Authors: Uwe Naumann
Abstract: The efficient computation of Jacobians represents a fundamental challenge in computational science and engineering. Large-scale modular numerical simulation programs can be regarded as sequences of evaluations of in our case differentiable subprograms with corresponding elemental Jacobians. The latter are typically not available. Tangent and adjoint versions of the individual subprograms are assumed to be given as results of algorithmic differentiation instead. The classical (Jacobian) Matrix Chain Product problem is reformulated in terms of matrix-free Jacobian-matrix (tangents) and matrix-Jacobian products (adjoints), subject to limited memory for storing information required by latter. All numerical results can be reproduced using an open-source reference implementation.
Authors: Adithya Vasudev
Abstract: The Lottery Ticket hypothesis proposes that ideal sparse subnetworks called lottery tickets exist in the untrained dense network. The Early Bird hypothesis proposes an efficient algorithm to find these winning lottery tickets in convolutional neural networks using the novel concept of distance between subnetworks to detect convergence in the subnetworks of a model. However, this approach overlooks unchanging groups of unimportant neurons near the end of the search. We propose WORM, a method that exploits these static groups by truncating their gradients, forcing the model to rely on other neurons. Experiments show WORM achieves faster ticket identification training and uses fewer FLOPs, despite the additional computational overhead. Additionally WORM pruned models lose less accuracy during pruning and recover accuracy faster, improving the robustness of the model. Furthermore, WORM is also able to generalize the Early Bird hypothesis reasonably well to larger models such as transformers, displaying its flexibility to adapt to various architectures.
Authors: Haoren Zhu, Pengfei Zhao, Wilfred Siu Hung NG, Dik Lun Lee
Abstract: Financial assets exhibit complex dependency structures, which are crucial for investors to create diversified portfolios to mitigate risk in volatile financial markets. To explore the financial asset dependencies dynamics, we propose a novel approach that models the dependencies of assets as an Asset Dependency Matrix (ADM) and treats the ADM sequences as image sequences. This allows us to leverage deep learning-based video prediction methods to capture the spatiotemporal dependencies among assets. However, unlike images where neighboring pixels exhibit explicit spatiotemporal dependencies due to the natural continuity of object movements, assets in ADM do not have a natural order. This poses challenges to organizing the relational assets to reveal better the spatiotemporal dependencies among neighboring assets for ADM forecasting. To tackle the challenges, we propose the Asset Dependency Neural Network (ADNN), which employs the Convolutional Long Short-Term Memory (ConvLSTM) network, a highly successful method for video prediction. ADNN can employ static and dynamic transformation functions to optimize the representations of the ADM. Through extensive experiments, we demonstrate that our proposed framework consistently outperforms the baselines in the ADM prediction and downstream application tasks. This research contributes to understanding and predicting asset dependencies, offering valuable insights for financial market participants.
Authors: Hui Liu, Wenya Wang, Hao Sun, Chris Xing Tian, Chenqi Kong, Xin Dong, Haoliang Li
Abstract: Large Language Models (LLMs) have demonstrated impressive in-context learning (ICL) capabilities from few-shot demonstration exemplars. While recent learning-based demonstration selection methods have proven beneficial to ICL by choosing more useful exemplars, their underlying mechanisms are opaque, hindering efforts to address limitations such as high training costs and poor generalization across tasks. These methods generally assume the selection process captures similarities between the exemplar and the target instance, however, it remains unknown what kinds of similarities are captured and vital to performing ICL. To dive into this question, we analyze the working mechanisms of the learning-based demonstration selection methods and empirically identify two important factors related to similarity measurement: 1) The ability to integrate different levels of task-agnostic text similarities between the input of exemplars and test cases enhances generalization power across different tasks. 2) Incorporating task-specific labels when measuring the similarities significantly improves the performance on each specific task. We validate these two findings through extensive quantitative and qualitative analyses across ten datasets and various LLMs. Based on our findings, we introduce two effective yet simplified exemplar selection methods catering to task-agnostic and task-specific demands, eliminating the costly LLM inference overhead.
Authors: Hao Bai, Yifei Zhou, Mert Cemri, Jiayi Pan, Alane Suhr, Sergey Levine, Aviral Kumar
Abstract: Training corpuses for vision language models (VLMs) typically lack sufficient amounts of decision-centric data. This renders off-the-shelf VLMs sub-optimal for decision-making tasks such as in-the-wild device control through graphical user interfaces (GUIs). While training with static demonstrations has shown some promise, we show that such methods fall short for controlling real GUIs due to their failure to deal with real-world stochasticity and non-stationarity not captured in static observational data. This paper introduces a novel autonomous RL approach, called DigiRL, for training in-the-wild device control agents through fine-tuning a pre-trained VLM in two stages: offline RL to initialize the model, followed by offline-to-online RL. To do this, we build a scalable and parallelizable Android learning environment equipped with a VLM-based evaluator and develop a simple yet effective RL approach for learning in this domain. Our approach runs advantage-weighted RL with advantage estimators enhanced to account for stochasticity along with an automatic curriculum for deriving maximal learning signal. We demonstrate the effectiveness of DigiRL using the Android-in-the-Wild (AitW) dataset, where our 1.3B VLM trained with RL achieves a 49.5% absolute improvement -- from 17.7 to 67.2% success rate -- over supervised fine-tuning with static human demonstration data. These results significantly surpass not only the prior best agents, including AppAgent with GPT-4V (8.3% success rate) and the 17B CogAgent trained with AitW data (38.5%), but also the prior best autonomous RL approach based on filtered behavior cloning (57.8%), thereby establishing a new state-of-the-art for digital agents for in-the-wild device control.
Authors: Taiqiang Wu, Jiahao Wang, Zhe Zhao, Ngai Wong
Abstract: In this paper, we introduce a subspace-inspired Low-Rank Adaptation (LoRA) method, which is computationally efficient, easy to implement, and readily applicable to large language, multimodal, and diffusion models. Initially, we equivalently decompose the weights of LoRA into two subspaces, and find that simply mixing them can enhance performance. To study such a phenomenon, we revisit it through a fine-grained subspace lens, showing that such modification is equivalent to employing a fixed mixer to fuse the subspaces. To be more flexible, we jointly learn the mixer with the original LoRA weights, and term the method Mixture-of-Subspaces LoRA (MoSLoRA). MoSLoRA consistently outperforms LoRA on tasks in different modalities, including commonsense reasoning, visual instruction tuning, and subject-driven text-to-image generation, demonstrating its effectiveness and robustness. Codes are available at \href{https://github.com/wutaiqiang/MoSLoRA}{github}.
Authors: Mengxi Liu, Daniel Gei{\ss}ler, Dominique Nshimyimana, Sizhen Bian, Bo Zhou, Paul Lukowicz
Abstract: In this work, we explore the use of a novel neural network architecture, the Kolmogorov-Arnold Networks (KANs) as feature extractors for sensor-based (specifically IMU) Human Activity Recognition (HAR). Where conventional networks perform a parameterized weighted sum of the inputs at each node and then feed the result into a statically defined nonlinearity, KANs perform non-linear computations represented by B-SPLINES on the edges leading to each node and then just sum up the inputs at the node. Instead of learning weights, the system learns the spline parameters. In the original work, such networks have been shown to be able to more efficiently and exactly learn sophisticated real valued functions e.g. in regression or PDE solution. We hypothesize that such an ability is also advantageous for computing low-level features for IMU-based HAR. To this end, we have implemented KAN as the feature extraction architecture for IMU-based human activity recognition tasks, including four architecture variations. We present an initial performance investigation of the KAN feature extractor on four public HAR datasets. It shows that the KAN-based feature extractor outperforms CNN-based extractors on all datasets while being more parameter efficient.
Authors: Chao He, Hongmei Shi, Ruixin Li, Jianbo Li, ZuJun Yu
Abstract: The service conditions of wheelset bearings has a direct impact on the safe operation of railway heavy haul freight trains as the key components. However, speed fluctuation of the trains and few fault samples are the two main problems that restrict the accuracy of bearing fault diagnosis. Therefore, a cross-machine transfer diagnosis (pyDSN) network coupled with interpretable modulated differentiable short-time Fourier transform (STFT) and physics-informed balanced spectrum quality metric is proposed to learn domain-invariant and discriminative features under time-varying speeds. Firstly, due to insufficiency in extracting extract frequency components of time-varying speed signals using fixed windows, a modulated differentiable STFT (MDSTFT) that is interpretable with STFT-informed theoretical support, is proposed to extract the robust time-frequency spectrum (TFS). During training process, multiple windows with different lengths dynamically change. Also, in addition to the classification metric and domain discrepancy metric, we creatively introduce a third kind of metric, referred to as the physics-informed metric, to enhance transferable TFS. A physics-informed balanced spectrum quality (BSQ) regularization loss is devised to guide an optimization direction for MDSTFT and model. With it, not only can model acquire high-quality TFS, but also a physics-restricted domain adaptation network can be also acquired, making it learn real-world physics knowledge, ultimately diminish the domain discrepancy across different datasets. The experiment is conducted in the scenario of migrating from the laboratory datasets to the freight train dataset, indicating that the hybrid-driven pyDSN outperforms existing methods and has practical value.
Authors: Pavel Rumiantsev, Mark Coates
Abstract: In terms of accuracy, Graph Neural Networks (GNNs) are the best architectural choice for the node classification task. Their drawback in real-world deployment is the latency that emerges from the neighbourhood processing operation. One solution to the latency issue is to perform knowledge distillation from a trained GNN to a Multi-Layer Perceptron (MLP), where the MLP processes only the features of the node being classified (and possibly some pre-computed structural information). However, the performance of such MLPs in both transductive and inductive settings remains inconsistent for existing knowledge distillation techniques. We propose to address the performance concerns by using a specially-designed student model instead of an MLP. Our model, named Routing-by-Memory (RbM), is a form of Mixture-of-Experts (MoE), with a design that enforces expert specialization. By encouraging each expert to specialize on a certain region on the hidden representation space, we demonstrate experimentally that it is possible to derive considerably more consistent performance across multiple datasets.
Authors: Xi Chen, Chuan Qin, Chuyu Fang, Chao Wang, Chen Zhu, Fuzhen Zhuang, Hengshu Zhu, Hui Xiong
Abstract: In a rapidly evolving job market, skill demand forecasting is crucial as it enables policymakers and businesses to anticipate and adapt to changes, ensuring that workforce skills align with market needs, thereby enhancing productivity and competitiveness. Additionally, by identifying emerging skill requirements, it directs individuals towards relevant training and education opportunities, promoting continuous self-learning and development. However, the absence of comprehensive datasets presents a significant challenge, impeding research and the advancement of this field. To bridge this gap, we present Job-SDF, a dataset designed to train and benchmark job-skill demand forecasting models. Based on 10.35 million public job advertisements collected from major online recruitment platforms in China between 2021 and 2023, this dataset encompasses monthly recruitment demand for 2,324 types of skills across 521 companies. Our dataset uniquely enables evaluating skill demand forecasting models at various granularities, including occupation, company, and regional levels. We benchmark a range of models on this dataset, evaluating their performance in standard scenarios, in predictions focused on lower value ranges, and in the presence of structural breaks, providing new insights for further research. Our code and dataset are publicly accessible via the https://github.com/Job-SDF/benchmark.
Authors: Jiaqi Lin, Qianqian Ren
Abstract: Traffic prediction is a challenging spatio-temporal forecasting problem that involves highly complex spatio-temporal correlations. This paper proposes a Multi-level Multi-view Augmented Spatio-temporal Transformer (LVSTformer) for traffic prediction. The model aims to capture spatial dependencies from three different levels: local geographic, global semantic, and pivotal nodes, along with long- and short-term temporal dependencies. Specifically, we design three spatial augmented views to delve into the spatial information from the perspectives of local, global, and pivotal nodes. By combining three spatial augmented views with three parallel spatial self-attention mechanisms, the model can comprehensively captures spatial dependencies at different levels. We design a gated temporal self-attention mechanism to effectively capture long- and short-term temporal dependencies. Furthermore, a spatio-temporal context broadcasting module is introduced between two spatio-temporal layers to ensure a well-distributed allocation of attention scores, alleviating overfitting and information loss, and enhancing the generalization ability and robustness of the model. A comprehensive set of experiments is conducted on six well-known traffic benchmarks, the experimental results demonstrate that LVSTformer achieves state-of-the-art performance compared to competing baselines, with the maximum improvement reaching up to 4.32%.
