consumers instead of raw data. We adapt foundation models (FM) through learned neural compression to generate multi-task embeddings while navigating the tradeoff between compression rate and embedding utility. We update only a small fraction of the FM parameters (10%) for a short training period (1% of the iterations of pre-training). We evaluate NEC on two EO tasks: scene classification and semantic segmentation. Compared with applying traditional compression to the raw data, NEC achieves similar accuracy with a 75% to 90% reduction in data. Even at 99.7% compression, performance drops by only 5% on the scene classification task. Overall, NEC is a data-efficient yet performant approach for multi-task EO modelling.
Authors: Shawn Im, Yixuan Li
Abstract: Aligning large language models (LLMs) with human intentions has become a critical task for safely deploying models in real-world systems. While existing alignment approaches have seen empirical success, theoretically understanding how these methods affect model behavior remains an open question. Our work provides an initial attempt to theoretically analyze the learning dynamics of human preference alignment. We formally show how the distribution of preference datasets influences the rate of model updates and provide rigorous guarantees on the training accuracy. Our theory also reveals an intricate phenomenon where the optimization is prone to prioritizing certain behaviors with higher preference distinguishability. We empirically validate our findings on contemporary LLMs and alignment tasks, reinforcing our theoretical insights and shedding light on considerations for future alignment approaches. Disclaimer: This paper contains potentially offensive text; reader discretion is advised.
Authors: Xiangfei Qiu, Jilin Hu, Lekui Zhou, Xingjian Wu, Junyang Du, Buang Zhang, Chenjuan Guo, Aoying Zhou, Christian S. Jensen, Zhenli Sheng, Bin Yang
Abstract: Time series are generated in diverse domains such as economic, traffic, health, and energy, where forecasting of future values has numerous important applications. Not surprisingly, many forecasting methods are being proposed. To ensure progress, it is essential to be able to study and compare such methods empirically in a comprehensive and reliable manner. To achieve this, we propose TFB, an automated benchmark for Time Series Forecasting (TSF) methods. TFB advances the state-of-the-art by addressing shortcomings related to datasets, comparison methods, and evaluation pipelines: 1) insufficient coverage of data domains, 2) stereotype bias against traditional methods, and 3) inconsistent and inflexible pipelines. To achieve better domain coverage, we include datasets from 10 different domains: traffic, electricity, energy, the environment, nature, economic, stock markets, banking, health, and the web. We also provide a time series characterization to ensure that the selected datasets are comprehensive. To remove biases against some methods, we include a diverse range of methods, including statistical learning, machine learning, and deep learning methods, and we also support a variety of evaluation strategies and metrics to ensure a more comprehensive evaluations of different methods. To support the integration of different methods into the benchmark and enable fair comparisons, TFB features a flexible and scalable pipeline that eliminates biases. Next, we employ TFB to perform a thorough evaluation of 21 Univariate Time Series Forecasting (UTSF) methods on 8,068 univariate time series and 14 Multivariate Time Series Forecasting (MTSF) methods on 25 datasets. The benchmark code and data are available at https://github.com/decisionintelligence/TFB.
Authors: Yazheng Yang, Yuqi Wang, Sankalok Sen, Lei Li, Qi Liu
Abstract: In the domain of data science, the predictive tasks of classification, regression, and imputation of missing values are commonly encountered challenges associated with tabular data. This research endeavors to apply Large Language Models (LLMs) towards addressing these predictive tasks. Despite their proficiency in comprehending natural language, LLMs fall short in dealing with structured tabular data. This limitation stems from their lacking exposure to the intricacies of tabular data during their foundational training. Our research aims to mitigate this gap by compiling a comprehensive corpus of tables annotated with instructions and executing large-scale training of Llama-2 on this enriched dataset. Furthermore, we investigate the practical application of applying the trained model to zero-shot prediction, few-shot prediction, and in-context learning scenarios. Through extensive experiments, our methodology has shown significant improvements over existing benchmarks. These advancements highlight the efficacy of tailoring LLM training to solve table-related problems in data science, thereby establishing a new benchmark in the utilization of LLMs for enhancing tabular intelligence.
Authors: Yao Ni, Piotr Koniusz
Abstract: Generative Adversarial Networks (GANs) significantly advanced image generation but their performance heavily depends on abundant training data. In scenarios with limited data, GANs often struggle with discriminator overfitting and unstable training. Batch Normalization (BN), despite being known for enhancing generalization and training stability, has rarely been used in the discriminator of Data-Efficient GANs. Our work addresses this gap by identifying a critical flaw in BN: the tendency for gradient explosion during the centering and scaling steps. To tackle this issue, we present CHAIN (lipsCHitz continuity constrAIned Normalization), which replaces the conventional centering step with zero-mean regularization and integrates a Lipschitz continuity constraint in the scaling step. CHAIN further enhances GAN training by adaptively interpolating the normalized and unnormalized features, effectively avoiding discriminator overfitting. Our theoretical analyses firmly establishes CHAIN's effectiveness in reducing gradients in latent features and weights, improving stability and generalization in GAN training. Empirical evidence supports our theory. CHAIN achieves state-of-the-art results in data-limited scenarios on CIFAR-10/100, ImageNet, five low-shot and seven high-resolution few-shot image datasets. Code: https://github.com/MaxwellYaoNi/CHAIN
Authors: Ali Akram, Marija Stanojevic, Malikeh Ehghaghi, Jekaterina Novikova
Abstract: Due to the substantial number of clinicians, patients, and data collection environments involved in clinical trials, gathering data of superior quality poses a significant challenge. In clinical trials, patients are assessed based on their speech data to detect and monitor cognitive and mental health disorders. We propose using these speech recordings to verify the identities of enrolled patients and identify and exclude the individuals who try to enroll multiple times in the same trial. Since clinical studies are often conducted across different countries, creating a system that can perform speaker verification in diverse languages without additional development effort is imperative. We evaluate pre-trained TitaNet, ECAPA-TDNN, and SpeakerNet models by enrolling and testing with speech-impaired patients speaking English, German, Danish, Spanish, and Arabic languages. Our results demonstrate that tested models can effectively generalize to clinical speakers, with less than 2.7% EER for European Languages and 8.26% EER for Arabic. This represents a significant step in developing more versatile and efficient speaker verification systems for cognitive and mental health clinical trials that can be used across a wide range of languages and dialects, substantially reducing the effort required to develop speaker verification systems for multiple languages. We also evaluate how speech tasks and number of speakers involved in the trial influence the performance and show that the type of speech tasks impacts the model performance.
Authors: Kamalika Chaudhuri, Chuan Guo, Laurens van der Maaten, Saeed Mahloujifar, Mark Tygert
Abstract: Protecting privacy during inference with deep neural networks is possible by adding noise to the activations in the last layers prior to the final classifiers or other task-specific layers. The activations in such layers are known as "features" (or, less commonly, as "embeddings" or "feature embeddings"). The added noise helps prevent reconstruction of the inputs from the noisy features. Lower bounding the variance of every possible unbiased estimator of the inputs quantifies the confidentiality arising from such added noise. Convenient, computationally tractable bounds are available from classic inequalities of Hammersley and of Chapman and Robbins -- the HCR bounds. Numerical experiments indicate that the HCR bounds are on the precipice of being effectual for small neural nets with the data sets, "MNIST" and "CIFAR-10," which contain 10 classes each for image classification. The HCR bounds appear to be insufficient on their own to guarantee confidentiality of the inputs to inference with standard deep neural nets, "ResNet-18" and "Swin-T," pre-trained on the data set, "ImageNet-1000," which contains 1000 classes. Supplementing the addition of noise to features with other methods for providing confidentiality may be warranted in the case of ImageNet. In all cases, the results reported here limit consideration to amounts of added noise that incur little degradation in the accuracy of classification from the noisy features. Thus, the added noise enhances confidentiality without much reduction in the accuracy on the task of image classification.
Authors: Zecheng Zhang
Abstract: The study of operator learning involves the utilization of neural networks to approximate operators. Traditionally, the focus has been on single-operator learning (SOL). However, recent advances have rapidly expanded this to include the approximation of multiple operators using foundation models equipped with millions or billions of trainable parameters, leading to the research of multi-operator learning (MOL). In this paper, we present a novel distributed training approach aimed at enabling a single neural operator with significantly fewer parameters to effectively tackle multi-operator learning challenges, all without incurring additional average costs. Our method is applicable to various neural operators, such as Deep Operator Neural Networks (DON). The core idea is to independently learn the output basis functions for each operator using its dedicated data, while simultaneously centralizing the learning of the input function encoding shared by all operators using the entire dataset. Through a systematic study of five numerical examples, we compare the accuracy and cost of training a single neural operator for each operator independently versus training a MOL model using our proposed method. Our results demonstrate enhanced efficiency and satisfactory accuracy. Moreover, our approach illustrates that some operators with limited data can be more effectively constructed with the aid of data from analogous operators through MOL learning. This highlights another MOL's potential to bolster operator learning.
Authors: Xusen Guo (Frank), Qiming Zhang (Frank), Mingxing Peng (Frank), Meixin Zhu (Frank), Hao (Frank), Yang
Abstract: Traffic flow prediction is crucial for urban planning, transportation management, and infrastructure development. However, achieving both accuracy and interpretability in prediction models remains challenging due to the complexity of traffic data and the inherent opacity of deep learning methodologies. In this paper, we propose a novel approach, Traffic Flow Prediction LLM (TF-LLM), which leverages large language models (LLMs) to generate interpretable traffic flow predictions. By transferring multi-modal traffic data into natural language descriptions, TF-LLM captures complex spatial-temporal patterns and external factors such as weather conditions, Points of Interest (PoIs), date, and holidays. We fine-tune the LLM framework using language-based instructions to align with spatial-temporal traffic flow data. Our comprehensive multi-modal traffic flow dataset (CATraffic) in California enables the evaluation of TF-LLM against state-of-the-art deep learning baselines. Results demonstrate TF-LLM's competitive accuracy while providing intuitive and interpretable predictions. We discuss the spatial-temporal and input dependencies for explainable future flow forecasting, showcasing TF-LLM's potential for diverse city prediction tasks. This paper contributes to advancing explainable traffic prediction models and lays a foundation for future exploration of LLM applications in transportation.
Authors: Wolfgang Roth, G\"unther Schindler, Bernhard Klein, Robert Peharz, Sebastian Tschiatschek, Holger Fr\"oning, Franz Pernkopf, Zoubin Ghahramani
Abstract: While machine learning is traditionally a resource intensive task, embedded systems, autonomous navigation, and the vision of the Internet of Things fuel the interest in resource-efficient approaches. These approaches aim for a carefully chosen trade-off between performance and resource consumption in terms of computation and energy. The development of such approaches is among the major challenges in current machine learning research and key to ensure a smooth transition of machine learning technology from a scientific environment with virtually unlimited computing resources into everyday's applications. In this article, we provide an overview of the current state of the art of machine learning techniques facilitating these real-world requirements. In particular, we focus on resource-efficient inference based on deep neural networks (DNNs), the predominant machine learning models of the past decade. We give a comprehensive overview of the vast literature that can be mainly split into three non-mutually exclusive categories: (i) quantized neural networks, (ii) network pruning, and (iii) structural efficiency. These techniques can be applied during training or as post-processing, and they are widely used to reduce the computational demands in terms of memory footprint, inference speed, and energy efficiency. We also briefly discuss different concepts of embedded hardware for DNNs and their compatibility with machine learning techniques as well as potential for energy and latency reduction. We substantiate our discussion with experiments on well-known benchmark data sets using compression techniques (quantization, pruning) for a set of resource-constrained embedded systems, such as CPUs, GPUs and FPGAs. The obtained results highlight the difficulty of finding good trade-offs between resource efficiency and prediction quality.