Authors: Muhao Xu, Zhenfeng Zhu, Youru Li, Shuai Zheng, Yawei Zhao, Kunlun He, Yao Zhao
Abstract: Multimodal electronic health record (EHR) data can offer a holistic assessment of a patient's health status, supporting various predictive healthcare tasks. Recently, several studies have embraced the multitask learning approach in the healthcare domain, exploiting the inherent correlations among clinical tasks to predict multiple outcomes simultaneously. However, existing methods necessitate samples to possess complete labels for all tasks, which places heavy demands on the data and restricts the flexibility of the model. Meanwhile, within a multitask framework with multimodal inputs, how to comprehensively consider the information disparity among modalities and among tasks still remains a challenging problem. To tackle these issues, a unified healthcare prediction model, also named by \textbf{FlexCare}, is proposed to flexibly accommodate incomplete multimodal inputs, promoting the adaption to multiple healthcare tasks. The proposed model breaks the conventional paradigm of parallel multitask prediction by decomposing it into a series of asynchronous single-task prediction. Specifically, a task-agnostic multimodal information extraction module is presented to capture decorrelated representations of diverse intra- and inter-modality patterns. Taking full account of the information disparities between different modalities and different tasks, we present a task-guided hierarchical multimodal fusion module that integrates the refined modality-level representations into an individual patient-level representation. Experimental results on multiple tasks from MIMIC-IV/MIMIC-CXR/MIMIC-NOTE datasets demonstrate the effectiveness of the proposed method. Additionally, further analysis underscores the feasibility and potential of employing such a multitask strategy in the healthcare domain. The source code is available at https://github.com/mhxu1998/FlexCare.
Authors: Victor Priser, Pascal Bianchi, Adil Salim
Abstract: Stein Variational Gradient Descent (SVGD) is a widely used sampling algorithm that has been successfully applied in several areas of Machine Learning. SVGD operates by iteratively moving a set of interacting particles (which represent the samples) to approximate the target distribution. Despite recent studies on the complexity of SVGD and its variants, their long-time asymptotic behavior (i.e., after numerous iterations ) is still not understood in the finite number of particles regime. We study the long-time asymptotic behavior of a noisy variant of SVGD. First, we establish that the limit set of noisy SVGD for large is well-defined. We then characterize this limit set, showing that it approaches the target distribution as increases. In particular, noisy SVGD provably avoids the variance collapse observed for SVGD. Our approach involves demonstrating that the trajectories of noisy SVGD closely resemble those described by a McKean-Vlasov process.
Authors: Rui Zhou, Chenyang Yuan, Frank Permenter, Yanxia Zhang, Nikos Arechiga, Matt Klenk, Faez Ahmed
Abstract: This study introduces a generative imputation model leveraging graph attention networks and tabular diffusion models for completing missing parametric data in engineering designs. This model functions as an AI design co-pilot, providing multiple design options for incomplete designs, which we demonstrate using the bicycle design CAD dataset. Through comparative evaluations, we demonstrate that our model significantly outperforms existing classical methods, such as MissForest, hotDeck, PPCA, and tabular generative method TabCSDI in both the accuracy and diversity of imputation options. Generative modeling also enables a broader exploration of design possibilities, thereby enhancing design decision-making by allowing engineers to explore a variety of design completions. The graph model combines GNNs with the structural information contained in assembly graphs, enabling the model to understand and predict the complex interdependencies between different design parameters. The graph model helps accurately capture and impute complex parametric interdependencies from an assembly graph, which is key for design problems. By learning from an existing dataset of designs, the imputation capability allows the model to act as an intelligent assistant that autocompletes CAD designs based on user-defined partial parametric design, effectively bridging the gap between ideation and realization. The proposed work provides a pathway to not only facilitate informed design decisions but also promote creative exploration in design.
Authors: Tianle Li, Wei-Lin Chiang, Evan Frick, Lisa Dunlap, Tianhao Wu, Banghua Zhu, Joseph E. Gonzalez, Ion Stoica
Abstract: The rapid evolution of language models has necessitated the development of more challenging benchmarks. Current static benchmarks often struggle to consistently distinguish between the capabilities of different models and fail to align with real-world user preferences. On the other hand, live crowd-sourced platforms like the Chatbot Arena collect a wide range of natural prompts and user feedback. However, these prompts vary in sophistication and the feedback cannot be applied offline to new models. In order to ensure that benchmarks keep up with the pace of LLM development, we address how one can evaluate benchmarks on their ability to confidently separate models and their alignment with human preference. Under these principles, we developed BenchBuilder, a living benchmark that filters high-quality prompts from live data sources to enable offline evaluation on fresh, challenging prompts. BenchBuilder identifies seven indicators of a high-quality prompt, such as the requirement for domain knowledge, and utilizes an LLM annotator to select a high-quality subset of prompts from various topic clusters. The LLM evaluation process employs an LLM judge to ensure a fully automated, high-quality, and constantly updating benchmark. We apply BenchBuilder on prompts from the Chatbot Arena to create Arena-Hard-Auto v0.1: 500 challenging user prompts from a wide range of tasks. Arena-Hard-Auto v0.1 offers 3x tighter confidence intervals than MT-Bench and achieves a state-of-the-art 89.1% agreement with human preference rankings, all at a cost of only $25 and without human labelers. The BenchBuilder pipeline enhances evaluation benchmarks and provides a valuable tool for developers, enabling them to extract high-quality benchmarks from extensive data with minimal effort.
Authors: Junwei You, Haotian Shi, Keshu Wu, Keke Long, Sicheng Fu, Sikai Chen, Bin Ran
Abstract: Vehicle trajectory prediction is crucial for advancing autonomous driving and advanced driver assistance systems (ADAS), enhancing road safety and traffic efficiency. While traditional methods have laid foundational work, modern deep learning techniques, particularly transformer-based models and generative approaches, have significantly improved prediction accuracy by capturing complex and non-linear patterns in vehicle motion and traffic interactions. However, these models often overlook the detailed car-following behaviors and inter-vehicle interactions essential for real-world driving scenarios. This study introduces a Cross-Attention Transformer Enhanced Conditional Diffusion Model (Crossfusor) specifically designed for car-following trajectory prediction. Crossfusor integrates detailed inter-vehicular interactions and car-following dynamics into a robust diffusion framework, improving both the accuracy and realism of predicted trajectories. The model leverages a novel temporal feature encoding framework combining GRU, location-based attention mechanisms, and Fourier embedding to capture historical vehicle dynamics. It employs noise scaled by these encoded historical features in the forward diffusion process, and uses a cross-attention transformer to model intricate inter-vehicle dependencies in the reverse denoising process. Experimental results on the NGSIM dataset demonstrate that Crossfusor outperforms state-of-the-art models, particularly in long-term predictions, showcasing its potential for enhancing the predictive capabilities of autonomous driving systems.
Authors: Jacob Dunefsky, Philippe Chlenski, Neel Nanda
Abstract: A key goal in mechanistic interpretability is circuit analysis: finding sparse subgraphs of models corresponding to specific behaviors or capabilities. However, MLP sublayers make fine-grained circuit analysis on transformer-based language models difficult. In particular, interpretable features -- such as those found by sparse autoencoders (SAEs) -- are typically linear combinations of extremely many neurons, each with its own nonlinearity to account for. Circuit analysis in this setting thus either yields intractably large circuits or fails to disentangle local and global behavior. To address this we explore transcoders, which seek to faithfully approximate a densely activating MLP layer with a wider, sparsely-activating MLP layer. We successfully train transcoders on language models with 120M, 410M, and 1.4B parameters, and find them to perform at least on par with SAEs in terms of sparsity, faithfulness, and human-interpretability. We then introduce a novel method for using transcoders to perform weights-based circuit analysis through MLP sublayers. The resulting circuits neatly factorize into input-dependent and input-invariant terms. Finally, we apply transcoders to reverse-engineer unknown circuits in the model, and we obtain novel insights regarding the greater-than circuit in GPT2-small. Our results suggest that transcoders can prove effective in decomposing model computations involving MLPs into interpretable circuits. Code is available at https://github.com/jacobdunefsky/transcoder_circuits.
Authors: Yi Fang, Dongzhe Fan, Daochen Zha, Qiaoyu Tan
Abstract: This work studies self-supervised graph learning for text-attributed graphs (TAGs) where nodes are represented by textual attributes. Unlike traditional graph contrastive methods that perturb the numerical feature space and alter the graph's topological structure, we aim to improve view generation through language supervision. This is driven by the prevalence of textual attributes in real applications, which complement graph structures with rich semantic information. However, this presents challenges because of two major reasons. First, text attributes often vary in length and quality, making it difficulty to perturb raw text descriptions without altering their original semantic meanings. Second, although text attributes complement graph structures, they are not inherently well-aligned. To bridge the gap, we introduce GAugLLM, a novel framework for augmenting TAGs. It leverages advanced large language models like Mistral to enhance self-supervised graph learning. Specifically, we introduce a mixture-of-prompt-expert technique to generate augmented node features. This approach adaptively maps multiple prompt experts, each of which modifies raw text attributes using prompt engineering, into numerical feature space. Additionally, we devise a collaborative edge modifier to leverage structural and textual commonalities, enhancing edge augmentation by examining or building connections between nodes. Empirical results across five benchmark datasets spanning various domains underscore our framework's ability to enhance the performance of leading contrastive methods as a plug-in tool. Notably, we observe that the augmented features and graph structure can also enhance the performance of standard generative methods, as well as popular graph neural networks. The open-sourced implementation of our GAugLLM is available at Github.
Authors: Mitchell Ostrow, Adam Eisen, Ila Fiete
Abstract: To generate coherent responses, language models infer unobserved meaning from their input text sequence. One potential explanation for this capability arises from theories of delay embeddings in dynamical systems, which prove that unobserved variables can be recovered from the history of only a handful of observed variables. To test whether language models are effectively constructing delay embeddings, we measure the capacities of sequence models to reconstruct unobserved dynamics. We trained 1-layer transformer decoders and state-space sequence models on next-step prediction from noisy, partially-observed time series data. We found that each sequence layer can learn a viable embedding of the underlying system. However, state-space models have a stronger inductive bias than transformers-in particular, they more effectively reconstruct unobserved information at initialization, leading to more parameter-efficient models and lower error on dynamics tasks. Our work thus forges a novel connection between dynamical systems and deep learning sequence models via delay embedding theory.
Authors: Muhammed Razzak, Andreas Kirsch, Yarin Gal
Abstract: Recently, transductive learning methods, which leverage holdout sets during training, have gained popularity for their potential to improve speed, accuracy, and fairness in machine learning models. Despite this, the composition of the holdout set itself, particularly the balance of sensitive sub-groups, has been largely overlooked. Our experiments on CIFAR and CelebA datasets show that compositional changes in the holdout set can substantially influence fairness metrics. Imbalanced holdout sets exacerbate existing disparities, while balanced holdouts can mitigate issues introduced by imbalanced training data. These findings underline the necessity of constructing holdout sets that are both diverse and representative.
Authors: Seungwoo Son, Wonpyo Park, Woohyun Han, Kyuyeun Kim, Jaeho Lee
Abstract: Despite recent advances in LLM quantization, activation quantization remains to be challenging due to the activation outliers. Conventional remedies, e.g., mixing precisions for different channels, introduce extra overhead and reduce the speedup. In this work, we develop a simple yet effective strategy to facilitate per-tensor activation quantization by preventing the generation of problematic tokens. Precisely, we propose a method to find a set of key-value cache, coined CushionCache, which mitigates outliers in subsequent tokens when inserted as a prefix. CushionCache works in two steps: First, we greedily search for a prompt token sequence that minimizes the maximum activation values in subsequent tokens. Then, we further tune the token cache to regularize the activations of subsequent tokens to be more quantization-friendly. The proposed method successfully addresses activation outliers of LLMs, providing a substantial performance boost for per-tensor activation quantization methods. We thoroughly evaluate our method over a wide range of models and benchmarks and find that it significantly surpasses the established baseline of per-tensor W8A8 quantization and can be seamlessly integrated with the recent activation quantization method.
Authors: M\'elanie Raymond (Universit\'e du Qu\'ebec \`a Montr\'eal), Marie-H\'el\`ene Descary (Universit\'e du Qu\'ebec \`a Montr\'eal), C\'edric Beaulac (Universit\'e du Qu\'ebec \`a Montr\'eal), Fabrice Larribe (Universit\'e du Qu\'ebec \`a Montr\'eal)
Abstract: Over the years, many approaches have been proposed to build ancestral recombination graphs (ARGs), graphs used to represent the genetic relationship between individuals. Among these methods, many rely on the assumption that the most likely graph is among the shortest ones. In this paper, we propose a new approach to build short ARGs: Reinforcement Learning (RL). We exploit the similarities between finding the shortest path between a set of genetic sequences and their most recent common ancestor and finding the shortest path between the entrance and exit of a maze, a classic RL problem. In the maze problem, the learner, called the agent, must learn the directions to take in order to escape as quickly as possible, whereas in our problem, the agent must learn the actions to take between coalescence, mutation, and recombination in order to reach the most recent common ancestor as quickly as possible. Our results show that RL can be used to build ARGs as short as those built with a heuristic algorithm optimized to build short ARGs, and sometimes even shorter. Moreover, our method allows to build a distribution of short ARGs for a given sample, and can also generalize learning to new samples not used during the learning process.