Authors: Jinglin Zhang, Wenjun Xu, Hui Gao, Miao Pan, Zhu Han, Ping Zhang
Abstract: Millimeter wave (mmWave) communications can potentially meet the high data-rate requirements of unmanned aerial vehicle (UAV) networks. However, as the prerequisite of mmWave communications, the narrow directional beam tracking is very challenging because of the three-dimensional (3D) mobility and attitude variation of UAVs. Aiming to address the beam tracking difficulties, we propose to integrate the conformal array (CA) with the surface of each UAV, which enables the full spatial coverage and the agile beam tracking in highly dynamic UAV mmWave networks. More specifically, the key contributions of our work are three-fold. 1) A new mmWave beam tracking framework is established for the CA-enabled UAV mmWave network. 2) A specialized hierarchical codebook is constructed to drive the directional radiating element (DRE)-covered cylindrical conformal array (CCA), which contains both the angular beam pattern and the subarray pattern to fully utilize the potential of the CA. 3) A codebook-based multiuser beam tracking scheme is proposed, where the Gaussian process machine learning enabled UAV position/attitude predication is developed to improve the beam tracking efficiency in conjunction with the tracking-error aware adaptive beamwidth control. Simulation results validate the effectiveness of the proposed codebook-based beam tracking scheme in the CA-enabled UAV mmWave network, and demonstrate the advantages of CA over the conventional planner array in terms of spectrum efficiency and outage probability in the highly dynamic scenarios.
Authors: Joaquim Dias Garcia, Alexandre Street, Tito Homem-de-Mello, Francisco D. Mu\~noz
Abstract: Forecasting and decision-making are generally modeled as two sequential steps with no feedback, following an open-loop approach. In this paper, we present application-driven learning, a new closed-loop framework in which the processes of forecasting and decision-making are merged and co-optimized through a bilevel optimization problem. We present our methodology in a general format and prove that the solution converges to the best estimator in terms of the expected cost of the selected application. Then, we propose two solution methods: an exact method based on the KKT conditions of the second-level problem and a scalable heuristic approach suitable for decomposition methods. The proposed methodology is applied to the relevant problem of defining dynamic reserve requirements and conditional load forecasts, offering an alternative approach to current ad hoc procedures implemented in industry practices. We benchmark our methodology with the standard sequential least-squares forecast and dispatch planning process. We apply the proposed methodology to an illustrative system and to a wide range of instances, from dozens of buses to large-scale realistic systems with thousands of buses. Our results show that the proposed methodology is scalable and yields consistently better performance than the standard open-loop approach.
Authors: Yujun Shi, Kuangqi Zhou, Jian Liang, Zihang Jiang, Jiashi Feng, Philip Torr, Song Bai, Vincent Y. F. Tan
Abstract: Class Incremental Learning (CIL) aims at learning a multi-class classifier in a phase-by-phase manner, in which only data of a subset of the classes are provided at each phase. Previous works mainly focus on mitigating forgetting in phases after the initial one. However, we find that improving CIL at its initial phase is also a promising direction. Specifically, we experimentally show that directly encouraging CIL Learner at the initial phase to output similar representations as the model jointly trained on all classes can greatly boost the CIL performance. Motivated by this, we study the difference between a na\"ively-trained initial-phase model and the oracle model. Specifically, since one major difference between these two models is the number of training classes, we investigate how such difference affects the model representations. We find that, with fewer training classes, the data representations of each class lie in a long and narrow region; with more training classes, the representations of each class scatter more uniformly. Inspired by this observation, we propose Class-wise Decorrelation (CwD) that effectively regularizes representations of each class to scatter more uniformly, thus mimicking the model jointly trained with all classes (i.e., the oracle model). Our CwD is simple to implement and easy to plug into existing methods. Extensive experiments on various benchmark datasets show that CwD consistently and significantly improves the performance of existing state-of-the-art methods by around 1\% to 3\%. Code will be released.
Authors: Veeranjaneyulu Sadhanala, Yu-Xiang Wang, Addison J. Hu, Ryan J. Tibshirani
Abstract: We study a multivariate version of trend filtering, called Kronecker trend filtering or KTF, for the case in which the design points form a lattice in $d$ dimensions. KTF is a natural extension of univariate trend filtering (Steidl et al., 2006; Kim et al., 2009; Tibshirani, 2014), and is defined by minimizing a penalized least squares problem whose penalty term sums the absolute (higher-order) differences of the parameter to be estimated along each of the coordinate directions. The corresponding penalty operator can be written in terms of Kronecker products of univariate trend filtering penalty operators, hence the name Kronecker trend filtering. Equivalently, one can view KTF in terms of an $\ell_1$-penalized basis regression problem where the basis functions are tensor products of falling factorial functions, a piecewise polynomial (discrete spline) basis that underlies univariate trend filtering. This paper is a unification and extension of the results in Sadhanala et al. (2016, 2017). We develop a complete set of theoretical results that describe the behavior of $k^{\mathrm{th}}$ order Kronecker trend filtering in $d$ dimensions, for every $k \geq 0$ and $d \geq 1$. This reveals a number of interesting phenomena, including the dominance of KTF over linear smoothers in estimating heterogeneously smooth functions, and a phase transition at $d=2(k+1)$, a boundary past which (on the high dimension-to-smoothness side) linear smoothers fail to be consistent entirely. We also leverage recent results on discrete splines from Tibshirani (2020), in particular, discrete spline interpolation results that enable us to extend the KTF estimate to any off-lattice location in constant-time (independent of the size of the lattice $n$).
Authors: Francisco Caldas, Cl\'audia Soares
Abstract: Since the late 1950s, when the first artificial satellite was launched, the number of Resident Space Objects has steadily increased. It is estimated that around one million objects larger than one cm are currently orbiting the Earth, with only thirty thousand larger than ten cm being tracked. To avert a chain reaction of collisions, known as Kessler Syndrome, it is essential to accurately track and predict debris and satellites' orbits. Current approximate physics-based methods have errors in the order of kilometers for seven-day predictions, which is insufficient when considering space debris, typically with less than one meter. This failure is usually due to uncertainty around the state of the space object at the beginning of the trajectory, forecasting errors in environmental conditions such as atmospheric drag, and unknown characteristics such as the mass or geometry of the space object. Operators can enhance Orbit Prediction accuracy by deriving unmeasured objects' characteristics and improving non-conservative forces' effects by leveraging data-driven techniques, such as Machine Learning. In this survey, we provide an overview of the work in applying Machine Learning for Orbit Determination, Orbit Prediction, and atmospheric density modeling.
Authors: Yu Lu, Yi-Jia Wang, Ying Chen, Jia-Jun Wu
Abstract: We present that by predicting the spectrum in discrete space from the phase shift in continuous space, the neural network can remarkably reproduce the numerical L\"uscher's formula to a high precision. The model-independent property of the L\"uscher's formula is naturally realized by the generalizability of the neural network. This exhibits the great potential of the neural network to extract model-independent relation between model-dependent quantities, and this data-driven approach could greatly facilitate the discovery of the physical principles underneath the intricate data.
Authors: Samir M. Perlaza, Gaetan Bisson, I\~naki Esnaola, Alain Jean-Marie, Stefano Rini
Abstract: The empirical risk minimization (ERM) problem with relative entropy regularization (ERM-RER) is investigated under the assumption that the reference measure is a $\sigma$-finite measure, and not necessarily a probability measure. Under this assumption, which leads to a generalization of the ERM-RER problem allowing a larger degree of flexibility for incorporating prior knowledge, numerous relevant properties are stated. Among these properties, the solution to this problem, if it exists, is shown to be a unique probability measure, mutually absolutely continuous with the reference measure. Such a solution exhibits a probably-approximately-correct guarantee for the ERM problem independently of whether the latter possesses a solution. For a fixed dataset and under a specific condition, the empirical risk is shown to be a sub-Gaussian random variable when the models are sampled from the solution to the ERM-RER problem. The generalization capabilities of the solution to the ERM-RER problem (the Gibbs algorithm) are studied via the sensitivity of the expected empirical risk to deviations from such a solution towards alternative probability measures. Finally, an interesting connection between sensitivity, generalization error, and lautum information is established.
Authors: Christian Rack, Konstantin Kobs, Tamara Fernando, Andreas Hotho, Marc Erich Latoschik
Abstract: Different approaches in machine learning have proven useful for user verification and identification based on motion data in eXtended Reality (XR). However, their real-world application still faces significant challenges concerning versatility, i.e., in terms of extensibility and generalization capability. This article presents a solution that is both extensible to new users without expensive retraining, and that generalizes well across different sessions, devices, and user tasks. To this end, we developed a similarity-learning model and pretrained it on the "Who Is Alyx?" dataset. This dataset features a wide array of actions from users playing the VR game "Half-Life: Alyx". In contrast to previous works, we used a dedicated set of users for model validation and final evaluation. Furthermore, we extended this evaluation using an independent dataset that features completely different users, tasks, and three different XR devices. In comparison with a traditional classification-learning baseline, our model shows superior performance, especially in scenarios with limited enrollment data. The pretraining process allows easy deployment in a diverse range of XR applications while maintaining high versatility. Looking ahead, our methodology can facilitate the easy integration of pretrained motion-based identification models into 3D engines: dedicated plugins for Unreal and Unity promise straightforward and quick adoption of these advanced models in the future, even for those lacking expertise in motion analysis and machine learning.
Authors: Yiqun Yao, Siqi fan, Xiusheng Huang, Xuezhi Fang, Xiang Li, Ziyi Ni, Xin Jiang, Xuying Meng, Peng Han, Shuo Shang, Kang Liu, Aixin Sun, Yequan Wang
Abstract: As language models scale up, it becomes increasingly expensive to verify research ideas because conclusions on small models do not trivially transfer to large ones. A possible solution is to establish a generic system that accurately predicts certain metrics for large models without training them. Existing scaling laws require hyperparameter search on the largest models, limiting their predicative capability. In this paper, we present an approach (namely {\mu}Scaling) to predict the pre-training loss, based on our observations that Maximal Update Parametrization ({\mu}P) enables accurate fitting of scaling laws close to common loss basins in hyperparameter space. With {\mu}Scaling, different model designs can be compared on large scales by training only their smaller counterparts. Further, we introduce nanoLM: an affordable LLM pre-training benchmark that facilitates this new research paradigm. With around 14% of the one-time pre-training cost, we can accurately forecast the loss for models up to 52B. Our goal with nanoLM is to empower researchers with limited resources to reach meaningful conclusions on large models. We also aspire for our benchmark to serve as a bridge between the academic community and the industry. Code for {\mu}Scaling is available at https://github.com/cofe-ai/Mu-scaling. Code for nanoLLM will be available later.
Authors: Yizhou Zhang, Loc Trinh, Defu Cao, Zijun Cui, Yan Liu
Abstract: Recent years have witnessed the sustained evolution of misinformation that aims at manipulating public opinions. Unlike traditional rumors or fake news editors who mainly rely on generated and/or counterfeited images, text and videos, current misinformation creators now more tend to use out-of-context multimedia contents (e.g. mismatched images and captions) to deceive the public and fake news detection systems. This new type of misinformation increases the difficulty of not only detection but also clarification, because every individual modality is close enough to true information. To address this challenge, in this paper we explore how to achieve interpretable cross-modal de-contextualization detection that simultaneously identifies the mismatched pairs and the cross-modal contradictions, which is helpful for fact-check websites to document clarifications. The proposed model first symbolically disassembles the text-modality information to a set of fact queries based on the Abstract Meaning Representation of the caption and then forwards the query-image pairs into a pre-trained large vision-language model select the ``evidences" that are helpful for us to detect misinformation. Extensive experiments indicate that the proposed methodology can provide us with much more interpretable predictions while maintaining the accuracy same as the state-of-the-art model on this task.
Authors: Sungwoo Kang, Jong-Kook Kim
Abstract: Despite the efficient market hypothesis, many studies suggest the existence of inefficiencies in the stock market leading to the development of techniques to gain above-market returns. Systematic trading has undergone significant advances in recent decades with deep learning schemes emerging as a powerful tool for analyzing and predicting market behavior. In this paper, a method is proposed that is inspired by how professional technical analysts trade. This scheme looks at stock prices of the previous 600 days and predicts whether the stock price will rise or fall 10% or 20% within the next D days. The proposed method uses the Resnet's (a deep learning model) skip connections and logits to increase the probability of the prediction. The model was trained and tested using historical data from both the Korea and US stock markets. The backtest is done using the data from 2020 to 2022. Using the proposed method for the Korea market it gave return of 75.36% having Sharpe ratio of 1.57, which far exceeds the market return by 36% and 0.61, respectively. On the US market it gives total return of 27.17% with Sharpe ratio of 0.61, which outperforms other benchmarks such as NASDAQ, S&P500, DOW JONES index by 17.69% and 0.27, respectively.