Authors: Josh Gardner, Juan C. Perdomo, Ludwig Schmidt
Abstract: Tabular data -- structured, heterogeneous, spreadsheet-style data with rows and columns -- is widely used in practice across many domains. However, while recent foundation models have reduced the need for developing task-specific datasets and predictors in domains such as language modeling and computer vision, this transfer learning paradigm has not had similar impact in the tabular domain. In this work, we seek to narrow this gap and present TabuLa-8B, a language model for tabular prediction. We define a process for extracting a large, high-quality training dataset from the TabLib corpus, proposing methods for tabular data filtering and quality control. Using the resulting dataset, which comprises over 1.6B rows from 3.1M unique tables, we fine-tune a Llama 3-8B large language model (LLM) for tabular data prediction (classification and binned regression) using a novel packing and attention scheme for tabular prediction. Through evaluation across a test suite of 329 datasets, we find that TabuLa-8B has zero-shot accuracy on unseen tables that is over 15 percentage points (pp) higher than random guessing, a feat that is not possible with existing state-of-the-art tabular prediction models (e.g. XGBoost, TabPFN). In the few-shot setting (1-32 shots), without any fine-tuning on the target datasets, TabuLa-8B is 5-15 pp more accurate than XGBoost and TabPFN models that are explicitly trained on equal, or even up to 16x more data. We release our model, code, and data along with the publication of this paper.
Authors: Gang Liu, Srijit Seal, John Arevalo, Zhenwen Liang, Anne E. Carpenter, Meng Jiang, Shantanu Singh
Abstract: Predicting drug efficacy and safety in vivo requires information on biological responses (e.g., cell morphology and gene expression) to small molecule perturbations. However, current molecular representation learning methods do not provide a comprehensive view of cell states under these perturbations and struggle to remove noise, hindering model generalization. We introduce the Information Alignment (InfoAlign) approach to learn molecular representations through the information bottleneck method in cells. We integrate molecules and cellular response data as nodes into a context graph, connecting them with weighted edges based on chemical, biological, and computational criteria. For each molecule in a training batch, InfoAlign optimizes the encoder's latent representation with a minimality objective to discard redundant structural information. A sufficiency objective decodes the representation to align with different feature spaces from the molecule's neighborhood in the context graph. We demonstrate that the proposed sufficiency objective for alignment is tighter than existing encoder-based contrastive methods. Empirically, we validate representations from InfoAlign in two downstream tasks: molecular property prediction against up to 19 baseline methods across four datasets, plus zero-shot molecule-morphology matching.
Authors: Kaan Sancak, Zhigang Hua, Jin Fang, Yan Xie, Andrey Malevich, Bo Long, Muhammed Fatih Balin, \"Umit V. \c{C}ataly\"urek
Abstract: Graph Neural Networks (GNNs) have shown impressive performance in graph representation learning, but they face challenges in capturing long-range dependencies due to their limited expressive power. To address this, Graph Transformers (GTs) were introduced, utilizing self-attention mechanism to effectively model pairwise node relationships. Despite their advantages, GTs suffer from quadratic complexity w.r.t. the number of nodes in the graph, hindering their applicability to large graphs. In this work, we present Graph-Enhanced Contextual Operator (GECO), a scalable and effective alternative to GTs that leverages neighborhood propagation and global convolutions to effectively capture local and global dependencies in quasilinear time. Our study on synthetic datasets reveals that GECO reaches 169x speedup on a graph with 2M nodes w.r.t. optimized attention. Further evaluations on diverse range of benchmarks showcase that GECO scales to large graphs where traditional GTs often face memory and time limitations. Notably, GECO consistently achieves comparable or superior quality compared to baselines, improving the SOTA up to 4.5%, and offering a scalable and effective solution for large-scale graph learning.
Authors: Jiyao Wang, Nicha C. Dvornek, Peiyu Duan, Lawrence H. Staib, Pamela Ventola, James S. Duncan
Abstract: Task-based fMRI uses actions or stimuli to trigger task-specific brain responses and measures them using BOLD contrast. Despite the significant task-induced spatiotemporal brain activation fluctuations, most studies on task-based fMRI ignore the task context information aligned with fMRI and consider task-based fMRI a coherent sequence. In this paper, we show that using the task structures as data-driven guidance is effective for spatiotemporal analysis. We propose STNAGNN, a GNN-based spatiotemporal architecture, and validate its performance in an autism classification task. The trained model is also interpreted for identifying autism-related spatiotemporal brain biomarkers.
Authors: Pankayaraj Pathmanathan, Souradip Chakraborty, Xiangyu Liu, Yongyuan Liang, Furong Huang
Abstract: Recent advancements in Reinforcement Learning with Human Feedback (RLHF) have significantly impacted the alignment of Large Language Models (LLMs). The sensitivity of reinforcement learning algorithms such as Proximal Policy Optimization (PPO) has led to new line work on Direct Policy Optimization (DPO), which treats RLHF in a supervised learning framework. The increased practical use of these RLHF methods warrants an analysis of their vulnerabilities. In this work, we investigate the vulnerabilities of DPO to poisoning attacks under different scenarios and compare the effectiveness of preference poisoning, a first of its kind. We comprehensively analyze DPO's vulnerabilities under different types of attacks, i.e., backdoor and non-backdoor attacks, and different poisoning methods across a wide array of language models, i.e., LLama 7B, Mistral 7B, and Gemma 7B. We find that unlike PPO-based methods, which, when it comes to backdoor attacks, require at least 4\% of the data to be poisoned to elicit harmful behavior, we exploit the true vulnerabilities of DPO more simply so we can poison the model with only as much as 0.5\% of the data. We further investigate the potential reasons behind the vulnerability and how well this vulnerability translates into backdoor vs non-backdoor attacks.
Authors: Huiqun Huang, Sihong He, Fei Miao
Abstract: Trajectory prediction models that can infer both finite future trajectories and their associated uncertainties of the target vehicles in an online setting (e.g., real-world application scenarios) is crucial for ensuring the safe and robust navigation and path planning of autonomous vehicle motion. However, the majority of existing trajectory prediction models have neither considered reducing the uncertainty as one objective during the training stage nor provided reliable uncertainty quantification during inference stage under potential distribution shift. Therefore, in this paper, we propose the Conformal Uncertainty Quantification under Distribution Shift framework, CUQDS, to quantify the uncertainty of the predicted trajectories of existing trajectory prediction models under potential data distribution shift, while considering improving the prediction accuracy of the models and reducing the estimated uncertainty during the training stage. Specifically, CUQDS includes 1) a learning-based Gaussian process regression module that models the output distribution of the base model (any existing trajectory prediction or time series forecasting neural networks) and reduces the estimated uncertainty by additional loss term, and 2) a statistical-based Conformal P control module to calibrate the estimated uncertainty from the Gaussian process regression module in an online setting under potential distribution shift between training and testing data.
Authors: Qinhua Jiang, Brian Yueshuai He, Changju Lee, Jiaqi Ma
Abstract: Accurate traffic prediction is vital for effective traffic management during hurricane evacuation. This paper proposes a predictive modeling system that integrates Multilayer Perceptron (MLP) and Long-Short Term Memory (LSTM) models to capture both long-term congestion patterns and short-term speed patterns. Leveraging various input variables, including archived traffic data, spatial-temporal road network information, and hurricane forecast data, the framework is designed to address challenges posed by heterogeneous human behaviors, limited evacuation data, and hurricane event uncertainties. Deployed in a real-world traffic prediction system in Louisiana, the model achieved an 82% accuracy in predicting long-term congestion states over a 6-hour period during a 7-day hurricane-impacted duration. The short-term speed prediction model exhibited Mean Absolute Percentage Errors (MAPEs) ranging from 7% to 13% across evacuation horizons from 1 to 6 hours. Evaluation results underscore the model's potential to enhance traffic management during hurricane evacuations, and real-world deployment highlights its adaptability and scalability in diverse hurricane scenarios within extensive transportation networks.
Authors: Yulai Zhao, Masatoshi Uehara, Gabriele Scalia, Tommaso Biancalani, Sergey Levine, Ehsan Hajiramezanali
Abstract: Diffusion models are powerful generative models that allow for precise control over the characteristics of the generated samples. While these diffusion models trained on large datasets have achieved success, there is often a need to introduce additional controls in downstream fine-tuning processes, treating these powerful models as pre-trained diffusion models. This work presents a novel method based on reinforcement learning (RL) to add additional controls, leveraging an offline dataset comprising inputs and corresponding labels. We formulate this task as an RL problem, with the classifier learned from the offline dataset and the KL divergence against pre-trained models serving as the reward functions. We introduce our method, $\textbf{CTRL}$ ($\textbf{C}$onditioning pre-$\textbf{T}$rained diffusion models with $\textbf{R}$einforcement $\textbf{L}$earning), which produces soft-optimal policies that maximize the abovementioned reward functions. We formally demonstrate that our method enables sampling from the conditional distribution conditioned on additional controls during inference. Our RL-based approach offers several advantages over existing methods. Compared to commonly used classifier-free guidance, our approach improves sample efficiency, and can greatly simplify offline dataset construction by exploiting conditional independence between the inputs and additional controls. Furthermore, unlike classifier guidance, we avoid the need to train classifiers from intermediate states to additional controls.
Authors: Dingyang Chen, Qi Zhang, Yinglun Zhu
Abstract: This paper focuses on extending the success of large language models (LLMs) to sequential decision making. Existing efforts either (i) re-train or finetune LLMs for decision making, or (ii) design prompts for pretrained LLMs. The former approach suffers from the computational burden of gradient updates, and the latter approach does not show promising results. In this paper, we propose a new approach that leverages online model selection algorithms to efficiently incorporate LLMs agents into sequential decision making. Statistically, our approach significantly outperforms both traditional decision making algorithms and vanilla LLM agents. Computationally, our approach avoids the need for expensive gradient updates of LLMs, and throughout the decision making process, it requires only a small number of LLM calls. We conduct extensive experiments to verify the effectiveness of our proposed approach. As an example, on a large-scale Amazon dataset, our approach achieves more than a $6$x performance gain over baselines while calling LLMs in only $1.5$\% of the time steps.
Authors: Vincent Olesen, Nina Weng, Aasa Feragen, Eike Petersen
Abstract: Machine learning models have achieved high overall accuracy in medical image analysis. However, performance disparities on specific patient groups pose challenges to their clinical utility, safety, and fairness. This can affect known patient groups - such as those based on sex, age, or disease subtype - as well as previously unknown and unlabeled groups. Furthermore, the root cause of such observed performance disparities is often challenging to uncover, hindering mitigation efforts. In this paper, to address these issues, we leverage Slice Discovery Methods (SDMs) to identify interpretable underperforming subsets of data and formulate hypotheses regarding the cause of observed performance disparities. We introduce a novel SDM and apply it in a case study on the classification of pneumothorax and atelectasis from chest x-rays. Our study demonstrates the effectiveness of SDMs in hypothesis formulation and yields an explanation of previously observed but unexplained performance disparities between male and female patients in widely used chest X-ray datasets and models. Our findings indicate shortcut learning in both classification tasks, through the presence of chest drains and ECG wires, respectively. Sex-based differences in the prevalence of these shortcut features appear to cause the observed classification performance gap, representing a previously underappreciated interaction between shortcut learning and model fairness analyses.
Authors: Ge Shi, Ziwen Kan, Jason Smucny, Ian Davidson
Abstract: In this study, we examine the efficacy of post-hoc local attribution methods in identifying features with predictive power from irrelevant ones in domains characterized by a low signal-to-noise ratio (SNR), a common scenario in real-world machine learning applications. We developed synthetic datasets encompassing symbolic functional, image, and audio data, incorporating a benchmark on the {\it (Model \(\times\) Attribution\(\times\) Noise Condition)} triplet. By rigorously testing various classic models trained from scratch, we gained valuable insights into the performance of these attribution methods in multiple conditions. Based on these findings, we introduce a novel extension to the notable recursive feature elimination (RFE) algorithm, enhancing its applicability for neural networks. Our experiments highlight its strengths in prediction and feature selection, alongside limitations in scalability. Further details and additional minor findings are included in the appendix, with extensive discussions. The codes and resources are available at \href{https://github.com/geshijoker/ChaosMining/}{URL}.