Authors: Anqi Wang, Jiahua Dong, Lik-Hang Lee, Jiachuan Shen, Pan Hui
Abstract: 3D shape generation techniques leveraging deep learning have garnered significant interest from both the computer vision and architectural design communities, promising to enrich the content of the future metaverse. However, research on virtual architectural design remains limited, particularly regarding human-AI collaboration and deep learning-assisted design. We first illuminate the principles, generation techniques, and current literature of virtual architecture, focusing on challenges such as datasets, multimodality, design intuition, and generative frameworks. In our survey, we reviewed 187 related articles (80.7\% of articles published between 2018 and 2022) covering architectural research, virtual environments, and technical approaches. This survey investigates the latest approaches to 3D object generation with deep generative models (DGMs) and summarizes four characteristics of deep-learning generation approaches for virtual architecture. According to our analysis of the survey, we expound on four research agendas, including agency, communication, user consideration, and integrating tools, and highlight three important enablers of ubiquitous interaction with immersive systems in deep learning-assisted architectural generation. Our work contributes to fostering understanding between designers and deep learning techniques, broadening access to human-AI collaboration. We advocate for interdisciplinary efforts to address this timely research topic, facilitating content designing and generation in the metaverse.
Authors: Junsol Kim, Byungkyu Lee
Abstract: Large language models (LLMs) that produce human-like responses have begun to revolutionize research practices in the social sciences. We develop a novel methodological framework that fine-tunes LLMs with repeated cross-sectional surveys to incorporate the meaning of survey questions, individual beliefs, and temporal contexts for opinion prediction. We introduce two new emerging applications of the AI-augmented survey: retrodiction (i.e., predict year-level missing responses) and unasked opinion prediction (i.e., predict entirely missing responses). Among 3,110 binarized opinions from 68,846 Americans in the General Social Survey from 1972 to 2021, our models based on Alpaca-7b excel in retrodiction (AUC = 0.86 for personal opinion prediction, $\rho$ = 0.98 for public opinion prediction). These remarkable prediction capabilities allow us to fill in missing trends with high confidence and pinpoint when public attitudes changed, such as the rising support for same-sex marriage. On the other hand, our fine-tuned Alpaca-7b models show modest success in unasked opinion prediction (AUC = 0.73, $\rho$ = 0.67). We discuss practical constraints and ethical concerns regarding individual autonomy and privacy when using LLMs for opinion prediction. Our study demonstrates that LLMs and surveys can mutually enhance each other's capabilities: LLMs can broaden survey potential, while surveys can improve the alignment of LLMs.
Authors: Jordan Meadows, Marco Valentino, Damien Teney, Andre Freitas
Abstract: This paper proposes a methodology for generating and perturbing detailed derivations of equations at scale, aided by a symbolic engine, to evaluate the generalisability of Transformers to out-of-distribution mathematical reasoning problems. Instantiating the framework in the context of sequence classification tasks, we compare the capabilities of GPT-4, GPT-3.5, and a canon of fine-tuned BERT models, exploring the relationship between specific operators and generalisation failure via the perturbation of reasoning aspects such as symmetry and variable surface forms. Surprisingly, our empirical evaluation reveals that the average in-distribution performance of fine-tuned models surpasses GPT-3.5, and rivals GPT-4. However, perturbations to input reasoning can reduce their performance by up to 80 F1 points. Overall, the results suggest that the in-distribution performance of smaller open-source models may potentially rival GPT by incorporating appropriately structured derivation dependencies during training, and highlight a shared weakness between BERT and GPT involving a relative inability to decode indirect references to mathematical entities. We release the full codebase, constructed datasets, and fine-tuned models to encourage future progress in the field.
Authors: Jaehun Jung, Peter West, Liwei Jiang, Faeze Brahman, Ximing Lu, Jillian Fisher, Taylor Sorensen, Yejin Choi
Abstract: We present Impossible Distillation, a novel framework for paraphrasing and sentence summarization, that distills a high-quality dataset and model from a low-quality teacher that itself cannot perform these tasks. Unlike prior works that rely on an extreme-scale teacher model (e.g., GPT3) or task-specific architecture, we hypothesize and verify the paraphrastic proximity intrinsic to pre-trained LMs (e.g., GPT2), where paraphrases occupy a proximal subspace in the LM distribution. By identifying and distilling generations from these subspaces, Impossible Distillation produces a high-quality dataset and model even from GPT2-scale LMs. We evaluate our method on multiple benchmarks spanning unconstrained / syntax-controlled paraphrase generation and sentence summarization. Our model with 770M parameters consistently outperforms strong baselines, including models distilled from ChatGPT, and sometimes, even ChatGPT itself. Also, we find that our distilled dataset from 1.5B LMs exhibits higher diversity and fidelity than up to 13 times larger datasets.
Authors: Hongjie Wang, Bhishma Dedhia, Niraj K. Jha
Abstract: Deployment of Transformer models on edge devices is becoming increasingly challenging due to the exponentially growing inference cost that scales quadratically with the number of tokens in the input sequence. Token pruning is an emerging solution to address this challenge due to its ease of deployment on various Transformer backbones. However, most token pruning methods require computationally expensive fine-tuning, which is undesirable in many edge deployment cases. In this work, we propose Zero-TPrune, the first zero-shot method that considers both the importance and similarity of tokens in performing token pruning. It leverages the attention graph of pre-trained Transformer models to produce an importance distribution for tokens via our proposed Weighted Page Rank (WPR) algorithm. This distribution further guides token partitioning for efficient similarity-based pruning. Due to the elimination of the fine-tuning overhead, Zero-TPrune can prune large models at negligible computational cost, switch between different pruning configurations at no computational cost, and perform hyperparameter tuning efficiently. We evaluate the performance of Zero-TPrune on vision tasks by applying it to various vision Transformer backbones and testing them on ImageNet. Without any fine-tuning, Zero-TPrune reduces the FLOPs cost of DeiT-S by 34.7% and improves its throughput by 45.3% with only 0.4% accuracy loss. Compared with state-of-the-art pruning methods that require fine-tuning, Zero-TPrune not only eliminates the need for fine-tuning after pruning but also does so with only 0.1% accuracy loss. Compared with state-of-the-art fine-tuning-free pruning methods, Zero-TPrune reduces accuracy loss by up to 49% with similar FLOPs budgets. Project webpage: https://jha-lab.github.io/zerotprune.
Authors: Yubo Zhuang, Xiaohui Chen, Yun Yang, Richard Y. Zhang
Abstract: $K$-means clustering is a widely used machine learning method for identifying patterns in large datasets. Semidefinite programming (SDP) relaxations have recently been proposed for solving the $K$-means optimization problem that enjoy strong statistical optimality guarantees, but the prohibitive cost of implementing an SDP solver renders these guarantees inaccessible to practical datasets. By contrast, nonnegative matrix factorization (NMF) is a simple clustering algorithm that is widely used by machine learning practitioners, but without a solid statistical underpinning nor rigorous guarantees. In this paper, we describe an NMF-like algorithm that works by solving a nonnegative low-rank restriction of the SDP relaxed $K$-means formulation using a nonconvex Burer--Monteiro factorization approach. The resulting algorithm is just as simple and scalable as state-of-the-art NMF algorithms, while also enjoying the same strong statistical optimality guarantees as the SDP. In our experiments, we observe that our algorithm achieves substantially smaller mis-clustering errors compared to the existing state-of-the-art.
Authors: Moe Kayali, Anton Lykov, Ilias Fountalis, Nikolaos Vasiloglou, Dan Olteanu, Dan Suciu
Abstract: We apply foundation models to data discovery and exploration tasks. Foundation models include large language models (LLMs) that show promising performance on a range of diverse tasks unrelated to their training. We show that these models are highly applicable to the data discovery and data exploration domain. When carefully used, they have superior capability on three representative tasks: table-class detection, column-type annotation and join-column prediction. On all three tasks, we show that a foundation-model-based approach outperforms the task-specific models and so the state of the art. Further, our approach often surpasses human-expert task performance. We investigate the fundamental characteristics of this approach including generalizability to several foundation models and the impact of non-determinism on the outputs. All in all, this suggests a future direction in which disparate data management tasks can be unified under foundation models.
Authors: Lifan Zhao, Shuming Kong, Yanyan Shen
Abstract: Stock trend forecasting is a fundamental task of quantitative investment where precise predictions of price trends are indispensable. As an online service, stock data continuously arrive over time. It is practical and efficient to incrementally update the forecast model with the latest data which may reveal some new patterns recurring in the future stock market. However, incremental learning for stock trend forecasting still remains under-explored due to the challenge of distribution shifts (a.k.a. concept drifts). With the stock market dynamically evolving, the distribution of future data can slightly or significantly differ from incremental data, hindering the effectiveness of incremental updates. To address this challenge, we propose DoubleAdapt, an end-to-end framework with two adapters, which can effectively adapt the data and the model to mitigate the effects of distribution shifts. Our key insight is to automatically learn how to adapt stock data into a locally stationary distribution in favor of profitable updates. Complemented by data adaptation, we can confidently adapt the model parameters under mitigated distribution shifts. We cast each incremental learning task as a meta-learning task and automatically optimize the adapters for desirable data adaptation and parameter initialization. Experiments on real-world stock datasets demonstrate that DoubleAdapt achieves state-of-the-art predictive performance and shows considerable efficiency.
Authors: Lukas Haas, Michal Skreta, Silas Alberti, Chelsea Finn
Abstract: Planet-scale image geolocalization remains a challenging problem due to the diversity of images originating from anywhere in the world. Although approaches based on vision transformers have made significant progress in geolocalization accuracy, success in prior literature is constrained to narrow distributions of images of landmarks, and performance has not generalized to unseen places. We present a new geolocalization system that combines semantic geocell creation, multi-task contrastive pretraining, and a novel loss function. Additionally, our work is the first to perform retrieval over location clusters for guess refinements. We train two models for evaluations on street-level data and general-purpose image geolocalization; the first model, PIGEON, is trained on data from the game of Geoguessr and is capable of placing over 40% of its guesses within 25 kilometers of the target location globally. We also develop a bot and deploy PIGEON in a blind experiment against humans, ranking in the top 0.01% of players. We further challenge one of the world's foremost professional Geoguessr players to a series of six matches with millions of viewers, winning all six games. Our second model, PIGEOTTO, differs in that it is trained on a dataset of images from Flickr and Wikipedia, achieving state-of-the-art results on a wide range of image geolocalization benchmarks, outperforming the previous SOTA by up to 7.7 percentage points on the city accuracy level and up to 38.8 percentage points on the country level. Our findings suggest that PIGEOTTO is the first image geolocalization model that effectively generalizes to unseen places and that our approach can pave the way for highly accurate, planet-scale image geolocalization systems. Our code is available on GitHub.
Authors: Xiaohao Yang, He Zhao, Dinh Phung, Lan Du
Abstract: Topic models have evolved from conventional Bayesian probabilistic models to recent Neural Topic Models (NTMs). Although NTMs have shown promising performance when trained and tested on a specific corpus, their generalisation ability across corpora has yet to be studied. In practice, we often expect that an NTM trained on a source corpus can still produce quality topical representation (i.e., latent distribution over topics) for the document from different target corpora. In this work, we aim to improve NTMs further so that their representation power for documents generalises reliably across corpora and tasks. To do so, we propose to enhance NTMs by narrowing the semantical distance between similar documents, with the underlying assumption that documents from different corpora may share similar semantics. Specifically, we obtain a similar document for each training document by text data augmentation. Then, we optimise NTMs further by minimising the semantical distance between each pair, measured by the Hierarchical Topic Transport Distance, which computes the Optimal Transport (OT) distance between their topical representations. Our framework can be readily applied to most NTMs as a plug-and-play module. Extensive experiments show that our framework significantly improves the generalisation ability regarding neural topical representation across corpora. Our code and datasets are available at: https://github.com/Xiaohao-Yang/Topic_Model_Generalisation
URLs: https://github.com/Xiaohao-Yang/Topic_Model_Generalisation
Authors: Shashank Jere, Lizhong Zheng, Karim Said, Lingjia Liu
Abstract: Recurrent neural networks (RNNs) are known to be universal approximators of dynamic systems under fairly mild and general assumptions. However, RNNs usually suffer from the issues of vanishing and exploding gradients in standard RNN training. Reservoir computing (RC), a special RNN where the recurrent weights are randomized and left untrained, has been introduced to overcome these issues and has demonstrated superior empirical performance especially in scenarios where training samples are extremely limited. On the other hand, the theoretical grounding to support this observed performance has yet been fully developed. In this work, we show that RC can universally approximate a general linear time-invariant (LTI) system. Specifically, we present a clear signal processing interpretation of RC and utilize this understanding in the problem of approximating a generic LTI system. Under this setup, we analytically characterize the optimum probability density function for configuring (instead of training and/or randomly generating) the recurrent weights of the underlying RNN of the RC. Extensive numerical evaluations are provided to validate the optimality of the derived distribution for configuring the recurrent weights of the RC to approximate a general LTI system. Our work results in clear signal processing-based model interpretability of RC and provides theoretical explanation/justification for the power of randomness in randomly generating instead of training RC's recurrent weights. Furthermore, it provides a complete optimum analytical characterization for configuring the untrained recurrent weights, marking an important step towards explainable machine learning (XML) to incorporate domain knowledge for efficient learning.
Authors: Frederic Wang, Han Qi, Alfredo De Goyeneche, Reinhard Heckel, Michael Lustig, Efrat Shimron
Abstract: Although deep learning (DL) methods are powerful for solving inverse problems, their reliance on high-quality training data is a major hurdle. This is significant in high-dimensional (dynamic/volumetric) magnetic resonance imaging (MRI), where acquisition of high-resolution fully sampled k-space data is impractical. We introduce a novel mathematical framework, dubbed k-band, that enables training DL models using only partial, limited-resolution k-space data. Specifically, we introduce training with stochastic gradient descent (SGD) over k-space subsets. In each training iteration, rather than using the fully sampled k-space for computing gradients, we use only a small k-space portion. This concept is compatible with different sampling strategies; here we demonstrate the method for k-space "bands", which have limited resolution in one dimension and can hence be acquired rapidly. We prove analytically that our method stochastically approximates the gradients computed in a fully-supervised setup, when two simple conditions are met: (i) the limited-resolution axis is chosen randomly-uniformly for every new scan, hence k-space is fully covered across the entire training set, and (ii) the loss function is weighed with a mask, derived here analytically, which facilitates accurate reconstruction of high-resolution details. Numerical experiments with raw MRI data indicate that k-band outperforms two other methods trained on limited-resolution data and performs comparably to state-of-the-art (SoTA) methods trained on high-resolution data. k-band hence obtains SoTA performance, with the advantage of training using only limited-resolution data. This work hence introduces a practical, easy-to-implement, self-supervised training framework, which involves fast acquisition and self-supervised reconstruction and offers theoretical guarantees.
Authors: Xingfang Wu, Heng Li, Foutse Khomh
Abstract: Logs are an essential source of information for people to understand the running status of a software system. Due to the evolving modern software architecture and maintenance methods, more research efforts have been devoted to automated log analysis. In particular, machine learning (ML) has been widely used in log analysis tasks. In ML-based log analysis tasks, converting textual log data into numerical feature vectors is a critical and indispensable step. However, the impact of using different log representation techniques on the performance of the downstream models is not clear, which limits researchers and practitioners' opportunities of choosing the optimal log representation techniques in their automated log analysis workflows. Therefore, this work investigates and compares the commonly adopted log representation techniques from previous log analysis research. Particularly, we select six log representation techniques and evaluate them with seven ML models and four public log datasets (i.e., HDFS, BGL, Spirit and Thunderbird) in the context of log-based anomaly detection. We also examine the impacts of the log parsing process and the different feature aggregation approaches when they are employed with log representation techniques. From the experiments, we provide some heuristic guidelines for future researchers and developers to follow when designing an automated log analysis workflow. We believe our comprehensive comparison of log representation techniques can help researchers and practitioners better understand the characteristics of different log representation techniques and provide them with guidance for selecting the most suitable ones for their ML-based log analysis workflow.
Authors: Guilherme Schardong, Tiago Novello, Hallison Paz, Iurii Medvedev, Vin\'icius da Silva, Luiz Velho, Nuno Gon\c{c}alves
Abstract: Face morphing is a problem in computer graphics with numerous artistic and forensic applications. It is challenging due to variations in pose, lighting, gender, and ethnicity. This task consists of a warping for feature alignment and a blending for a seamless transition between the warped images. We propose to leverage coord-based neural networks to represent such warpings and blendings of face images. During training, we exploit the smoothness and flexibility of such networks by combining energy functionals employed in classical approaches without discretizations. Additionally, our method is time-dependent, allowing a continuous warping/blending of the images. During morphing inference, we need both direct and inverse transformations of the time-dependent warping. The first (second) is responsible for warping the target (source) image into the source (target) image. Our neural warping stores those maps in a single network dismissing the need for inverting them. The results of our experiments indicate that our method is competitive with both classical and generative models under the lens of image quality and face-morphing detectors. Aesthetically, the resulting images present a seamless blending of diverse faces not yet usual in the literature.
Authors: Md Kaykobad Reza, Ashley Prater-Bennette, M. Salman Asif
Abstract: Leveraging information across diverse modalities is known to enhance performance on multimodal segmentation tasks. However, effectively fusing information from different modalities remains challenging due to the unique characteristics of each modality. In this paper, we propose a novel fusion strategy that can effectively fuse information from different modality combinations. We also propose a new model named Multi-Modal Segmentation TransFormer (MMSFormer) that incorporates the proposed fusion strategy to perform multimodal material and semantic segmentation tasks. MMSFormer outperforms current state-of-the-art models on three different datasets. As we begin with only one input modality, performance improves progressively as additional modalities are incorporated, showcasing the effectiveness of the fusion block in combining useful information from diverse input modalities. Ablation studies show that different modules in the fusion block are crucial for overall model performance. Furthermore, our ablation studies also highlight the capacity of different input modalities to improve performance in the identification of different types of materials. The code and pretrained models will be made available at https://github.com/csiplab/MMSFormer.
Authors: Karl Schrader, Joachim Weickert, Michael Krause
Abstract: Anisotropic diffusion processes with a diffusion tensor are important in image analysis, physics, and engineering. However, their numerical approximation has a strong impact on dissipative artefacts and deviations from rotation invariance. In this work, we study a large family of finite difference discretisations on a 3 x 3 stencil. We derive it by splitting 2-D anisotropic diffusion into four 1-D diffusions. The resulting stencil class involves one free parameter and covers a wide range of existing discretisations. It comprises the full stencil family of Weickert et al. (2013) and shows that their two parameters contain redundancy. Furthermore, we establish a bound on the spectral norm of the matrix corresponding to the stencil. This gives time step size limits that guarantee stability of an explicit scheme in the Euclidean norm. Our directional splitting also allows a very natural translation of the explicit scheme into ResNet blocks. Employing neural network libraries enables simple and highly efficient parallel implementations on GPUs.
Authors: Purui Zhang, Xingchao Jian, Feng Ji, Wee Peng Tay, Bihan Wen
Abstract: Topological signal processing (TSP) utilizes simplicial complexes to model structures with higher order than vertices and edges. In this paper, we study the transferability of TSP via a generalized higher-order version of graphon, known as complexon. We recall the notion of a complexon as the limit of a simplicial complex sequence [1]. Inspired by the graphon shift operator and message-passing neural network, we construct a marginal complexon and complexon shift operator (CSO) according to components of all possible dimensions from the complexon. We investigate the CSO's eigenvalues and eigenvectors and relate them to a new family of weighted adjacency matrices. We prove that when a simplicial complex signal sequence converges to a complexon signal, the eigenvalues, eigenspaces, and Fourier transform of the corresponding CSOs converge to that of the limit complexon signal. This conclusion is further verified by two numerical experiments. These results hint at learning transferability on large simplicial complexes or simplicial complex sequences, which generalize the graphon signal processing framework.
Authors: Shengyu Mao, Xiaohan Wang, Mengru Wang, Yong Jiang, Pengjun Xie, Fei Huang, Ningyu Zhang
Abstract: This paper introduces an innovative task focused on editing the personality traits of Large Language Models (LLMs). This task seeks to adjust the models' responses to opinion-related questions on specified topics since an individual's personality often manifests in the form of their expressed opinions, thereby showcasing different personality traits. Specifically, we construct a new benchmark dataset PersonalityEdit to address this task. Drawing on the theory in Social Psychology, we isolate three representative traits, namely Neuroticism, Extraversion, and Agreeableness, as the foundation for our benchmark. We then gather data using GPT-4, generating responses that not only align with a specified topic but also embody the targeted personality trait. We conduct comprehensive experiments involving various baselines and discuss the representation of personality behavior in LLMs. Our intriguing findings uncover potential challenges of the proposed task, illustrating several remaining issues. We anticipate that our work can provide the NLP community with insights. Code and datasets are available at https://github.com/zjunlp/EasyEdit.
Authors: Jing Liu, Ruihao Gong, Xiuying Wei, Zhiwei Dong, Jianfei Cai, Bohan Zhuang
Abstract: Large Language Models (LLMs) excel in NLP, but their demands hinder their widespread deployment. While Quantization-Aware Training (QAT) offers a solution, its extensive training costs make Post-Training Quantization (PTQ) a more practical approach for LLMs. In existing studies, activation outliers in particular channels are identified as the bottleneck to PTQ accuracy. They propose to transform the magnitudes from activations to weights, which however offers limited alleviation or suffers from unstable gradients, resulting in a severe performance drop at low-bitwidth. In this paper, we propose QLLM, an accurate and efficient low-bitwidth PTQ method designed for LLMs. QLLM introduces an adaptive channel reassembly technique that reallocates the magnitude of outliers to other channels, thereby mitigating their impact on the quantization range. This is achieved by channel disassembly and channel assembly, which first breaks down the outlier channels into several sub-channels to ensure a more balanced distribution of activation magnitudes. Then similar channels are merged to maintain the original channel number for efficiency. Additionally, an adaptive strategy is designed to autonomously determine the optimal number of sub-channels for channel disassembly. To further compensate for the performance loss caused by quantization, we propose an efficient tuning method that only learns a small number of low-rank weights while freezing the pre-trained quantized model. After training, these low-rank parameters can be fused into the frozen weights without affecting inference. Extensive experiments on LLaMA-1 and LLaMA-2 show that QLLM can obtain accurate quantized models efficiently. For example, QLLM quantizes the 4-bit LLaMA-2-70B within 10 hours on a single A100-80G GPU, outperforming the previous state-of-the-art method by 7.89% on the average accuracy across five zero-shot tasks.
Authors: Yoshio Ebihara, Xin Dai, Victor Magron, Dimitri Peaucelle, Sophie Tarbouriech
Abstract: This paper is concerned with the computation of the local Lipschitz constant of feedforward neural networks (FNNs) with activation functions being rectified linear units (ReLUs). The local Lipschitz constant of an FNN for a target input is a reasonable measure for its quantitative evaluation of the reliability. By following a standard procedure using multipliers that capture the behavior of ReLUs,we first reduce the upper bound computation problem of the local Lipschitz constant into a semidefinite programming problem (SDP). Here we newly introduce copositive multipliers to capture the ReLU behavior accurately. Then, by considering the dual of the SDP for the upper bound computation, we second derive a viable test to conclude the exactness of the computed upper bound. However, these SDPs are intractable for practical FNNs with hundreds of ReLUs. To address this issue, we further propose a method to construct a reduced order model whose input-output property is identical to the original FNN over a neighborhood of the target input. We finally illustrate the effectiveness of the model reduction and exactness verification methods with numerical examples of practical FNNs.
Authors: Franz Nowak, Anej Svete, Li Du, Ryan Cotterell
Abstract: This work investigates the computational expressivity of language models (LMs) based on recurrent neural networks (RNNs). Siegelmann and Sontag (1992) famously showed that RNNs with rational weights and hidden states and unbounded computation time are Turing complete. However, LMs define weightings over strings in addition to just (unweighted) language membership and the analysis of the computational power of RNN LMs (RLMs) should reflect this. We extend the Turing completeness result to the probabilistic case, showing how a rationally weighted RLM with unbounded computation time can simulate any deterministic probabilistic Turing machine (PTM) with rationally weighted transitions. Since, in practice, RLMs work in real-time, processing a symbol at every time step, we treat the above result as an upper bound on the expressivity of RLMs. We also provide a lower bound by showing that under the restriction to real-time computation, such models can simulate deterministic real-time rational PTMs.