Authors: Wenda Xu, Jiachen Li, William Yang Wang, Lei Li
Abstract: Direct alignment from preferences (DAP) has emerged as a promising paradigm for aligning large language models (LLMs) to human desiderata from pre-collected, offline preference datasets. While recent studies indicate that existing offline DAP methods can directly benefit from online training samples, we highlight the need to develop specific online DAP algorithms to fully harness the power of online training. Specifically, we identify that the learned LLM should adhere to the proximity of the behavior LLM, which collects the training samples. To this end, we propose online Preference Optimization in proximity to the Behavior LLM (BPO), emphasizing the importance of constructing a proper trust region for LLM alignment. We conduct extensive experiments to validate the effectiveness and applicability of our approach by integrating it with various DAP methods, resulting in significant performance improvements across a wide range of tasks when training with the same amount of preference data. Even when only introducing one additional data collection phase, our online BPO improves its offline DAP baseline from 72.0% to 80.2% on TL;DR and from 82.2% to 89.1% on Anthropic Helpfulness in terms of win rate against human reference text.
Authors: Yanyong Huang, Li Yang, Dongjie Wang, Ke Li, Xiuwen Yi, Fengmao Lv, Tianrui Li
Abstract: Semi-supervised multi-label feature selection has recently been developed to solve the curse of dimensionality problem in high-dimensional multi-label data with certain samples missing labels. Although many efforts have been made, most existing methods use a predefined graph approach to capture the sample similarity or the label correlation. In this manner, the presence of noise and outliers within the original feature space can undermine the reliability of the resulting sample similarity graph. It also fails to precisely depict the label correlation due to the existence of unknown labels. Besides, these methods only consider the discriminative power of selected features, while neglecting their redundancy. In this paper, we propose an Adaptive Collaborative Correlation lEarning-based Semi-Supervised Multi-label Feature Selection (Access-MFS) method to address these issues. Specifically, a generalized regression model equipped with an extended uncorrelated constraint is introduced to select discriminative yet irrelevant features and maintain consistency between predicted and ground-truth labels in labeled data, simultaneously. Then, the instance correlation and label correlation are integrated into the proposed regression model to adaptively learn both the sample similarity graph and the label similarity graph, which mutually enhance feature selection performance. Extensive experimental results demonstrate the superiority of the proposed Access-MFS over other state-of-the-art methods.
Authors: Haowei Ni, Shuchen Meng, Xieming Geng, Panfeng Li, Zhuoying Li, Xupeng Chen, Xiaotong Wang, Shiyao Zhang
Abstract: Cardiovascular disease (CVD) is a leading cause of death globally, necessitating precise forecasting models for monitoring vital signs like heart rate, blood pressure, and ECG. Traditional models, such as ARIMA and Prophet, are limited by their need for manual parameter tuning and challenges in handling noisy, sparse, and highly variable medical data. This study investigates advanced deep learning models, including LSTM, and transformer-based architectures, for predicting heart rate time series from the MIT-BIH Database. Results demonstrate that deep learning models, particularly PatchTST, significantly outperform traditional models across multiple metrics, capturing complex patterns and dependencies more effectively. This research underscores the potential of deep learning to enhance patient monitoring and CVD management, suggesting substantial clinical benefits. Future work should extend these findings to larger, more diverse datasets and real-world clinical applications to further validate and optimize model performance.
Authors: Qiugang Zhan, Jinbo Cao, Xiurui Xie, Malu Zhang, Huajin Tang, Guisong Liu
Abstract: Spiking federated learning is an emerging distributed learning paradigm that allows resource-constrained devices to train collaboratively at low power consumption without exchanging local data. It takes advantage of both the privacy computation property in federated learning (FL) and the energy efficiency in spiking neural networks (SNN). Thus, it is highly promising to revolutionize the efficient processing of multimedia data. However, existing spiking federated learning methods employ a random selection approach for client aggregation, assuming unbiased client participation. This neglect of statistical heterogeneity affects the convergence and accuracy of the global model significantly. In our work, we propose a credit assignment-based active client selection strategy, the SFedCA, to judiciously aggregate clients that contribute to the global sample distribution balance. Specifically, the client credits are assigned by the firing intensity state before and after local model training, which reflects the local data distribution difference from the global model. Comprehensive experiments are conducted on various non-identical and independent distribution (non-IID) scenarios. The experimental results demonstrate that the SFedCA outperforms the existing state-of-the-art spiking federated learning methods, and requires fewer communication rounds.
Authors: Zhirui Chen, Vincent Y. F. Tan
Abstract: We consider offline reinforcement learning (RL) with preference feedback in which the implicit reward is a linear function of an unknown parameter. Given an offline dataset, our objective consists in ascertaining the optimal action for each state, with the ultimate goal of minimizing the {\em simple regret}. We propose an algorithm, \underline{RL} with \underline{L}ocally \underline{O}ptimal \underline{W}eights or {\sc RL-LOW}, which yields a simple regret of $\exp ( - \Omega(n/H) )$ where $n$ is the number of data samples and $H$ denotes an instance-dependent hardness quantity that depends explicitly on the suboptimality gap of each action. Furthermore, we derive a first-of-its-kind instance-dependent lower bound in offline RL with preference feedback. Interestingly, we observe that the lower and upper bounds on the simple regret match order-wise in the exponent, demonstrating order-wise optimality of {\sc RL-LOW}. In view of privacy considerations in practical applications, we also extend {\sc RL-LOW} to the setting of $(\varepsilon,\delta)$-differential privacy and show, somewhat surprisingly, that the hardness parameter $H$ is unchanged in the asymptotic regime as $n$ tends to infinity; this underscores the inherent efficiency of {\sc RL-LOW} in terms of preserving the privacy of the observed rewards. Given our focus on establishing instance-dependent bounds, our work stands in stark contrast to previous works that focus on establishing worst-case regrets for offline RL with preference feedback.
Authors: Ryo Karakida, Toshihiro Ota, Masato Taki
Abstract: Transformers have established themselves as the leading neural network model in natural language processing and are increasingly foundational in various domains. In vision, the MLP-Mixer model has demonstrated competitive performance, suggesting that attention mechanisms might not be indispensable. Inspired by this, recent research has explored replacing attention modules with other mechanisms, including those described by MetaFormers. However, the theoretical framework for these models remains underdeveloped. This paper proposes a novel perspective by integrating Krotov's hierarchical associative memory with MetaFormers, enabling a comprehensive representation of the entire Transformer block, encompassing token-/channel-mixing modules, layer normalization, and skip connections, as a single Hopfield network. This approach yields a parallelized MLP-Mixer derived from a three-layer Hopfield network, which naturally incorporates symmetric token-/channel-mixing modules and layer normalization. Empirical studies reveal that symmetric interaction matrices in the model hinder performance in image recognition tasks. Introducing symmetry-breaking effects transitions the performance of the symmetric parallelized MLP-Mixer to that of the vanilla MLP-Mixer. This indicates that during standard training, weight matrices of the vanilla MLP-Mixer spontaneously acquire a symmetry-breaking configuration, enhancing their effectiveness. These findings offer insights into the intrinsic properties of Transformers and MLP-Mixers and their theoretical underpinnings, providing a robust framework for future model design and optimization.
Authors: Haque Ishfaq, Yixin Tan, Yu Yang, Qingfeng Lan, Jianfeng Lu, A. Rupam Mahmood, Doina Precup, Pan Xu
Abstract: Thompson sampling (TS) is one of the most popular exploration techniques in reinforcement learning (RL). However, most TS algorithms with theoretical guarantees are difficult to implement and not generalizable to Deep RL. While the emerging approximate sampling-based exploration schemes are promising, most existing algorithms are specific to linear Markov Decision Processes (MDP) with suboptimal regret bounds, or only use the most basic samplers such as Langevin Monte Carlo. In this work, we propose an algorithmic framework that incorporates different approximate sampling methods with the recently proposed Feel-Good Thompson Sampling (FGTS) approach (Zhang, 2022; Dann et al., 2021), which was previously known to be computationally intractable in general. When applied to linear MDPs, our regret analysis yields the best known dependency of regret on dimensionality, surpassing existing randomized algorithms. Additionally, we provide explicit sampling complexity for each employed sampler. Empirically, we show that in tasks where deep exploration is necessary, our proposed algorithms that combine FGTS and approximate sampling perform significantly better compared to other strong baselines. On several challenging games from the Atari 57 suite, our algorithms achieve performance that is either better than or on par with other strong baselines from the deep RL literature.
Authors: Fan Zhou, Chen Pan, Lintao Ma, Yu Liu, James Zhang, Jun Zhou, Hongyuan Mei, Weitao Lin, Zi Zhuang, Wenxin Ning, Yunhua Hu, Siqiao Xue
Abstract: Time series forecasts of different temporal granularity are widely used in real-world applications, e.g., sales prediction in days and weeks for making different inventory plans. However, these tasks are usually solved separately without ensuring coherence, which is crucial for aligning downstream decisions. Previous works mainly focus on ensuring coherence with some straightforward methods, e.g., aggregation from the forecasts of fine granularity to the coarse ones, and allocation from the coarse granularity to the fine ones. These methods merely take the temporal hierarchical structure to maintain coherence without improving the forecasting accuracy. In this paper, we propose a novel granularity message-passing mechanism (GMP) that leverages temporal hierarchy information to improve forecasting performance and also utilizes an adaptive reconciliation (AR) strategy to maintain coherence without performance loss. Furthermore, we introduce an optimization module to achieve task-based targets while adhering to more real-world constraints. Experiments on real-world datasets demonstrate that our framework (GMP-AR) achieves superior performances on temporal hierarchical forecasting tasks compared to state-of-the-art methods. In addition, our framework has been successfully applied to a real-world task of payment traffic management in Alipay by integrating with the task-based optimization module.
Authors: Byung-Kwan Lee, Sangyun Chung, Chae Won Kim, Beomchan Park, Yong Man Ro
Abstract: Large language and vision models (LLVMs) have been driven by the generalization power of large language models (LLMs) and the advent of visual instruction tuning. Along with scaling them up directly, these models enable LLVMs to showcase powerful vision language (VL) performances by covering diverse tasks via natural language instructions. However, existing open-source LLVMs that perform comparably to closed-source LLVMs such as GPT-4V are often considered too large (e.g., 26B, 34B, and 110B parameters), having a larger number of layers. These large models demand costly, high-end resources for both training and inference. To address this issue, we present a new efficient LLVM family with 1.8B, 3.8B, and 7B LLM model sizes, Traversal of Layers (TroL), which enables the reuse of layers in a token-wise manner. This layer traversing technique simulates the effect of looking back and retracing the answering stream while increasing the number of forward propagation layers without physically adding more layers. We demonstrate that TroL employs a simple layer traversing approach yet efficiently outperforms the open-source LLVMs with larger model sizes and rivals the performances of the closed-source LLVMs with substantial sizes.
Authors: Kukjin Choi, Jihun Yi, Jisoo Mok, Sungroh Yoon
Abstract: Continuous efforts are being made to advance anomaly detection in various manufacturing processes to increase the productivity and safety of industrial sites. Deep learning replaced rule-based methods and recently emerged as a promising method for anomaly detection in diverse industries. However, in the real world, the scarcity of abnormal data and difficulties in obtaining labeled data create limitations in the training of detection models. In this study, we addressed these shortcomings by proposing a learnable data augmentation-based time-series anomaly detection (LATAD) technique that is trained in a self-supervised manner. LATAD extracts discriminative features from time-series data through contrastive learning. At the same time, learnable data augmentation produces challenging negative samples to enhance learning efficiency. We measured anomaly scores of the proposed technique based on latent feature similarities. As per the results, LATAD exhibited comparable or improved performance to the state-of-the-art anomaly detection assessments on several benchmark datasets and provided a gradient-based diagnosis technique to help identify root causes.
Authors: Yizirui Fang, Anthony Bellotti
Abstract: Inductive conformal predictors (ICPs) are algorithms that are able to generate prediction sets, instead of point predictions, which are valid at a user-defined confidence level, only assuming exchangeability. These algorithms are useful for reliable machine learning and are increasing in popularity. The ICP development process involves dividing development data into three parts: training, calibration and test. With access to limited or expensive development data, it is an open question regarding the most efficient way to divide the data. This study provides several experiments to explore this question and consider the case for allowing overlap of examples between training and calibration sets. Conclusions are drawn that will be of value to academics and practitioners planning to use ICPs.
Authors: Yue Jiang, Xiucheng Li, Yile Chen, Shuai Liu, Weilong Kong, Antonis F. Lentzakis, Gao Cong
Abstract: Time series forecasting is essential for our daily activities and precise modeling of the complex correlations and shared patterns among multiple time series is essential for improving forecasting performance. Spatial-Temporal Graph Neural Networks (STGNNs) are widely used in multivariate time series forecasting tasks and have achieved promising performance on multiple real-world datasets for their ability to model the underlying complex spatial and temporal dependencies. However, existing studies have mainly focused on datasets comprising only a few hundred sensors due to the heavy computational cost and memory cost of spatial-temporal GNNs. When applied to larger datasets, these methods fail to capture the underlying complex spatial dependencies and exhibit limited scalability and performance. To this end, we present a Scalable Adaptive Graph Diffusion Forecasting Network (SAGDFN) to capture complex spatial-temporal correlation for large-scale multivariate time series and thereby, leading to exceptional performance in multivariate time series forecasting tasks. The proposed SAGDFN is scalable to datasets of thousands of nodes without the need of prior knowledge of spatial correlation. Extensive experiments demonstrate that SAGDFN achieves comparable performance with state-of-the-art baselines on one real-world dataset of 207 nodes and outperforms all state-of-the-art baselines by a significant margin on three real-world datasets of 2000 nodes.