Authors: Sam Dillavou, Benjamin D Beyer, Menachem Stern, Andrea J Liu, Marc Z Miskin, Douglas J Durian
Abstract: Standard deep learning algorithms require differentiating large nonlinear networks, a process that is slow and power-hungry. Electronic learning metamaterials offer potentially fast, efficient, and fault-tolerant hardware for analog machine learning, but existing implementations are linear, severely limiting their capabilities. These systems differ significantly from artificial neural networks as well as the brain, so the feasibility and utility of incorporating nonlinear elements have not been explored. Here we introduce a nonlinear learning metamaterial -- an analog electronic network made of self-adjusting nonlinear resistive elements based on transistors. We demonstrate that the system learns tasks unachievable in linear systems, including XOR and nonlinear regression, without a computer. We find our nonlinear learning metamaterial reduces modes of training error in order (mean, slope, curvature), similar to spectral bias in artificial neural networks. The circuitry is robust to damage, retrainable in seconds, and performs learned tasks in microseconds while dissipating only picojoules of energy across each transistor. This suggests enormous potential for fast, low-power computing in edge systems like sensors, robotic controllers, and medical devices, as well as manufacturability at scale for performing and studying emergent learning.
Authors: Gianluigi Lopardo, Frederic Precioso, Damien Garreau
Abstract: Interpretability is essential for machine learning models to be trusted and deployed in critical domains. However, existing methods for interpreting text models are often complex, lack mathematical foundations, and their performance is not guaranteed. In this paper, we propose FRED (Faithful and Robust Explainer for textual Documents), a novel method for interpreting predictions over text. FRED offers three key insights to explain a model prediction: (1) it identifies the minimal set of words in a document whose removal has the strongest influence on the prediction, (2) it assigns an importance score to each token, reflecting its influence on the model's output, and (3) it provides counterfactual explanations by generating examples similar to the original document, but leading to a different prediction. We establish the reliability of FRED through formal definitions and theoretical analyses on interpretable classifiers. Additionally, our empirical evaluation against state-of-the-art methods demonstrates the effectiveness of FRED in providing insights into text models.
Authors: Romuald A. Janik
Abstract: Large Language Models (LLMs) are huge artificial neural networks which primarily serve to generate text, but also provide a very sophisticated probabilistic model of language use. Since generating a semantically consistent text requires a form of effective memory, we investigate the memory properties of LLMs and find surprising similarities with key characteristics of human memory. We argue that the human-like memory properties of the Large Language Model do not follow automatically from the LLM architecture but are rather learned from the statistics of the training textual data. These results strongly suggest that the biological features of human memory leave an imprint on the way that we structure our textual narratives.
Authors: Yilin Wang, Xinyi Hu, Matthew R. Gormley
Abstract: Most pretrained language models rely on subword tokenization, which processes text as a sequence of subword tokens. However, different granularities of text, such as characters, subwords, and words, can contain different kinds of information. Previous studies have shown that incorporating multiple input granularities improves model generalization, yet very few of them outputs useful representations for each granularity. In this paper, we introduce the entanglement model, aiming to combine character and subword language models. Inspired by vision-language models, our model treats characters and subwords as separate modalities, and it generates mutually informed representations for both granularities as output. We evaluate our model on text classification, named entity recognition, POS-tagging, and character-level sequence labeling (intraword code-switching). Notably, the entanglement model outperforms its backbone language models, particularly in the presence of noisy texts and low-resource languages. Furthermore, the entanglement model even outperforms larger pre-trained models on all English sequence labeling tasks and classification tasks. We make our code publically available.
Authors: Ehsan Asali, Prashant Doshi, Jin Sun
Abstract: The learn-from-observation (LfO) paradigm is a human-inspired mode for a robot to learn to perform a task simply by watching it being performed. LfO can facilitate robot integration on factory floors by minimizing disruption and reducing tedious programming. A key component of the LfO pipeline is a transformation of the depth camera frames to the corresponding task state and action pairs, which are then relayed to learning techniques such as imitation or inverse reinforcement learning for understanding the task parameters. While several existing computer vision models analyze videos for activity recognition, SA-Net specifically targets robotic LfO from RGB-D data. However, SA-Net and many other models analyze frame data captured from a single viewpoint. Their analysis is therefore highly sensitive to occlusions of the observed task, which are frequent in deployments. An obvious way of reducing occlusions is to simultaneously observe the task from multiple viewpoints and synchronously fuse the multiple streams in the model. Toward this, we present multi-view SA-Net, which generalizes the SA-Net model to allow the perception of multiple viewpoints of the task activity, integrate them, and better recognize the state and action in each frame. Performance evaluations on two distinct domains establish that MVSA-Net recognizes the state-action pairs under occlusion more accurately compared to single-view MVSA-Net and other baselines. Our ablation studies further evaluate its performance under different ambient conditions and establish the contribution of the architecture components. As such, MVSA-Net offers a significantly more robust and deployable state-action trajectory generation compared to previous methods.
Authors: Kecen Li, Chen Gong, Zhixiang Li, Yuzhong Zhao, Xinwen Hou, Tianhao Wang
Abstract: Differential Privacy (DP) image data synthesis, which leverages the DP technique to generate synthetic data to replace the sensitive data, allowing organizations to share and utilize synthetic images without privacy concerns. Previous methods incorporate the advanced techniques of generative models and pre-training on a public dataset to produce exceptional DP image data, but suffer from problems of unstable training and massive computational resource demands. This paper proposes a novel DP image synthesis method, termed PRIVIMAGE, which meticulously selects pre-training data, promoting the efficient creation of DP datasets with high fidelity and utility. PRIVIMAGE first establishes a semantic query function using a public dataset. Then, this function assists in querying the semantic distribution of the sensitive dataset, facilitating the selection of data from the public dataset with analogous semantics for pre-training. Finally, we pre-train an image generative model using the selected data and then fine-tune this model on the sensitive dataset using Differentially Private Stochastic Gradient Descent (DP-SGD). PRIVIMAGE allows us to train a lightly parameterized generative model, reducing the noise in the gradient during DP-SGD training and enhancing training stability. Extensive experiments demonstrate that PRIVIMAGE uses only 1% of the public dataset for pre-training and 7.6% of the parameters in the generative model compared to the state-of-the-art method, whereas achieves superior synthetic performance and conserves more computational resources. On average, PRIVIMAGE achieves 30.1% lower FID and 12.6% higher Classification Accuracy than the state-of-the-art method. The replication package and datasets can be accessed online.
Authors: Ayush K. Rai, Tarun Krishna, Feiyan Hu, Alexandru Drimbarean, Kevin McGuinness, Alan F. Smeaton, Noel E. O'Connor
Abstract: Video Anomaly Detection (VAD) is an open-set recognition task, which is usually formulated as a one-class classification (OCC) problem, where training data is comprised of videos with normal instances while test data contains both normal and anomalous instances. Recent works have investigated the creation of pseudo-anomalies (PAs) using only the normal data and making strong assumptions about real-world anomalies with regards to abnormality of objects and speed of motion to inject prior information about anomalies in an autoencoder (AE) based reconstruction model during training. This work proposes a novel method for generating generic spatio-temporal PAs by inpainting a masked out region of an image using a pre-trained Latent Diffusion Model and further perturbing the optical flow using mixup to emulate spatio-temporal distortions in the data. In addition, we present a simple unified framework to detect real-world anomalies under the OCC setting by learning three types of anomaly indicators, namely reconstruction quality, temporal irregularity and semantic inconsistency. Extensive experiments on four VAD benchmark datasets namely Ped2, Avenue, ShanghaiTech and UBnormal demonstrate that our method performs on par with other existing state-of-the-art PAs generation and reconstruction based methods under the OCC setting. Our analysis also examines the transferability and generalisation of PAs across these datasets, offering valuable insights by identifying real-world anomalies through PAs.
Authors: H. C. Donker, D. Neijzen, J. de Jong, G. A. Lunter
Abstract: Healthcare data from patient or population cohorts are often characterized by sparsity, high missingness and relatively small sample sizes. In addition, being able to quantify uncertainty is often important in a medical context. To address these analytical requirements we propose a deep generative Bayesian model for multinomial count data. We develop a collapsed Gibbs sampling procedure that takes advantage of a series of augmentation relations, inspired by the Zhou$\unicode{x2013}$Cong$\unicode{x2013}$Chen model. We visualise the model's ability to identify coherent substructures in the data using a dataset of handwritten digits. We then apply it to a large experimental dataset of DNA mutations in cancer and show that we can identify biologically meaningful clusters of mutational signatures in a fully data-driven way.
Authors: Lorenzo Bianchi, Fabio Carrara, Nicola Messina, Claudio Gennaro, Fabrizio Falchi
Abstract: Recent advancements in large vision-language models enabled visual object detection in open-vocabulary scenarios, where object classes are defined in free-text formats during inference. In this paper, we aim to probe the state-of-the-art methods for open-vocabulary object detection to determine to what extent they understand fine-grained properties of objects and their parts. To this end, we introduce an evaluation protocol based on dynamic vocabulary generation to test whether models detect, discern, and assign the correct fine-grained description to objects in the presence of hard-negative classes. We contribute with a benchmark suite of increasing difficulty and probing different properties like color, pattern, and material. We further enhance our investigation by evaluating several state-of-the-art open-vocabulary object detectors using the proposed protocol and find that most existing solutions, which shine in standard open-vocabulary benchmarks, struggle to accurately capture and distinguish finer object details. We conclude the paper by highlighting the limitations of current methodologies and exploring promising research directions to overcome the discovered drawbacks. Data and code are available at https://lorebianchi98.github.io/FG-OVD/.
Authors: Anaelia Ovalle, Ninareh Mehrabi, Palash Goyal, Jwala Dhamala, Kai-Wei Chang, Richard Zemel, Aram Galstyan, Yuval Pinter, Rahul Gupta
Abstract: Gender-inclusive NLP research has documented the harmful limitations of gender binary-centric large language models (LLM), such as the inability to correctly use gender-diverse English neopronouns (e.g., xe, zir, fae). While data scarcity is a known culprit, the precise mechanisms through which scarcity affects this behavior remain underexplored. We discover LLM misgendering is significantly influenced by Byte-Pair Encoding (BPE) tokenization, the tokenizer powering many popular LLMs. Unlike binary pronouns, BPE overfragments neopronouns, a direct consequence of data scarcity during tokenizer training. This disparate tokenization mirrors tokenizer limitations observed in multilingual and low-resource NLP, unlocking new misgendering mitigation strategies. We propose two techniques: (1) pronoun tokenization parity, a method to enforce consistent tokenization across gendered pronouns, and (2) utilizing pre-existing LLM pronoun knowledge to improve neopronoun proficiency. Our proposed methods outperform finetuning with standard BPE, improving neopronoun accuracy from 14.1% to 58.4%. Our paper is the first to link LLM misgendering to tokenization and deficient neopronoun grammar, indicating that LLMs unable to correctly treat neopronouns as pronouns are more prone to misgender.
Authors: Ziqian Zeng, Yihuai Hong, Hongliang Dai, Huiping Zhuang, Cen Chen
Abstract: Early Exiting is one of the most popular methods to achieve efficient inference. Current early exiting methods adopt the (weighted) sum of the cross entropy loss of all internal classifiers during training, imposing all these classifiers to predict all instances correctly. However, during inference, as long as one internal classifier predicts an instance correctly, it can accelerate without losing accuracy. Thus, there is a notable gap between training and inference. We propose ConsistentEE, an early exiting method that is consistent in training and inference. ConsistentEE formulates the early exiting process as a reinforcement learning problem. A policy network is added to decide whether an instance should exit or continue. The training objective of ConsistentEE only require each instance to be predicted correctly by one internal classifier. Additionally, we introduce the concept Memorize Layer to measure the hardness of an instance. We incorporate memorized layer into reward function design, which allows "easy" instances to focus more on acceleration while "hard" instances to focus more on accuracy. Experimental results show that our method outperforms other baselines on various natural language understanding and generation tasks.