Authors: Brett Daley, Marlos C. Machado, Martha White
Abstract: The recency heuristic in reinforcement learning is the assumption that stimuli that occurred closer in time to an acquired reward should be more heavily reinforced. The recency heuristic is one of the key assumptions made by TD($\lambda$), which reinforces recent experiences according to an exponentially decaying weighting. In fact, all other widely used return estimators for TD learning, such as $n$-step returns, satisfy a weaker (i.e., non-monotonic) recency heuristic. Why is the recency heuristic effective for temporal credit assignment? What happens when credit is assigned in a way that violates this heuristic? In this paper, we analyze the specific mathematical implications of adopting the recency heuristic in TD learning. We prove that any return estimator satisfying this heuristic: 1) is guaranteed to converge to the correct value function, 2) has a relatively fast contraction rate, and 3) has a long window of effective credit assignment, yet bounded worst-case variance. We also give a counterexample where on-policy, tabular TD methods violating the recency heuristic diverge. Our results offer some of the first theoretical evidence that credit assignment based on the recency heuristic facilitates learning.
Authors: Yu-hsuan Chen, Jonathan Cagan, Levent Burak kara
Abstract: Designing for manufacturing poses significant challenges in part due to the computation bottleneck of Computer-Aided Manufacturing (CAM) simulations. Although deep learning as an alternative offers fast inference, its performance is dependently bounded by the need for abundant training data. Representation learning, particularly through pre-training, offers promise for few-shot learning, aiding in manufacturability tasks where data can be limited. This work introduces VIRL, a Volume-Informed Representation Learning approach to pre-train a 3D geometric encoder. The pretrained model is evaluated across four manufacturability indicators obtained from CAM simulations: subtractive machining (SM) time, additive manufacturing (AM) time, residual von Mises stress, and blade collisions during Laser Power Bed Fusion process. Across all case studies, the model pre-trained by VIRL shows substantial enhancements on demonstrating improved generalizability with limited data and superior performance with larger datasets. Regarding deployment strategy, case-specific phenomenon exists where finetuning VIRL-pretrained models adversely affects AM tasks with limited data but benefits SM time prediction. Moreover, the efficacy of Low-rank adaptation (LoRA), which balances between probing and finetuning, is explored. LoRA shows stable performance akin to probing with limited data, while achieving a higher upper bound than probing as data size increases, without the computational costs of finetuning. Furthermore, static normalization of manufacturing indicators consistently performs well across tasks, while dynamic normalization enhances performance when a reliable task dependent input is available.
Authors: Ji Xu, Tianlong Xiao, Jinye Yang, Panpan Zhu
Abstract: Density peaks clustering (DP) has the ability of detecting clusters of arbitrary shape and clustering non-Euclidean space data, but its quadratic complexity in both computing and storage makes it difficult to scale for big data. Various approaches have been proposed in this regard, including MapReduce based distribution computing, multi-core parallelism, presentation transformation (e.g., kd-tree, Z-value), granular computing, and so forth. However, most of these existing methods face two limitations. One is their target datasets are mostly constrained to be in Euclidian space, the other is they emphasize only on local neighbors while ignoring global data distribution due to restriction to cut-off kernel when computing density. To address the two issues, we present a faithful and parallel DP method that makes use of two types of vector-like distance matrices and an inverse leading-node-finding policy. The method is implemented on a message passing interface (MPI) system. Extensive experiments showed that our method is capable of clustering non-Euclidean data such as in community detection, while outperforming the state-of-the-art counterpart methods in accuracy when clustering large Euclidean data. Our code is publicly available at https://github.com/alanxuji/FaithPDP.
Authors: Dongwon Jo, Taesu Kim, Yulhwa Kim, Jae-Joon Kim
Abstract: Binarization, which converts weight parameters to binary values, has emerged as an effective strategy to reduce the size of large language models (LLMs). However, typical binarization techniques significantly diminish linguistic effectiveness of LLMs. To address this issue, we introduce a novel binarization technique called Mixture of Scales (BinaryMoS). Unlike conventional methods, BinaryMoS employs multiple scaling experts for binary weights, dynamically merging these experts for each token to adaptively generate scaling factors. This token-adaptive approach boosts the representational power of binarized LLMs by enabling contextual adjustments to the values of binary weights. Moreover, because this adaptive process only involves the scaling factors rather than the entire weight matrix, BinaryMoS maintains compression efficiency similar to traditional static binarization methods. Our experimental results reveal that BinaryMoS surpasses conventional binarization techniques in various natural language processing tasks and even outperforms 2-bit quantization methods, all while maintaining similar model size to static binarization techniques.
Authors: Federico Errica, Giuseppe Siracusano, Davide Sanvito, Roberto Bifulco
Abstract: Large Language Models (LLMs) changed the way we design and interact with software systems. Their ability to process and extract information from text has drastically improved productivity in a number of routine tasks. Developers that want to include these models in their software stack, however, face a dreadful challenge: debugging their inconsistent behavior across minor variations of the prompt. We therefore introduce two metrics for classification tasks, namely sensitivity and consistency, which are complementary to task performance. First, sensitivity measures changes of predictions across rephrasings of the prompt, and does not require access to ground truth labels. Instead, consistency measures how predictions vary across rephrasings for elements of the same class. We perform an empirical comparison of these metrics on text classification tasks, using them as guideline for understanding failure modes of the LLM. Our hope is that sensitivity and consistency will be powerful allies in automatic prompt engineering frameworks to obtain LLMs that balance robustness with performance.
Authors: Carla Schenker, Xiulin Wang, David Horner, Morten A. Rasmussen, Evrim Acar
Abstract: Data fusion models based on Coupled Matrix and Tensor Factorizations (CMTF) have been effective tools for joint analysis of data from multiple sources. While the vast majority of CMTF models are based on the strictly multilinear CANDECOMP/PARAFAC (CP) tensor model, recently also the more flexible PARAFAC2 model has been integrated into CMTF models. PARAFAC2 tensor models can handle irregular/ragged tensors and have shown to be especially useful for modelling dynamic data with unaligned or irregular time profiles. However, existing PARAFAC2-based CMTF models have limitations in terms of possible regularizations on the factors and/or types of coupling between datasets. To address these limitations, in this paper we introduce a flexible algorithmic framework that fits PARAFAC2-based CMTF models using Alternating Optimization (AO) and the Alternating Direction Method of Multipliers (ADMM). The proposed framework allows to impose various constraints on all modes and linear couplings to other matrix-, CP- or PARAFAC2-models. Experiments on various simulated and a real dataset demonstrate the utility and versatility of the proposed framework as well as its benefits in terms of accuracy and efficiency in comparison with state-of-the-art methods.
Authors: Menglong Zhang, Fuyuan Qian, Quanying Liu
Abstract: Fast adaptation to new tasks is extremely important for embodied agents in the real world. Meta-reinforcement learning (meta-RL) has emerged as an effective method to enable fast adaptation in unknown environments. Compared to on-policy meta-RL algorithms, off-policy algorithms rely heavily on efficient data sampling strategies to extract and represent the historical trajectories. However, little is known about how different data sampling methods impact the ability of meta-RL agents to represent unknown environments. Here, we investigate the impact of data sampling strategies on the exploration and adaptability of meta-RL agents. Specifically, we conducted experiments with two types of off-policy meta-RL algorithms based on Thompson sampling and Bayes-optimality theories in continuous control tasks within the MuJoCo environment and sparse reward navigation tasks. Our analysis revealed the long-memory and short-memory sequence sampling strategies affect the representation and adaptive capabilities of meta-RL agents. We found that the algorithm based on Bayes-optimality theory exhibited more robust and better adaptability than the algorithm based on Thompson sampling, highlighting the importance of appropriate data sampling strategies for the agent's representation of an unknown environment, especially in the case of sparse rewards.
Authors: Yue Jiang, Qin Chao, Yile Chen, Xiucheng Li, Shuai Liu, Gao Cong
Abstract: Location-based services play an critical role in improving the quality of our daily lives. Despite the proliferation of numerous specialized AI models within spatio-temporal context of location-based services, these models struggle to autonomously tackle problems regarding complex urban planing and management. To bridge this gap, we introduce UrbanLLM, a fine-tuned large language model (LLM) designed to tackle diverse problems in urban scenarios. UrbanLLM functions as a problem-solver by decomposing urban-related queries into manageable sub-tasks, identifying suitable spatio-temporal AI models for each sub-task, and generating comprehensive responses to the given queries. Our experimental results indicate that UrbanLLM significantly outperforms other established LLMs, such as Llama and the GPT series, in handling problems concerning complex urban activity planning and management. UrbanLLM exhibits considerable potential in enhancing the effectiveness of solving problems in urban scenarios, reducing the workload and reliance for human experts.
Authors: Pierre Boudart (DI-ENS, PSL), Alessandro Rudi (PSL, DI-ENS, Inria), Pierre Gaillard (UGA, LJK)
Abstract: We study a theoretical and algorithmic framework for structured prediction in the online learning setting. The problem of structured prediction, i.e. estimating function where the output space lacks a vectorial structure, is well studied in the literature of supervised statistical learning. We show that our algorithm is a generalisation of optimal algorithms from the supervised learning setting, and achieves the same excess risk upper bound also when data are not i.i.d. Moreover, we consider a second algorithm designed especially for non-stationary data distributions, including adversarial data. We bound its stochastic regret in function of the variation of the data distributions.
Authors: Haoze Wu, Zihan Qiu, Zili Wang, Hang Zhao, Jie Fu
Abstract: Mixture-of-Experts (MoE) has been demonstrated as an efficient method to scale up models. By dynamically and sparsely selecting activated experts, MoE can effectively reduce computational costs. Despite the success, we observe that many tokens in the MoE models have uncertain routing results. These tokens have nearly equal scores for choosing each expert, and we demonstrate that this uncertainty can lead to incorrect selections. Inspired by the Global Workspace Theory (GWT), we propose a new fine-tuning method, GW-MoE, to address this issue. The core idea is to broadcast the uncertain tokens across experts during fine-tuning. Therefore, these tokens can acquire the necessary knowledge from any expert during inference and become less sensitive to the choice. GW-MoE does not introduce additional inference overhead. We validate that GW can mitigate the uncertain problem and consistently improve in different tasks (text classification, question answering, summarization, code generation, and mathematical problem solving) and model sizes (650M and 8B parameters).
Authors: Liad Erez, Alon Cohen, Tomer Koren, Yishay Mansour, Shay Moran
Abstract: We study multiclass PAC learning with bandit feedback, where inputs are classified into one of $K$ possible labels and feedback is limited to whether or not the predicted labels are correct. Our main contribution is in designing a novel learning algorithm for the agnostic $(\varepsilon,\delta)$-PAC version of the problem, with sample complexity of $O\big( (\operatorname{poly}(K) + 1 / \varepsilon^2) \log (|H| / \delta) \big)$ for any finite hypothesis class $H$. In terms of the leading dependence on $\varepsilon$, this improves upon existing bounds for the problem, that are of the form $O(K/\varepsilon^2)$. We also provide an extension of this result to general classes and establish similar sample complexity bounds in which $\log |H|$ is replaced by the Natarajan dimension. This matches the optimal rate in the full-information version of the problem and resolves an open question studied by Daniely, Sabato, Ben-David, and Shalev-Shwartz (2011) who demonstrated that the multiplicative price of bandit feedback in realizable PAC learning is $\Theta(K)$. We complement this by revealing a stark contrast with the agnostic case, where the price of bandit feedback is only $O(1)$ as $\varepsilon \to 0$. Our algorithm utilizes a stochastic optimization technique to minimize a log-barrier potential based on Frank-Wolfe updates for computing a low-variance exploration distribution over the hypotheses, and is made computationally efficient provided access to an ERM oracle over $H$.