Authors: Kun Chen, Lei Bai, Fenghua Ling, Peng Ye, Tao Chen, Jing-Jia Luo, Hao Chen, Yi Xiao, Kang Chen, Tao Han, Wanli Ouyang
Abstract: The weather forecasting system is important for science and society, and significant achievements have been made in applying artificial intelligence (AI) to medium-range weather forecasting. However, existing AI-based weather forecasting models rely on analysis or reanalysis products from traditional numerical weather prediction (NWP) systems as initial conditions for making predictions. Initial states are typically generated by traditional data assimilation components, which are computational expensive and time-consuming. Here we present an AI-based data assimilation model, i.e., Adas, for global weather variables. By introducing the confidence matrix, Adas employs gated convolution to handle sparse observations and gated cross-attention for capturing the interactions between the background and observations. Further, we combine Adas with the advanced AI-based forecasting model (i.e., FengWu) to construct the first end-to-end AI-based global weather forecasting system: FengWu-Adas. We demonstrate that Adas can assimilate global observations to produce high-quality analysis, enabling the system operate stably for long term. Moreover, we are the first to apply the methods to real-world scenarios, which is more challenging and has considerable practical application potential. We have also achieved the forecasts based on the analyses generated by AI with a skillful forecast lead time exceeding that of the IFS for the first time.
Authors: Mohamed Elmahallawy, Tie Luo
Abstract: In the ambitious realm of space AI, the integration of federated learning (FL) with low Earth orbit (LEO) satellite constellations holds immense promise. However, many challenges persist in terms of feasibility, learning efficiency, and convergence. These hurdles stem from the bottleneck in communication, characterized by sporadic and irregular connectivity between LEO satellites and ground stations, coupled with the limited computation capability of satellite edge computing (SEC). This paper proposes a novel FL-SEC framework that empowers LEO satellites to execute large-scale machine learning (ML) tasks onboard efficiently. Its key components include i) personalized learning via divide-and-conquer, which identifies and eliminates redundant satellite images and converts complex multi-class classification problems to simple binary classification, enabling rapid and energy-efficient training of lightweight ML models suitable for IoT/edge devices on satellites; ii) orbital model retraining, which generates an aggregated "orbital model" per orbit and retrains it before sending to the ground station, significantly reducing the required communication rounds. We conducted experiments using Jetson Nano, an edge device closely mimicking the limited compute on LEO satellites, and a real satellite dataset. The results underscore the effectiveness of our approach, highlighting SEC's ability to run lightweight ML models on real and high-resolution satellite imagery. Our approach dramatically reduces FL convergence time by nearly 30 times, and satellite energy consumption down to as low as 1.38 watts, all while maintaining an exceptional accuracy of up to 96%.
Authors: Jian Gu, Aldeida Aleti, Chunyang Chen, Hongyu Zhang
Abstract: Understanding the latent space of language models (LM) is crucial to refining their performance and interpretability. Existing analyses often fall short in providing disentangled (model-centric) insights into LM semantics, and neglect essential aspects of LM adaptation. In response, we introduce a pioneering method called vocabulary-defined semantics, which establishes a reference frame within the LM latent space, ensuring disentangled semantic analysis grounded in LM vocabulary. Our approach transcends prior entangled analysis, leveraging LM vocabulary for model-centric insights. Furthermore, we propose a novel technique to compute logits, emphasising differentiability and local isotropy, and introduce a neural clustering module for semantically calibrating data representations during LM adaptation. Through extensive experiments across diverse text understanding datasets, our approach outperforms state-of-the-art methods of retrieval-augmented generation and parameter-efficient finetuning, showcasing its efficacy and broad applicability. Our findings not only shed light on LM mechanics, but also offer practical solutions to enhance LM performance and interpretability.
Authors: Jinghong Li, Huy Phan, Wen Gu, Koichi Ota, Shinobu Hasegawa
Abstract: Research surveys have always posed a challenge for beginner researchers who lack of research training. These researchers struggle to understand the directions within their research topic, and the discovery of new research findings within a short time. One way to provide intuitive assistance to beginner researchers is by offering relevant knowledge graphs(KG) and recommending related academic papers. However, existing navigation knowledge graphs primarily rely on keywords in the research field and often fail to present the logical hierarchy among multiple related papers clearly. Moreover, most recommendation systems for academic papers simply rely on high text similarity, which can leave researchers confused as to why a particular article is being recommended. They may lack of grasp important information about the insight connection between "Issue resolved" and "Issue finding" that they hope to obtain. To address these issues, this study aims to support research insight surveys for beginner researchers by establishing a hierarchical tree-structured knowledge graph that reflects the inheritance insight of research topics and the relevance insight among the academic papers.
Authors: Yizhou Zhang, Lun Du, Defu Cao, Qiang Fu, Yan Liu
Abstract: Foundation models, such as Large language Models (LLMs), have attracted significant amount of interest due to their large number of applications. Existing works show that appropriate prompt design, such as Chain-of-Thoughts, can unlock LLM's powerful capacity in diverse areas. However, when handling tasks involving repetitive sub-tasks and/or deceptive contents, such as arithmetic calculation and article-level fake news detection, existing prompting strategies either suffers from insufficient expressive power or intermediate errors triggered by hallucination. To make LLM more discerning to such intermediate errors, we propose to guide LLM with a Divide-and-Conquer program that simultaneously ensures superior expressive power and disentangles task decomposition, sub-task resolution, and resolution assembly process. Theoretic analysis reveals that our strategy can guide LLM to extend the expressive power of fixed-depth Transformer. Experiments indicate that our proposed method can achieve better performance than typical prompting strategies in tasks bothered by intermediate errors and deceptive contents, such as large integer multiplication, hallucination detection and misinformation detection.
Authors: Yu Liu, Ibrahim Al-Nahhal, Octavia A. Dobre, Fanggang Wang
Abstract: Integrated sensing and communication (ISAC) and intelligent reflecting surface (IRS) are viewed as promising technologies for future generations of wireless networks. This paper investigates the channel estimation problem in an IRS-assisted ISAC system. A deep-learning framework is proposed to estimate the sensing and communication (S&C) channels in such a system. Considering different propagation environments of the S&C channels, two deep neural network (DNN) architectures are designed to realize this framework. The first DNN is devised at the ISAC base station for estimating the sensing channel, while the second DNN architecture is assigned to each downlink user equipment to estimate its communication channel. Moreover, the input-output pairs to train the DNNs are carefully designed. Simulation results show the superiority of the proposed estimation approach compared to the benchmark scheme under various signal-to-noise ratio conditions and system parameters.
Authors: Yu Liu, Ibrahim Al-Nahhal, Octavia A. Dobre, Fanggang Wang, Hyundong Shin
Abstract: Multi-user integrated sensing and communication (ISAC) assisted by intelligent reflecting surface (IRS) has been recently investigated to provide a high spectral and energy efficiency transmission. This paper proposes a practical channel estimation approach for the first time to an IRS-assisted multiuser ISAC system. The estimation problem in such a system is challenging since the sensing and communication (SAC) signals interfere with each other, and the passive IRS lacks signal processing ability. A two-stage approach is proposed to transfer the overall estimation problem into sub-ones, successively including the direct and reflected channels estimation. Based on this scheme, the ISAC base station (BS) estimates all the SAC channels associated with the target and uplink users, while each downlink user estimates the downlink communication channels individually. Considering a low-cost demand of the ISAC BS and downlink users, the proposed two-stage approach is realized by an efficient neural network (NN) framework that contains two different extreme learning machine (ELM) structures to estimate the above SAC channels. Moreover, two types of input-output pairs to train the ELMs are carefully devised, which impact the estimation accuracy and computational complexity under different system parameters. Simulation results reveal a substantial performance improvement achieved by the proposed ELM-based approach over the least-squares and NN-based benchmarks, with reduced training complexity and faster training speed.
Authors: Yu Liu, Ibrahim Al-Nahhal, Octavia A. Dobre, Fanggang Wang
Abstract: Integrated sensing and communication (ISAC), and intelligent reflecting surface (IRS) are envisioned as revolutionary technologies to enhance spectral and energy efficiencies for next wireless system generations. For the first time, this paper focuses on the channel estimation problem in an IRS-assisted ISAC system. This problem is challenging due to the lack of signal processing capacity in passive IRS, as well as the presence of mutual interference between sensing and communication (SAC) signals in ISAC systems. A three-stage approach is proposed to decouple the estimation problem into sub-ones, including the estimation of the direct SAC channels in the first stage, reflected communication channel in the second stage, and reflected sensing channel in the third stage. The proposed three-stage approach is based on a deep-learning framework, which involves two different convolutional neural network (CNN) architectures to estimate the channels at the full-duplex ISAC base station. Furthermore, two types of input-output pairs to train the CNNs are carefully designed, which affect the estimation performance under various signal-to-noise ratio conditions and system parameters. Simulation results validate the superiority of the proposed estimation approach compared to the least-squares baseline scheme, and its computational complexity is also analyzed.
Authors: Hamza Mahdi, Eptehal Nashnoush, Rami Saab, Arjun Balachandar, Rishit Dagli, Lucas X. Perri, Houman Khosravani
Abstract: This study assesses deep learning models for audio classification in a clinical setting with the constraint of small datasets reflecting real-world prospective data collection. We analyze CNNs, including DenseNet and ConvNeXt, alongside transformer models like ViT, SWIN, and AST, and compare them against pre-trained audio models such as YAMNet and VGGish. Our method highlights the benefits of pre-training on large datasets before fine-tuning on specific clinical data. We prospectively collected two first-of-their-kind patient audio datasets from stroke patients. We investigated various preprocessing techniques, finding that RGB and grayscale spectrogram transformations affect model performance differently based on the priors they learn from pre-training. Our findings indicate CNNs can match or exceed transformer models in small dataset contexts, with DenseNet-Contrastive and AST models showing notable performance. This study highlights the significance of incremental marginal gains through model selection, pre-training, and preprocessing in sound classification; this offers valuable insights for clinical diagnostics that rely on audio classification.
Authors: Honghao Gui, Lin Yuan, Hongbin Ye, Ningyu Zhang, Mengshu Sun, Lei Liang, Huajun Chen
Abstract: Large Language Models (LLMs) demonstrate remarkable potential across various domains; however, they exhibit a significant performance gap in Information Extraction (IE). Note that high-quality instruction data is the vital key for enhancing the specific capabilities of LLMs, while current IE datasets tend to be small in scale, fragmented, and lack standardized schema. To this end, we introduce IEPile, a comprehensive bilingual (English and Chinese) IE instruction corpus, which contains approximately 0.32B tokens. We construct IEPile by collecting and cleaning 33 existing IE datasets, and introduce schema-based instruction generation to unearth a large-scale corpus. Experimental results on LLaMA, Baichuan and Qwen demonstrate that using IEPile can enhance the performance of LLMs for IE, especially the zero-shot generalization. We open-source the resource and pre-trained models, hoping to provide valuable support to the NLP community.
Authors: Donghwan Lee
Abstract: This paper analyzes multi-step TD-learning algorithms within the `deadly triad' scenario, characterized by linear function approximation, off-policy learning, and bootstrapping. In particular, we prove that n-step TD-learning algorithms converge to a solution as the sampling horizon n increases sufficiently. The paper is divided into two parts. In the first part, we comprehensively examine the fundamental properties of their model-based deterministic counterparts, including projected value iteration, gradient descent algorithms, and the control theoretic approach, which can be viewed as prototype deterministic algorithms whose analysis plays a pivotal role in understanding and developing their model-free reinforcement learning counterparts. In particular, we prove that these algorithms converge to meaningful solutions when n is sufficiently large. Based on these findings, two n-step TD-learning algorithms are proposed and analyzed, which can be seen as the model-free reinforcement learning counterparts of the gradient and control theoretic algorithms.
Authors: Shiming Zhang, Zhipeng Luo, Li Yang, Fei Teng, Tianrui Li
Abstract: Nowadays, with advanced information technologies deployed citywide, large data volumes and powerful computational resources are intelligentizing modern city development. As an important part of intelligent transportation, route recommendation and its applications are widely used, directly influencing citizens` travel habits. Developing smart and efficient travel routes based on big data (possibly multi-modal) has become a central challenge in route recommendation research. Our survey offers a comprehensive review of route recommendation work based on urban computing. It is organized by the following three parts: 1) Methodology-wise. We categorize a large volume of traditional machine learning and modern deep learning methods. Also, we discuss their historical relations and reveal the edge-cutting progress. 2) Application\-wise. We present numerous novel applications related to route commendation within urban computing scenarios. 3) We discuss current problems and challenges and envision several promising research directions. We believe that this survey can help relevant researchers quickly familiarize themselves with the current state of route recommendation research and then direct them to future research trends.
Authors: Yanjie Ze, Gu Zhang, Kangning Zhang, Chenyuan Hu, Muhan Wang, Huazhe Xu
Abstract: Imitation learning provides an efficient way to teach robots dexterous skills; however, learning complex skills robustly and generalizablely usually consumes large amounts of human demonstrations. To tackle this challenging problem, we present 3D Diffusion Policy (DP3), a novel visual imitation learning approach that incorporates the power of 3D visual representations into diffusion policies, a class of conditional action generative models. The core design of DP3 is the utilization of a compact 3D visual representation, extracted from sparse point clouds with an efficient point encoder. In our experiments involving 72 simulation tasks, DP3 successfully handles most tasks with just 10 demonstrations and surpasses baselines with a 24.2% relative improvement. In 4 real robot tasks, DP3 demonstrates precise control with a high success rate of 85%, given only 40 demonstrations of each task, and shows excellent generalization abilities in diverse aspects, including space, viewpoint, appearance, and instance. Interestingly, in real robot experiments, DP3 rarely violates safety requirements, in contrast to baseline methods which frequently do, necessitating human intervention. Our extensive evaluation highlights the critical importance of 3D representations in real-world robot learning. Videos, code, and data are available on https://3d-diffusion-policy.github.io .
Authors: Qibing Ren, Chang Gao, Jing Shao, Junchi Yan, Xin Tan, Yu Qiao, Wai Lam, Lizhuang Ma
Abstract: The rapid advancement of Large Language Models (LLMs) has brought about remarkable generative capabilities but also raised concerns about their potential misuse. While strategies like supervised fine-tuning and reinforcement learning from human feedback have enhanced their safety, these methods primarily focus on natural languages, which may not generalize to other domains. This paper introduces CodeAttack, a framework that transforms natural language inputs into code inputs, presenting a novel environment for testing the safety generalization of LLMs. Our comprehensive studies on state-of-the-art LLMs including GPT-4, Claude-2, and Llama-2 series reveal a common safety vulnerability of these models against code input: CodeAttack bypasses the safety guardrails of all models more than 80% of the time. We find that a larger distribution gap between CodeAttack and natural language leads to weaker safety generalization, such as encoding natural language input with data structures. Furthermore, we give two hypotheses about the success of CodeAttack: (1) the misaligned bias acquired by LLMs during code training, prioritizing code completion over avoiding the potential safety risk; (2) the limited self-evaluation capability regarding the safety of their code outputs. Finally, we analyze potential mitigation measures. These findings highlight new safety risks in the code domain and the need for more robust safety alignment algorithms to match the code capabilities of LLMs.
Authors: Shadi Iskander, Kira Radinsky, Yonatan Belinkov
Abstract: Mitigating social biases typically requires identifying the social groups associated with each data sample. In this paper, we present DAFair, a novel approach to address social bias in language models. Unlike traditional methods that rely on explicit demographic labels, our approach does not require any such information. Instead, we leverage predefined prototypical demographic texts and incorporate a regularization term during the fine-tuning process to mitigate bias in the model's representations. Our empirical results across two tasks and two models demonstrate the effectiveness of our method compared to previous approaches that do not rely on labeled data. Moreover, with limited demographic-annotated data, our approach outperforms common debiasing approaches.
Authors: Jordan Morris
Abstract: This paper proposes a machine learning pre-sort stage to traditional supervised learning using Tsetlin Machines. Initially, K data-points are identified from the dataset using an expedited genetic algorithm to solve the maximum dispersion problem. These are then used as the initial placement to run the K-Medoid clustering algorithm. Finally, an expedited genetic algorithm is used to align K independent Tsetlin Machines by maximising hamming distance. For MNIST level classification problems, results demonstrate up to 10% improvement in accuracy, approx. 383X reduction in training time and approx. 86X reduction in inference time.
Authors: Yaxi Hu, Amartya Sanyal, Bernhard Sch\"olkopf
Abstract: When analysing Differentially Private (DP) machine learning pipelines, the potential privacy cost of data-dependent pre-processing is frequently overlooked in privacy accounting. In this work, we propose a general framework to evaluate the additional privacy cost incurred by non-private data-dependent pre-processing algorithms. Our framework establishes upper bounds on the overall privacy guarantees by utilising two new technical notions: a variant of DP termed Smooth DP and the bounded sensitivity of the pre-processing algorithms. In addition to the generic framework, we provide explicit overall privacy guarantees for multiple data-dependent pre-processing algorithms, such as data imputation, quantization, deduplication and PCA, when used in combination with several DP algorithms. Notably, this framework is also simple to implement, allowing direct integration into existing DP pipelines.
Authors: Zhenyuan Yuan, Siyuan Xu, Minghui Zhu
Abstract: This paper considers the problem of learning a control policy for robot motion planning with zero-shot generalization, i.e., no data collection and policy adaptation is needed when the learned policy is deployed in new environments. We develop a federated reinforcement learning framework that enables collaborative learning of multiple learners and a central server, i.e., the Cloud, without sharing their raw data. In each iteration, each learner uploads its local control policy and the corresponding estimated normalized arrival time to the Cloud, which then computes the global optimum among the learners and broadcasts the optimal policy to the learners. Each learner then selects between its local control policy and that from the Cloud for next iteration. The proposed framework leverages on the derived zero-shot generalization guarantees on arrival time and safety. Theoretical guarantees on almost-sure convergence, almost consensus, Pareto improvement and optimality gap are also provided. Monte Carlo simulation is conducted to evaluate the proposed framework.
Authors: Hasan Abed Al Kader Hammoud, Tuhin Das, Fabio Pizzati, Philip Torr, Adel Bibi, Bernard Ghanem
Abstract: We explore the impact of training with more diverse datasets, characterized by the number of unique samples, on the performance of self-supervised learning (SSL) under a fixed computational budget. Our findings consistently demonstrate that increasing pretraining data diversity enhances SSL performance, albeit only when the distribution distance to the downstream data is minimal. Notably, even with an exceptionally large pretraining data diversity achieved through methods like web crawling or diffusion-generated data, among other ways, the distribution shift remains a challenge. Our experiments are comprehensive with seven SSL methods using large-scale datasets such as ImageNet and YFCC100M amounting to over 200 GPU days. Code and trained models will be available at https://github.com/hammoudhasan/DiversitySSL .
Authors: Narges Rashvand, Kenneth Witham, Gabriel Maldonado, Vinit Katariya, Nishanth Marer Prabhu, Gunar Schirner, Hamed Tabkhi
Abstract: Automatic modulation recognition (AMR) is vital for accurately identifying modulation types within incoming signals, a critical task for optimizing operations within edge devices in IoT ecosystems. This paper presents an innovative approach that leverages Transformer networks, initially designed for natural language processing, to address the challenges of efficient AMR. Our transformer network architecture is designed with the mindset of real-time edge computing on IoT devices. Four tokenization techniques are proposed and explored for creating proper embeddings of RF signals, specifically focusing on overcoming the limitations related to the model size often encountered in IoT scenarios. Extensive experiments reveal that our proposed method outperformed advanced deep learning techniques, achieving the highest recognition accuracy. Notably, our model achieves an accuracy of 65.75 on the RML2016 and 65.80 on the CSPB.ML.2018+ dataset.
Authors: Florian Chen, Felix Weitk\"amper, Sagar Malhotra
Abstract: We study the generalization behavior of Markov Logic Networks (MLNs) across relational structures of different sizes. Multiple works have noticed that MLNs learned on a given domain generalize poorly across domains of different sizes. This behavior emerges from a lack of internal consistency within an MLN when used across different domain sizes. In this paper, we quantify this inconsistency and bound it in terms of the variance of the MLN parameters. The parameter variance also bounds the KL divergence between an MLN's marginal distributions taken from different domain sizes. We use these bounds to show that maximizing the data log-likelihood while simultaneously minimizing the parameter variance corresponds to two natural notions of generalization across domain sizes. Our theoretical results apply to Exponential Random Graphs and other Markov network based relational models. Finally, we observe that solutions known to decrease the variance of the MLN parameters, like regularization and Domain-Size Aware MLNs, increase the internal consistency of the MLNs. We empirically verify our results on four different datasets, with different methods to control parameter variance, showing that controlling parameter variance leads to better generalization.
Authors: Ke Guo, Zhenwei Miao, Wei Jing, Weiwei Liu, Weizi Li, Dayang Hao, Jia Pan
Abstract: Microscopic traffic simulation plays a crucial role in transportation engineering by providing insights into individual vehicle behavior and overall traffic flow. However, creating a realistic simulator that accurately replicates human driving behaviors in various traffic conditions presents significant challenges. Traditional simulators relying on heuristic models often fail to deliver accurate simulations due to the complexity of real-world traffic environments. Due to the covariate shift issue, existing imitation learning-based simulators often fail to generate stable long-term simulations. In this paper, we propose a novel approach called learner-aware supervised imitation learning to address the covariate shift problem in multi-agent imitation learning. By leveraging a variational autoencoder simultaneously modeling the expert and learner state distribution, our approach augments expert states such that the augmented state is aware of learner state distribution. Our method, applied to urban traffic simulation, demonstrates significant improvements over existing state-of-the-art baselines in both short-term microscopic and long-term macroscopic realism when evaluated on the real-world dataset pNEUMA.
Authors: Robik Shrestha, Yang Zou, Qiuyu Chen, Zhiheng Li, Yusheng Xie, Siqi Deng
Abstract: Existing text-to-image generative models reflect or even amplify societal biases ingrained in their training data. This is especially concerning for human image generation where models are biased against certain demographic groups. Existing attempts to rectify this issue are hindered by the inherent limitations of the pre-trained models and fail to substantially improve demographic diversity. In this work, we introduce Fair Retrieval Augmented Generation (FairRAG), a novel framework that conditions pre-trained generative models on reference images retrieved from an external image database to improve fairness in human generation. FairRAG enables conditioning through a lightweight linear module that projects reference images into the textual space. To enhance fairness, FairRAG applies simple-yet-effective debiasing strategies, providing images from diverse demographic groups during the generative process. Extensive experiments demonstrate that FairRAG outperforms existing methods in terms of demographic diversity, image-text alignment, and image fidelity while incurring minimal computational overhead during inference.
Authors: Kaiyuan Gao, Qizhi Pei, Jinhua Zhu, Kun He, Lijun Wu
Abstract: Molecular docking is a pivotal process in drug discovery. While traditional techniques rely on extensive sampling and simulation governed by physical principles, these methods are often slow and costly. The advent of deep learning-based approaches has shown significant promise, offering increases in both accuracy and efficiency. Building upon the foundational work of FABind, a model designed with a focus on speed and accuracy, we present FABind+, an enhanced iteration that largely boosts the performance of its predecessor. We identify pocket prediction as a critical bottleneck in molecular docking and propose a novel methodology that significantly refines pocket prediction, thereby streamlining the docking process. Furthermore, we introduce modifications to the docking module to enhance its pose generation capabilities. In an effort to bridge the gap with conventional sampling/generative methods, we incorporate a simple yet effective sampling technique coupled with a confidence model, requiring only minor adjustments to the regression framework of FABind. Experimental results and analysis reveal that FABind+ remarkably outperforms the original FABind, achieves competitive state-of-the-art performance, and delivers insightful modeling strategies. This demonstrates FABind+ represents a substantial step forward in molecular docking and drug discovery. Our code is in https://github.com/QizhiPei/FABind.
Authors: Orson Mengara
Abstract: Audio-based machine learning systems frequently use public or third-party data, which might be inaccurate. This exposes deep neural network (DNN) models trained on such data to potential data poisoning attacks. In this type of assault, attackers can train the DNN model using poisoned data, potentially degrading its performance. Another type of data poisoning attack that is extremely relevant to our investigation is label flipping, in which the attacker manipulates the labels for a subset of data. It has been demonstrated that these assaults may drastically reduce system performance, even for attackers with minimal abilities. In this study, we propose a backdoor attack named 'DirtyFlipping', which uses dirty label techniques, "label-on-label", to input triggers (clapping) in the selected data patterns associated with the target class, thereby enabling a stealthy backdoor.