Authors: David Bergstr\"om, Mattias Tiger, Fredrik Heintz
Abstract: Many of today's data is time-series data originating from various sources, such as sensors, transaction systems, or production systems. Major challenges with such data include privacy and business sensitivity. Generative time-series models have the potential to overcome these problems, allowing representative synthetic data, such as people's movement in cities, to be shared openly and be used to the benefit of society at large. However, contemporary approaches are limited to prohibitively short sequences and small scales. Aside from major memory limitations, the models generate less accurate and less representative samples the longer the sequences are. This issue is further exacerbated by the lack of a comprehensive and accessible benchmark. Furthermore, a common need in practical applications is what-if analysis and dynamic adaptation to data distribution changes, for usage in decision making and to manage a changing world: What if this road is temporarily blocked or another road is added? The focus of this paper is on mobility data, such as people's movement in cities, requiring all these issues to be addressed. To this end, we propose a transformer-based diffusion model, TDDPM, for time-series which outperforms and scales substantially better than state-of-the-art. This is evaluated in a new comprehensive benchmark across several sequence lengths, standard datasets, and evaluation measures. We also demonstrate how the model can be conditioned on a prior over spatial occupancy frequency information, allowing the model to generate mobility data for previously unseen environments and for hypothetical scenarios where the underlying road network and its usage changes. This is evaluated by training on mobility data from part of a city. Then, using only aggregate spatial information as prior, we demonstrate out-of-distribution generalization to the unobserved remainder of the city.
Authors: Bisheng Tang, Xiaojun Chen, Shaopu Wang, Yuexin Xuan, Zhendong Zhao
Abstract: Subgraph federated learning (SFL) is a research methodology that has gained significant attention for its potential to handle distributed graph-structured data. In SFL, the local model comprises graph neural networks (GNNs) with a partial graph structure. However, some SFL models have overlooked the significance of missing cross-subgraph edges, which can lead to local GNNs being unable to message-pass global representations to other parties' GNNs. Moreover, existing SFL models require substantial labeled data, which limits their practical applications. To overcome these limitations, we present a novel SFL framework called FedMpa that aims to learn cross-subgraph node representations. FedMpa first trains a multilayer perceptron (MLP) model using a small amount of data and then propagates the federated feature to the local structures. To further improve the embedding representation of nodes with local subgraphs, we introduce the FedMpae method, which reconstructs the local graph structure with an innovation view that applies pooling operation to form super-nodes. Our extensive experiments on six graph datasets demonstrate that FedMpa is highly effective in node classification. Furthermore, our ablation experiments verify the effectiveness of FedMpa.
Authors: Tianqi Zhao, Ngan Thi Dong, Alan Hanjalic, Megha Khosla
Abstract: Graph Neural Networks (GNNs) have achieved state-of-the-art results in node classification tasks. However, most improvements are in multi-class classification, with less focus on the cases where each node could have multiple labels. The first challenge in studying multi-label node classification is the scarcity of publicly available datasets. To address this, we collected and released three real-world biological datasets and developed a multi-label graph generator with tunable properties. We also argue that traditional notions of homophily and heterophily do not apply well to multi-label scenarios. Therefore, we define homophily and Cross-Class Neighborhood Similarity for multi-label classification and investigate $9$ collected multi-label datasets. Lastly, we conducted a large-scale comparative study with $8$ methods across nine datasets to evaluate current progress in multi-label node classification. We release our code at \url{https://github.com/Tianqi-py/MLGNC}.
Authors: Pratik Gajane
Abstract: We introduce the problem of regret minimization in adversarial multi-dueling bandits. While adversarial preferences have been studied in dueling bandits, they have not been explored in multi-dueling bandits. In this setting, the learner is required to select $m \geq 2$ arms at each round and observes as feedback the identity of the most preferred arm which is based on an arbitrary preference matrix chosen obliviously. We introduce a novel algorithm, MiDEX (Multi Dueling EXP3), to learn from such preference feedback that is assumed to be generated from a pairwise-subset choice model. We prove that the expected cumulative $T$-round regret of MiDEX compared to a Borda-winner from a set of $K$ arms is upper bounded by $O((K \log K)^{1/3} T^{2/3})$. Moreover, we prove a lower bound of $\Omega(K^{1/3} T^{2/3})$ for the expected regret in this setting which demonstrates that our proposed algorithm is near-optimal.
Authors: Francesco Daghero, Alessio Burrello, Massimo Poncino, Enrico Macii, Daniele Jahier Pagliari
Abstract: Depthwise separable convolutions are a fundamental component in efficient Deep Neural Networks, as they reduce the number of parameters and operations compared to traditional convolutions while maintaining comparable accuracy. However, their low data reuse opportunities make deploying them notoriously difficult. In this work, we perform an extensive exploration of alternatives to fuse the depthwise and pointwise kernels that constitute the separable convolutional block. Our approach aims to minimize time-consuming memory transfers by combining different data layouts. When targeting a commercial ultra-low-power device with a three-level memory hierarchy, the GreenWaves GAP8 SoC, we reduce the latency of end-to-end network execution by up to 11.40%. Furthermore, our kernels reduce activation data movements between L2 and L1 memories by up to 52.97%.
Authors: Harry Robertshaw, Lennart Karstensen, Benjamin Jackson, Alejandro Granados, Thomas C. Booth
Abstract: Purpose: Autonomous navigation of catheters and guidewires can enhance endovascular surgery safety and efficacy, reducing procedure times and operator radiation exposure. Integrating tele-operated robotics could widen access to time-sensitive emergency procedures like mechanical thrombectomy (MT). Reinforcement learning (RL) shows potential in endovascular navigation, yet its application encounters challenges without a reward signal. This study explores the viability of autonomous navigation in MT vasculature using inverse RL (IRL) to leverage expert demonstrations. Methods: This study established a simulation-based training and evaluation environment for MT navigation. We used IRL to infer reward functions from expert behaviour when navigating a guidewire and catheter. We utilized soft actor-critic to train models with various reward functions and compared their performance in silico. Results: We demonstrated feasibility of navigation using IRL. When evaluating single versus dual device (i.e. guidewire versus catheter and guidewire) tracking, both methods achieved high success rates of 95% and 96%, respectively. Dual-tracking, however, utilized both devices mimicking an expert. A success rate of 100% and procedure time of 22.6 s were obtained when training with a reward function obtained through reward shaping. This outperformed a dense reward function (96%, 24.9 s) and an IRL-derived reward function (48%, 59.2 s). Conclusions: We have contributed to the advancement of autonomous endovascular intervention navigation, particularly MT, by employing IRL. The results underscore the potential of using reward shaping to train models, offering a promising avenue for enhancing the accessibility and precision of MT. We envisage that future research can extend our methodology to diverse anatomical structures to enhance generalizability.
Authors: Davide Italo Serramazza, Thach Le Nguyen, Georgiana Ifrim
Abstract: Explanation for Multivariate Time Series Classification (MTSC) is an important topic that is under explored. There are very few quantitative evaluation methodologies and even fewer examples of actionable explanation, where the explanation methods are shown to objectively improve specific computational tasks on time series data. In this paper we focus on analyzing InterpretTime, a recent evaluation methodology for attribution methods applied to MTSC. We reproduce the original paper results, showcase some significant weaknesses of the methodology and propose ideas to improve both its accuracy and efficiency. Unlike related work, we go beyond evaluation and also showcase the actionability of the produced explainer ranking, by using the best attribution methods for the task of channel selection in MTSC. We find that perturbation-based methods such as SHAP and Feature Ablation work well across a set of datasets, classifiers and tasks and outperform gradient-based methods. We apply the best ranked explainers to channel selection for MTSC and show significant data size reduction and improved classifier accuracy.
Authors: Lena Schmid, Daniel Biebert, Christian Hakert, Kuan-Hsun Chen, Michel Lang, Markus Pauly, Jian-Jia Chen
Abstract: The rise of machine learning methods on heavily resource constrained devices requires not only the choice of a suitable model architecture for the target platform, but also the optimization of the chosen model with regard to execution time consumption for inference in order to optimally utilize the available resources. Random forests and decision trees are shown to be a suitable model for such a scenario, since they are not only heavily tunable towards the total model size, but also offer a high potential for optimizing their executions according to the underlying memory architecture. In addition to the straightforward strategy of enforcing shorter paths through decision trees and hence reducing the execution time for inference, hardware-aware implementations can optimize the execution time in an orthogonal manner. One particular hardware-aware optimization is to layout the memory of decision trees in such a way, that higher probably paths are less likely to be evicted from system caches. This works particularly well when splits within tree nodes are uneven and have a high probability to visit one of the child nodes. In this paper, we present a method to reduce path lengths by rewarding uneven probability distributions during the training of decision trees at the cost of a minimal accuracy degradation. Specifically, we regularize the impurity computation of the CART algorithm in order to favor not only low impurity, but also highly asymmetric distributions for the evaluation of split criteria and hence offer a high optimization potential for a memory architecture-aware implementation. We show that especially for binary classification data sets and data sets with many samples, this form of regularization can lead to an reduction of up to approximately four times in the execution time with a minimal accuracy degradation.
Authors: Hongyi Zhou, Denis Blessing, Ge Li, Onur Celik, Xiaogang Jia, Gerhard Neumann, Rudolf Lioutikov
Abstract: This work introduces Variational Diffusion Distillation (VDD), a novel method that distills denoising diffusion policies into Mixtures of Experts (MoE) through variational inference. Diffusion Models are the current state-of-the-art in generative modeling due to their exceptional ability to accurately learn and represent complex, multi-modal distributions. This ability allows Diffusion Models to replicate the inherent diversity in human behavior, making them the preferred models in behavior learning such as Learning from Human Demonstrations (LfD). However, diffusion models come with some drawbacks, including the intractability of likelihoods and long inference times due to their iterative sampling process. The inference times, in particular, pose a significant challenge to real-time applications such as robot control. In contrast, MoEs effectively address the aforementioned issues while retaining the ability to represent complex distributions but are notoriously difficult to train. VDD is the first method that distills pre-trained diffusion models into MoE models, and hence, combines the expressiveness of Diffusion Models with the benefits of Mixture Models. Specifically, VDD leverages a decompositional upper bound of the variational objective that allows the training of each expert separately, resulting in a robust optimization scheme for MoEs. VDD demonstrates across nine complex behavior learning tasks, that it is able to: i) accurately distill complex distributions learned by the diffusion model, ii) outperform existing state-of-the-art distillation methods, and iii) surpass conventional methods for training MoE.
Authors: Kun Wang, Guibin Zhang, Xinnan Zhang, Junfeng Fang, Xun Wu, Guohao Li, Shirui Pan, Wei Huang, Yuxuan Liang
Abstract: Graph Neural Networks (GNNs) have become pivotal tools for a range of graph-based learning tasks. Notably, most current GNN architectures operate under the assumption of homophily, whether explicitly or implicitly. While this underlying assumption is frequently adopted, it is not universally applicable, which can result in potential shortcomings in learning effectiveness. In this paper, \textbf{for the first time}, we transfer the prevailing concept of ``one node one receptive field" to the heterophilic graph. By constructing a proxy label predictor, we enable each node to possess a latent prediction distribution, which assists connected nodes in determining whether they should aggregate their associated neighbors. Ultimately, every node can have its own unique aggregation hop and pattern, much like each snowflake is unique and possesses its own characteristics. Based on observations, we innovatively introduce the Heterophily Snowflake Hypothesis and provide an effective solution to guide and facilitate research on heterophilic graphs and beyond. We conduct comprehensive experiments including (1) main results on 10 graphs with varying heterophily ratios across 10 backbones; (2) scalability on various deep GNN backbones (SGC, JKNet, etc.) across various large number of layers (2,4,6,8,16,32 layers); (3) comparison with conventional snowflake hypothesis; (4) efficiency comparison with existing graph pruning algorithms. Our observations show that our framework acts as a versatile operator for diverse tasks. It can be integrated into various GNN frameworks, boosting performance in-depth and offering an explainable approach to choosing the optimal network depth. The source code is available at \url{https://github.com/bingreeky/HeteroSnoH}.
Authors: Jie-Jing Shao, Hao-Sen Shi, Lan-Zhe Guo, Yu-Feng Li
Abstract: In offline Imitation Learning (IL), one of the main challenges is the \textit{covariate shift} between the expert observations and the actual distribution encountered by the agent, because it is difficult to determine what action an agent should take when outside the state distribution of the expert demonstrations. Recently, the model-free solutions introduce the supplementary data and identify the latent expert-similar samples to augment the reliable samples during learning. Model-based solutions build forward dynamic models with conservatism quantification and then generate additional trajectories in the neighborhood of expert demonstrations. However, without reward supervision, these methods are often over-conservative in the out-of-expert-support regions, because only in states close to expert-observed states can there be a preferred action enabling policy optimization. To encourage more exploration on expert-unobserved states, we propose a novel model-based framework, called offline Imitation Learning with Self-paced Reverse Augmentation (SRA). Specifically, we build a reverse dynamic model from the offline demonstrations, which can efficiently generate trajectories leading to the expert-observed states in a self-paced style. Then, we use the subsequent reinforcement learning method to learn from the augmented trajectories and transit from expert-unobserved states to expert-observed states. This framework not only explores the expert-unobserved states but also guides maximizing long-term returns on these states, ultimately enabling generalization beyond the expert data. Empirical results show that our proposal could effectively mitigate the covariate shift and achieve the state-of-the-art performance on the offline imitation learning benchmarks. Project website: \url{https://www.lamda.nju.edu.cn/shaojj/KDD24_SRA/}.