Authors: Jakub Piskorski, Micha{\l} Marci\'nczuk, Roman Yangarber
Abstract: This paper presents a corpus manually annotated with named entities for six Slavic languages - Bulgarian, Czech, Polish, Slovenian, Russian, and Ukrainian. This work is the result of a series of shared tasks, conducted in 2017-2023 as a part of the Workshops on Slavic Natural Language Processing. The corpus consists of 5 017 documents on seven topics. The documents are annotated with five classes of named entities. Each entity is described by a category, a lemma, and a unique cross-lingual identifier. We provide two train-tune dataset splits - single topic out and cross topics. For each split, we set benchmarks using a transformer-based neural network architecture with the pre-trained multilingual models - XLM-RoBERTa-large for named entity mention recognition and categorization, and mT5-large for named entity lemmatization and linking.
Authors: Eric Guiffo Kaigom
Abstract: Metarobotics aims to combine next generation wireless communication, multi-sense immersion, and collective intelligence to provide a pervasive, itinerant, and non-invasive access and interaction with distant robotized applications. Industry and society are expected to benefit from these functionalities. For instance, robot programmers will no longer travel worldwide to plan and test robot motions, even collaboratively. Instead, they will have a personalized access to robots and their environments from anywhere, thus spending more time with family and friends. Students enrolled in robotics courses will be taught under authentic industrial conditions in real-time. This paper describes objectives of Metarobotics in society, industry, and in-between. It identifies and surveys technologies likely to enable their completion and provides an architecture to put forward the interplay of key components of Metarobotics. Potentials for self-determination, self-efficacy, and work-life-flexibility in robotics-related applications in Society 5.0, Industry 4.0, and Industry 5.0 are outlined.
Authors: Ming Xu, Stephen Gould
Abstract: We propose a novel approach to the action segmentation task for long, untrimmed videos, based on solving an optimal transport problem. By encoding a temporal consistency prior into a Gromov-Wasserstein problem, we are able to decode a temporally consistent segmentation from a noisy affinity/matching cost matrix between video frames and action classes. Unlike previous approaches, our method does not require knowing the action order for a video to attain temporal consistency. Furthermore, our resulting (fused) Gromov-Wasserstein problem can be efficiently solved on GPUs using a few iterations of projected mirror descent. We demonstrate the effectiveness of our method in an unsupervised learning setting, where our method is used to generate pseudo-labels for self-training. We evaluate our segmentation approach and unsupervised learning pipeline on the Breakfast, 50-Salads, YouTube Instructions and Desktop Assembly datasets, yielding state-of-the-art results for the unsupervised video action segmentation task.
Authors: Mamadou Keita, Wassim Hamidouche, Hessen Bougueffa Eutamene, Abdenour Hadid, Abdelmalik Taleb-Ahmed
Abstract: Advancements in deep image synthesis techniques, such as generative adversarial networks (GANs) and diffusion models (DMs), have ushered in an era of generating highly realistic images. While this technological progress has captured significant interest, it has also raised concerns about the potential difficulty in distinguishing real images from their synthetic counterparts. This paper takes inspiration from the potent convergence capabilities between vision and language, coupled with the zero-shot nature of vision-language models (VLMs). We introduce an innovative method called Bi-LORA that leverages VLMs, combined with low-rank adaptation (LORA) tuning techniques, to enhance the precision of synthetic image detection for unseen model-generated images. The pivotal conceptual shift in our methodology revolves around reframing binary classification as an image captioning task, leveraging the distinctive capabilities of cutting-edge VLM, notably bootstrapping language image pre-training (BLIP2). Rigorous and comprehensive experiments are conducted to validate the effectiveness of our proposed approach, particularly in detecting unseen diffusion-generated images from unknown diffusion-based generative models during training, showcasing robustness to noise, and demonstrating generalization capabilities to GANs. The obtained results showcase an impressive average accuracy of 93.41% in synthetic image detection on unseen generation models. The code and models associated with this research can be publicly accessed at https://github.com/Mamadou-Keita/VLM-DETECT.
Authors: Daisuke Kikuta, Hiroki Ikeuchi, Kengo Tajiri, Yuta Toyama, Masaki Nakamura, Yuusuke Nakano
Abstract: As a telecom provider, our company has a critical mission to maintain telecom services even during power outages. To accomplish the mission, it is essential to maintain the power of the telecom base stations. Here we consider a solution where electric vehicles (EVs) directly supply power to base stations by traveling to their locations. The goal is to find EV routes that minimize both the total travel distance of all EVs and the number of downed base stations. In this paper, we formulate this routing problem as a new variant of the Electric Vehicle Routing Problem (EVRP) and propose a solver that combines a rule-based vehicle selector and a reinforcement learning (RL)-based node selector. The rule of the vehicle selector ensures the exact environmental states when the selected EV starts to move. In addition, the node selection by the RL model enables fast route generation, which is critical in emergencies. We evaluate our solver on both synthetic datasets and real datasets. The results show that our solver outperforms baselines in terms of the objective value and computation time. Moreover, we analyze the generalization and scalability of our solver, demonstrating the capability toward unseen settings and large-scale problems. Check also our project page: https://ntt-dkiku.github.io/rl-evrpeps.
Authors: Ye Yuan, Kexin Tang, Jianhao Shen, Ming Zhang, Chenguang Wang
Abstract: We present a new challenge to examine whether large language models understand social norms. In contrast to existing datasets, our dataset requires a fundamental understanding of social norms to solve. Our dataset features the largest set of social norm skills, consisting of 402 skills and 12,383 questions covering a wide set of social norms ranging from opinions and arguments to culture and laws. We design our dataset according to the K-12 curriculum. This enables the direct comparison of the social understanding of large language models to humans, more specifically, elementary students. While prior work generates nearly random accuracy on our benchmark, recent large language models such as GPT3.5-Turbo and LLaMA2-Chat are able to improve the performance significantly, only slightly below human performance. We then propose a multi-agent framework based on large language models to improve the models' ability to understand social norms. This method further improves large language models to be on par with humans. Given the increasing adoption of large language models in real-world applications, our finding is particularly important and presents a unique direction for future improvements.
Authors: Zhongming Yu, Genghan Zhang, Hanxian Huang, Xin Chen, Jishen Zhao
Abstract: In recent years, Graph Neural Networks (GNNs) have ignited a surge of innovation, significantly enhancing the processing of geometric data structures such as graphs, point clouds, and meshes. As the domain continues to evolve, a series of frameworks and libraries are being developed to push GNN efficiency to new heights. While graph-centric libraries have achieved success in the past, the advent of efficient tensor compilers has highlighted the urgent need for tensor-centric libraries. Yet, efficient tensor-centric frameworks for GNNs remain scarce due to unique challenges and limitations encountered when implementing segment reduction in GNN contexts. We introduce GeoT, a cutting-edge tensor-centric library designed specifically for GNNs via efficient segment reduction. GeoT debuts innovative parallel algorithms that not only introduce new design principles but also expand the available design space. Importantly, GeoT is engineered for straightforward fusion within a computation graph, ensuring compatibility with contemporary tensor-centric machine learning frameworks and compilers. Setting a new performance benchmark, GeoT marks a considerable advancement by showcasing an average operator speedup of 1.80x and an end-to-end speedup of 1.68x.
Authors: Detai Xin, Xu Tan, Kai Shen, Zeqian Ju, Dongchao Yang, Yuancheng Wang, Shinnosuke Takamichi, Hiroshi Saruwatari, Shujie Liu, Jinyu Li, Sheng Zhao
Abstract: We present RALL-E, a robust language modeling method for text-to-speech (TTS) synthesis. While previous work based on large language models (LLMs) shows impressive performance on zero-shot TTS, such methods often suffer from poor robustness, such as unstable prosody (weird pitch and rhythm/duration) and a high word error rate (WER), due to the autoregressive prediction style of language models. The core idea behind RALL-E is chain-of-thought (CoT) prompting, which decomposes the task into simpler steps to enhance the robustness of LLM-based TTS. To accomplish this idea, RALL-E first predicts prosody features (pitch and duration) of the input text and uses them as intermediate conditions to predict speech tokens in a CoT style. Second, RALL-E utilizes the predicted duration prompt to guide the computing of self-attention weights in Transformer to enforce the model to focus on the corresponding phonemes and prosody features when predicting speech tokens. Results of comprehensive objective and subjective evaluations demonstrate that, compared to a powerful baseline method VALL-E, RALL-E significantly improves the WER of zero-shot TTS from $6.3\%$ (without reranking) and $2.1\%$ (with reranking) to $2.8\%$ and $1.0\%$, respectively. Furthermore, we demonstrate that RALL-E correctly synthesizes sentences that are hard for VALL-E and reduces the error rate from $68\%$ to $4\%$.
Authors: Farnaz Kohankhaki, Jacob-Junqi Tian, David Emerson, Laleh Seyyed-Kalantari, Faiza Khan Khattak
Abstract: Large language models (LLMs), trained on vast datasets, can carry biases that manifest in various forms, from overt discrimination to implicit stereotypes. One facet of bias is performance disparities in LLMs, often harming underprivileged groups, such as racial minorities. A common approach to quantifying bias is to use template-based bias probes, which explicitly state group membership (e.g. White) and evaluate if the outcome of a task, sentiment analysis for instance, is invariant to the change of group membership (e.g. change White race to Black). This approach is widely used in bias quantification. However, in this work, we find evidence of an unexpectedly overlooked consequence of using template-based probes for LLM bias quantification. We find that in doing so, text examples associated with White ethnicities appear to be classified as exhibiting negative sentiment at elevated rates. We hypothesize that the scenario arises artificially through a mismatch between the pre-training text of LLMs and the templates used to measure bias through reporting bias, unstated norms that imply group membership without explicit statement. Our finding highlights the potential misleading impact of varying group membership through explicit mention in bias quantification
Authors: Jiawei Guo, Ziming Li, Xueling Liu, Kaijing Ma, Tianyu Zheng, Zhouliang Yu, Ding Pan, Yizhi LI, Ruibo Liu, Yue Wang, Shuyue Guo, Xingwei Qu, Xiang Yue, Ge Zhang, Wenhu Chen, Jie Fu
Abstract: Large Language Models (LLMs) for code are rapidly evolving, with code editing emerging as a critical capability. We introduce CodeEditorBench, an evaluation framework designed to rigorously assess the performance of LLMs in code editing tasks, including debugging, translating, polishing, and requirement switching. Unlike existing benchmarks focusing solely on code generation, CodeEditorBench emphasizes real-world scenarios and practical aspects of software development. We curate diverse coding challenges and scenarios from five sources, covering various programming languages, complexity levels, and editing tasks. Evaluation of 19 LLMs reveals that closed-source models (particularly Gemini-Ultra and GPT-4), outperform open-source models in CodeEditorBench, highlighting differences in model performance based on problem types and prompt sensitivities. CodeEditorBench aims to catalyze advancements in LLMs by providing a robust platform for assessing code editing capabilities. We will release all prompts and datasets to enable the community to expand the dataset and benchmark emerging LLMs. By introducing CodeEditorBench, we contribute to the advancement of LLMs in code editing and provide a valuable resource for researchers and practitioners.
Authors: Zhengxuan Wu, Aryaman Arora, Zheng Wang, Atticus Geiger, Dan Jurafsky, Christopher D. Manning, Christopher Potts
Abstract: Parameter-efficient fine-tuning (PEFT) methods seek to adapt large models via updates to a small number of weights. However, much prior interpretability work has shown that representations encode rich semantic information, suggesting that editing representations might be a more powerful alternative. Here, we pursue this hypothesis by developing a family of $\textbf{Representation Finetuning (ReFT)}$ methods. ReFT methods operate on a frozen base model and learn task-specific interventions on hidden representations. We define a strong instance of the ReFT family, Low-rank Linear Subspace ReFT (LoReFT). LoReFT is a drop-in replacement for existing PEFTs and learns interventions that are 10x-50x more parameter-efficient than prior state-of-the-art PEFTs. We showcase LoReFT on eight commonsense reasoning tasks, four arithmetic reasoning tasks, Alpaca-Eval v1.0, and GLUE. In all these evaluations, LoReFT delivers the best balance of efficiency and performance, and almost always outperforms state-of-the-art PEFTs. We release a generic ReFT training library publicly at https://github.com/stanfordnlp/pyreft.