Authors: Miguel Vasco, Takuma Seno, Kenta Kawamoto, Kaushik Subramanian, Peter R. Wurman, Peter Stone
Abstract: Racing autonomous cars faster than the best human drivers has been a longstanding grand challenge for the fields of Artificial Intelligence and robotics. Recently, an end-to-end deep reinforcement learning agent met this challenge in a high-fidelity racing simulator, Gran Turismo. However, this agent relied on global features that require instrumentation external to the car. This paper introduces, to the best of our knowledge, the first super-human car racing agent whose sensor input is purely local to the car, namely pixels from an ego-centric camera view and quantities that can be sensed from on-board the car, such as the car's velocity. By leveraging global features only at training time, the learned agent is able to outperform the best human drivers in time trial (one car on the track at a time) races using only local input features. The resulting agent is evaluated in Gran Turismo 7 on multiple tracks and cars. Detailed ablation experiments demonstrate the agent's strong reliance on visual inputs, making it the first vision-based super-human car racing agent.
Authors: Chi Ma, Mincong Huang, Chao Wang, Yujie Wang, Lei Yu, Chuan Liu, Wei Lin
Abstract: Massive Over-activation Yielded Uplifts(MOYU) is an inherent property of large language models, and dynamic activation(DA) based on the MOYU property is a clever yet under-explored strategy designed to accelerate inference in these models. Existing methods that utilize MOYU often face a significant 'Impossible Trinity': struggling to simultaneously maintain model performance, enhance inference speed, and extend applicability across various architectures. Due to the theoretical ambiguities surrounding MOYU, this paper elucidates the root cause of the MOYU property and outlines the mechanisms behind two primary limitations encountered by current DA methods: 1) history-related activation uncertainty, and 2) semantic-irrelevant activation inertia. Our analysis not only underscores the limitations of current dynamic activation strategies within large-scale LLaMA models but also proposes opportunities for refining the design of future sparsity schemes.
Authors: Matthijs de Goede, Bart Cox, J\'er\'emie Decouchant
Abstract: The training of diffusion-based models for image generation is predominantly controlled by a select few Big Tech companies, raising concerns about privacy, copyright, and data authority due to their lack of transparency regarding training data. To ad-dress this issue, we propose a federated diffusion model scheme that enables the independent and collaborative training of diffusion models without exposing local data. Our approach adapts the Federated Averaging (FedAvg) algorithm to train a Denoising Diffusion Model (DDPM). Through a novel utilization of the underlying UNet backbone, we achieve a significant reduction of up to 74% in the number of parameters exchanged during training,compared to the naive FedAvg approach, whilst simultaneously maintaining image quality comparable to the centralized setting, as evaluated by the FID score.
Authors: Jirui Yang, Peng Chen, Zhihui Lu, Qiang Duan, Yubing Bao
Abstract: Vertical Federated Learning (VFL) facilitates collaborative machine learning without the need for participants to share raw private data. However, recent studies have revealed privacy risks where adversaries might reconstruct sensitive features through data leakage during the learning process. Although data reconstruction methods based on gradient or model information are somewhat effective, they reveal limitations in VFL application scenarios. This is because these traditional methods heavily rely on specific model structures and/or have strict limitations on application scenarios. To address this, our study introduces the Unified InverNet Framework into VFL, which yields a novel and flexible approach (dubbed UIFV) that leverages intermediate feature data to reconstruct original data, instead of relying on gradients or model details. The intermediate feature data is the feature exchanged by different participants during the inference phase of VFL. Experiments on four datasets demonstrate that our methods significantly outperform state-of-the-art techniques in attack precision. Our work exposes severe privacy vulnerabilities within VFL systems that pose real threats to practical VFL applications and thus confirms the necessity of further enhancing privacy protection in the VFL architecture.
Authors: Jarek Liesen, Chris Lu, Andrei Lupu, Jakob N. Foerster, Henning Sprekeler, Robert T. Lange
Abstract: Reinforcement learning (RL) agents are commonly trained and evaluated in the same environment. In contrast, humans often train in a specialized environment before being evaluated, such as studying a book before taking an exam. The potential of such specialized training environments is still vastly underexplored, despite their capacity to dramatically speed up training. The framework of synthetic environments takes a first step in this direction by meta-learning neural network-based Markov decision processes (MDPs). The initial approach was limited to toy problems and produced environments that did not transfer to unseen RL algorithms. We extend this approach in three ways: Firstly, we modify the meta-learning algorithm to discover environments invariant towards hyperparameter configurations and learning algorithms. Secondly, by leveraging hardware parallelism and introducing a curriculum on an agent's evaluation episode horizon, we can achieve competitive results on several challenging continuous control problems. Thirdly, we surprisingly find that contextual bandits enable training RL agents that transfer well to their evaluation environment, even if it is a complex MDP. Hence, we set up our experiments to train synthetic contextual bandits, which perform on par with synthetic MDPs, yield additional insights into the evaluation environment, and can speed up downstream applications.
Authors: Baptiste Abeles, Eugenio Clerico, Gergely Neu
Abstract: We study the generalization error of statistical learning algorithms in a non-i.i.d. setting, where the training data is sampled from a stationary mixing process. We develop an analytic framework for this scenario based on a reduction to online learning with delayed feedback. In particular, we show that the existence of an online learning algorithm with bounded regret (against a fixed statistical learning algorithm in a specially constructed game of online learning with delayed feedback) implies low generalization error of said statistical learning method even if the data sequence is sampled from a mixing time series. The rates demonstrate a trade-off between the amount of delay in the online learning game and the degree of dependence between consecutive data points, with near-optimal rates recovered in a number of well-studied settings when the delay is tuned appropriately as a function of the mixing time of the process.
Authors: Lijia Shi, Shihao Dong
Abstract: The challenge of WAD (web attack detection) is growing as hackers continuously refine their methods to evade traditional detection. Deep learning models excel in handling complex unknown attacks due to their strong generalization and adaptability. However, they are vulnerable to backdoor attacks, where contextually irrelevant fragments are inserted into requests, compromising model stability. While backdoor attacks are well studied in image recognition, they are largely unexplored in WAD. This paper introduces backdoor attacks in WAD, proposing five methods and corresponding defenses. Testing on textCNN, biLSTM, and tinybert models shows an attack success rate over 87%, reducible through fine-tuning. Future research should focus on backdoor defenses in WAD. All the code and data of this paper can be obtained at https://anonymous.4open.science/r/attackDefenceinDL-7E05
URLs: https://anonymous.4open.science/r/attackDefenceinDL-7E05
Authors: Yedi Zhang, Andrew Saxe, Peter E. Latham
Abstract: We investigate the expressivity and learning dynamics of bias-free ReLU networks. We firstly show that two-layer bias-free ReLU networks have limited expressivity: the only odd function two-layer bias-free ReLU networks can express is a linear one. We then show that, under symmetry conditions on the data, these networks have the same learning dynamics as linear networks. This allows us to give closed-form time-course solutions to certain two-layer bias-free ReLU networks, which has not been done for nonlinear networks outside the lazy learning regime. While deep bias-free ReLU networks are more expressive than their two-layer counterparts, they still share a number of similarities with deep linear networks. These similarities enable us to leverage insights from linear networks, leading to a novel understanding of bias-free ReLU networks. Overall, our results show that some properties established for bias-free ReLU networks arise due to equivalence to linear networks, and suggest that including bias or considering asymmetric data are avenues to engage with nonlinear behaviors.
Authors: Antonio Terpin, Nicolas Lanzetti, Florian D\"orfler
Abstract: Diffusion regulates a phenomenal number of natural processes and the dynamics of many successful generative models. Existing models to learn the diffusion terms from observational data rely on complex bilevel optimization problems and properly model only the drift of the system. We propose a new simple model, JKOnet*, which bypasses altogether the complexity of existing architectures while presenting significantly enhanced representational capacity: JKOnet* recovers the potential, interaction, and internal energy components of the underlying diffusion process. JKOnet* minimizes a simple quadratic loss, runs at lightspeed, and drastically outperforms other baselines in practice. Additionally, JKOnet* provides a closed-form optimal solution for linearly parametrized functionals. Our methodology is based on the interpretation of diffusion processes as energy-minimizing trajectories in the probability space via the so-called JKO scheme, which we study via its first-order optimality conditions, in light of few-weeks-old advancements in optimization in the probability space.
Authors: Jingzhao Gu (Beijing Institute of Technology), Haoyang Huang (Chongqing University)
Abstract: Data, algorithms, and arithmetic power are the three foundational conditions for deep learning to be effective in the application domain. Data is the focus for developing deep learning algorithms. In practical engineering applications, some data are affected by the conditions under which more data cannot be obtained or the cost of obtaining data is too high, resulting in smaller data sets (generally several hundred to several thousand) and data sizes that are far smaller than the size of large data sets (tens of thousands). The above two methods are based on the original dataset to generate, in the case of insufficient data volume of the original data may not reflect all the real environment, such as the real environment of the light, silhouette and other information, if the amount of data is not enough, it is difficult to use a simple transformation or neural network generative model to generate the required data. The research in this paper firstly analyses the key points of the data enhancement technology of graph neural network, and at the same time introduces the composition foundation of graph neural network in depth, on the basis of which the data enhancement technology of graph neural network is optimized and analysed.
Authors: Hengyi Wang, Shiwei Tan, Hao Wang
Abstract: Vision transformers (ViTs) have emerged as a significant area of focus, particularly for their capacity to be jointly trained with large language models and to serve as robust vision foundation models. Yet, the development of trustworthy explanation methods for ViTs has lagged, particularly in the context of post-hoc interpretations of ViT predictions. Existing sub-image selection approaches, such as feature-attribution and conceptual models, fall short in this regard. This paper proposes five desiderata for explaining ViTs -- faithfulness, stability, sparsity, multi-level structure, and parsimony -- and demonstrates the inadequacy of current methods in meeting these criteria comprehensively. We introduce a variational Bayesian explanation framework, dubbed ProbAbilistic Concept Explainers (PACE), which models the distributions of patch embeddings to provide trustworthy post-hoc conceptual explanations. Our qualitative analysis reveals the distributions of patch-level concepts, elucidating the effectiveness of ViTs by modeling the joint distribution of patch embeddings and ViT's predictions. Moreover, these patch-level explanations bridge the gap between image-level and dataset-level explanations, thus completing the multi-level structure of PACE. Through extensive experiments on both synthetic and real-world datasets, we demonstrate that PACE surpasses state-of-the-art methods in terms of the defined desiderata.
Authors: Joseph Chakar
Abstract: Bayesian optimization (BO) has emerged as a powerful tool for navigating complex search spaces, showcasing practical applications in the fields of science and engineering.However, since it typically relies on a surrogate model to approximate the objective function, BO grapples with heightened computational costs that tend to escalate as the number of parameters and experiments grows. Several methods such as parallelization, surrogate model approximations, and memory pruning have been proposed to cut down computing time, but they all fall short of resolving the core issue behind BO's curse of dimensionality. In this paper, a 1D reparametrization trick is proposed to break this curse and sustain linear time complexity for BO in high-dimensional landscapes. This fast and scalable approach named SCORE can successfully find the global minimum of needle-in-a-haystack optimization functions and fit real-world data without the high-performance computing resources typically required by state-of-the-art techniques.
Authors: Flora Angileri, Giulia Lombardi, Andrea Fois, Renato Faraone, Carlo Metta, Michele Salvi, Luigi Amedeo Bianchi, Marco Fantozzi, Silvia Giulia Galfr\`e, Daniele Pavesi, Maurizio Parton, Francesco Morandin
Abstract: In 2021, Adam Zsolt Wagner proposed an approach to disprove conjectures in graph theory using Reinforcement Learning (RL). Wagner's idea can be framed as follows: consider a conjecture, such as a certain quantity f(G) < 0 for every graph G; one can then play a single-player graph-building game, where at each turn the player decides whether to add an edge or not. The game ends when all edges have been considered, resulting in a certain graph G_T, and f(G_T) is the final score of the game; RL is then used to maximize this score. This brilliant idea is as simple as innovative, and it lends itself to systematic generalization. Several different single-player graph-building games can be employed, along with various RL algorithms. Moreover, RL maximizes the cumulative reward, allowing for step-by-step rewards instead of a single final score, provided the final cumulative reward represents the quantity of interest f(G_T). In this paper, we discuss these and various other choices that can be significant in Wagner's framework. As a contribution to this systematization, we present four distinct single-player graph-building games. Each game employs both a step-by-step reward system and a single final score. We also propose a principled approach to select the most suitable neural network architecture for any given conjecture, and introduce a new dataset of graphs labeled with their Laplacian spectra. Furthermore, we provide a counterexample for a conjecture regarding the sum of the matching number and the spectral radius, which is simpler than the example provided in Wagner's original paper. The games have been implemented as environments in the Gymnasium framework, and along with the dataset, are available as open-source supplementary materials.
Authors: Du Yin, Hao Xue, Arian Prabowo, Shuang Ao, Flora Salim
Abstract: Traffic forecasting is crucial for smart cities and intelligent transportation initiatives, where deep learning has made significant progress in modeling complex spatio-temporal patterns in recent years. However, current public datasets have limitations in reflecting the ultra-dynamic nature of real-world scenarios, characterized by continuously evolving infrastructures, varying temporal distributions, and temporal gaps due to sensor downtimes or changes in traffic patterns. These limitations inevitably restrict the practical applicability of existing traffic forecasting datasets. To bridge this gap, we present XXLTraffic, the largest available public traffic dataset with the longest timespan and increasing number of sensor nodes over the multiple years observed in the data, curated to support research in ultra-dynamic forecasting. Our benchmark includes both typical time-series forecasting settings with hourly and daily aggregated data and novel configurations that introduce gaps and down-sample the training size to better simulate practical constraints. We anticipate the new XXLTraffic will provide a fresh perspective for the time-series and traffic forecasting communities. It would also offer a robust platform for developing and evaluating models designed to tackle ultra-dynamic and extremely long forecasting problems. Our dataset supplements existing spatio-temporal data resources and leads to new research directions in this domain.
Authors: Du Yin, Jinliang Deng, Shuang Ao, Zechen Li, Hao Xue, Arian Prabowo, Renhe Jiang, Xuan Song, Flora Salim
Abstract: Training models on spatio-temporal (ST) data poses an open problem due to the complicated and diverse nature of the data itself, and it is challenging to ensure the model's performance directly trained on the original ST data. While limiting the variety of training data can make training easier, it can also lead to a lack of knowledge and information for the model, resulting in a decrease in performance. To address this challenge, we presented an innovative paradigm that incorporates three separate forms of curriculum learning specifically targeting from spatial, temporal, and quantile perspectives. Furthermore, our framework incorporates a stacking fusion module to combine diverse information from three types of curriculum learning, resulting in a strong and thorough learning process. We demonstrated the effectiveness of this framework with extensive empirical evaluations, highlighting its better performance in addressing complex ST challenges. We provided thorough ablation studies to investigate the effectiveness of our curriculum and to explain how it contributes to the improvement of learning efficiency on ST data.
Authors: Zahra Gharaee, Scott C. Lowe, ZeMing Gong, Pablo Millan Arias, Nicholas Pellegrino, Austin T. Wang, Joakim Bruslund Haurum, Iuliia Zarubiieva, Lila Kari, Dirk Steinke, Graham W. Taylor, Paul Fieguth, Angel X. Chang
Abstract: As part of an ongoing worldwide effort to comprehend and monitor insect biodiversity, this paper presents the BIOSCAN-5M Insect dataset to the machine learning community and establish several benchmark tasks. BIOSCAN-5M is a comprehensive dataset containing multi-modal information for over 5 million insect specimens, and it significantly expands existing image-based biological datasets by including taxonomic labels, raw nucleotide barcode sequences, assigned barcode index numbers, and geographical information. We propose three benchmark experiments to demonstrate the impact of the multi-modal data types on the classification and clustering accuracy. First, we pretrain a masked language model on the DNA barcode sequences of the \mbox{BIOSCAN-5M} dataset, and demonstrate the impact of using this large reference library on species- and genus-level classification performance. Second, we propose a zero-shot transfer learning task applied to images and DNA barcodes to cluster feature embeddings obtained from self-supervised learning, to investigate whether meaningful clusters can be derived from these representation embeddings. Third, we benchmark multi-modality by performing contrastive learning on DNA barcodes, image data, and taxonomic information. This yields a general shared embedding space enabling taxonomic classification using multiple types of information and modalities. The code repository of the BIOSCAN-5M Insect dataset is available at {\url{https://github.com/zahrag/BIOSCAN-5M}}
Authors: Wenjie Du, Jun Wang, Linglong Qian, Yiyuan Yang, Fanxing Liu, Zepu Wang, Zina Ibrahim, Haoxin Liu, Zhiyuan Zhao, Yingjie Zhou, Wenjia Wang, Kaize Ding, Yuxuan Liang, B. Aditya Prakash, Qingsong Wen
Abstract: Effective imputation is a crucial preprocessing step for time series analysis. Despite the development of numerous deep learning algorithms for time series imputation, the community lacks standardized and comprehensive benchmark platforms to effectively evaluate imputation performance across different settings. Moreover, although many deep learning forecasting algorithms have demonstrated excellent performance, whether their modeling achievements can be transferred to time series imputation tasks remains unexplored. To bridge these gaps, we develop TSI-Bench, the first (to our knowledge) comprehensive benchmark suite for time series imputation utilizing deep learning techniques. The TSI-Bench pipeline standardizes experimental settings to enable fair evaluation of imputation algorithms and identification of meaningful insights into the influence of domain-appropriate missingness ratios and patterns on model performance. Furthermore, TSI-Bench innovatively provides a systematic paradigm to tailor time series forecasting algorithms for imputation purposes. Our extensive study across 34,804 experiments, 28 algorithms, and 8 datasets with diverse missingness scenarios demonstrates TSI-Bench's effectiveness in diverse downstream tasks and potential to unlock future directions in time series imputation research and analysis. The source code and experiment logs are available at https://github.com/WenjieDu/AwesomeImputation.
Authors: Angel Daruna, Vasily Zadorozhnyy, Georgina Lukoczki, Han-Pang Chiu
Abstract: Machine Learning (ML) for Mineral Prospectivity Mapping (MPM) remains a challenging problem as it requires the analysis of associations between large-scale multi-modal geospatial data and few historical mineral commodity observations (positive labels). Recent MPM works have explored Deep Learning (DL) as a modeling tool with more representation capacity. However, these overparameterized methods may be more prone to overfitting due to their reliance on scarce labeled data. While a large quantity of unlabeled geospatial data exists, no prior MPM works have considered using such information in a self-supervised manner. Our MPM approach uses a masked image modeling framework to pretrain a backbone neural network in a self-supervised manner using unlabeled geospatial data alone. After pretraining, the backbone network provides feature extraction for downstream MPM tasks. We evaluated our approach alongside existing methods to assess mineral prospectivity of Mississippi Valley Type (MVT) and Clastic-Dominated (CD) Lead-Zinc deposits in North America and Australia. Our results demonstrate that self-supervision promotes robustness in learned features, improving prospectivity predictions. Additionally, we leverage explainable artificial intelligence techniques to demonstrate that individual predictions can be interpreted from a geological perspective.
Authors: Silvia Garc\'ia-M\'endez, Francisco de Arriba-P\'erez, Francisco J. Gonz\'alez-Casta\~no, Javier Vales-Alonso
Abstract: Artificial Intelligence (AI) has found application in Human Activity Recognition (HAR) in competitive sports. To date, most Machine Learning (ML) approaches for HAR have relied on offline (batch) training, imposing higher computational and tagging burdens compared to online processing unsupervised approaches. Additionally, the decisions behind traditional ML predictors are opaque and require human interpretation. In this work, we apply an online processing unsupervised clustering approach based on low-cost wearable Inertial Measurement Units (IMUs). The outcomes generated by the system allow for the automatic expansion of limited tagging available (e.g., by referees) within those clusters, producing pertinent information for the explainable classification stage. Specifically, our work focuses on achieving automatic explainability for predictions related to athletes' activities, distinguishing between correct, incorrect, and cheating practices in Nordic Walking. The proposed solution achieved performance metrics of close to 100 % on average.
Authors: Zhaoxian Wu, Tayfun Gokmen, Malte J. Rasch, Tianyi Chen
Abstract: Given the high economic and environmental costs of using large vision or language models, analog in-memory accelerators present a promising solution for energy-efficient AI. While inference on analog accelerators has been studied recently, the training perspective is underexplored. Recent studies have shown that the "workhorse" of digital AI training - stochastic gradient descent (SGD) algorithm converges inexactly when applied to model training on non-ideal devices. This paper puts forth a theoretical foundation for gradient-based training on analog devices. We begin by characterizing the non-convergent issue of SGD, which is caused by the asymmetric updates on the analog devices. We then provide a lower bound of the asymptotic error to show that there is a fundamental performance limit of SGD-based analog training rather than an artifact of our analysis. To address this issue, we study a heuristic analog algorithm called Tiki-Taka that has recently exhibited superior empirical performance compared to SGD and rigorously show its ability to exactly converge to a critical point and hence eliminates the asymptotic error. The simulations verify the correctness of the analyses.
Authors: Rylan Schaeffer, Mikail Khona, Sanmi Koyejo
Abstract: In-context learning is a powerful capability of certain machine learning models that arguably underpins the success of today's frontier AI models. However, in-context learning is critically limited to settings where the in-context distribution of interest $p_{\theta}^{ICL}( x|\mathcal{D})$ can be straightforwardly expressed and/or parameterized by the model; for instance, language modeling relies on expressing the next-token distribution as a categorical distribution parameterized by the network's output logits. In this work, we present a more general form of in-context learning without such a limitation that we call \textit{in-context learning of energy functions}. The idea is to instead learn the unconstrained and arbitrary in-context energy function $E_{\theta}^{ICL}(x|\mathcal{D})$ corresponding to the in-context distribution $p_{\theta}^{ICL}(x|\mathcal{D})$. To do this, we use classic ideas from energy-based modeling. We provide preliminary evidence that our method empirically works on synthetic data. Interestingly, our work contributes (to the best of our knowledge) the first example of in-context learning where the input space and output space differ from one another, suggesting that in-context learning is a more-general capability than previously realized.
Authors: Riccardo Zamboni, Duilio Cirino, Marcello Restelli, Mirco Mutti
Abstract: The problem of pure exploration in Markov decision processes has been cast as maximizing the entropy over the state distribution induced by the agent's policy, an objective that has been extensively studied. However, little attention has been dedicated to state entropy maximization under partial observability, despite the latter being ubiquitous in applications, e.g., finance and robotics, in which the agent only receives noisy observations of the true state governing the system's dynamics. How can we address state entropy maximization in those domains? In this paper, we study the simple approach of maximizing the entropy over observations in place of true latent states. First, we provide lower and upper bounds to the approximation of the true state entropy that only depends on some properties of the observation function. Then, we show how knowledge of the latter can be exploited to compute a principled regularization of the observation entropy to improve performance. With this work, we provide both a flexible approach to bring advances in state entropy maximization to the POMDP setting and a theoretical characterization of its intrinsic limits.
Authors: Leonardo Pellegrina, Fabio Vandin
Abstract: Learning interpretable models has become a major focus of machine learning research, given the increasing prominence of machine learning in socially important decision-making. Among interpretable models, rule lists are among the best-known and easily interpretable ones. However, finding optimal rule lists is computationally challenging, and current approaches are impractical for large datasets. We present a novel and scalable approach to learn nearly optimal rule lists from large datasets. Our algorithm uses sampling to efficiently obtain an approximation of the optimal rule list with rigorous guarantees on the quality of the approximation. In particular, our algorithm guarantees to find a rule list with accuracy very close to the optimal rule list when a rule list with high accuracy exists. Our algorithm builds on the VC-dimension of rule lists, for which we prove novel upper and lower bounds. Our experimental evaluation on large datasets shows that our algorithm identifies nearly optimal rule lists with a speed-up up to two orders of magnitude over state-of-the-art exact approaches. Moreover, our algorithm is as fast as, and sometimes faster than, recent heuristic approaches, while reporting higher quality rule lists. In addition, the rules reported by our algorithm are more similar to the rules in the optimal rule list than the rules from heuristic approaches.
Authors: Chen Henry Wu, Jing Yu Koh, Ruslan Salakhutdinov, Daniel Fried, Aditi Raghunathan
Abstract: Vision-enabled language models (VLMs) are now used to build autonomous multimodal agents capable of taking actions in real environments. In this paper, we show that multimodal agents raise new safety risks, even though attacking agents is more challenging than prior attacks due to limited access to and knowledge about the environment. Our attacks use adversarial text strings to guide gradient-based perturbation over one trigger image in the environment: (1) our captioner attack attacks white-box captioners if they are used to process images into captions as additional inputs to the VLM; (2) our CLIP attack attacks a set of CLIP models jointly, which can transfer to proprietary VLMs. To evaluate the attacks, we curated VisualWebArena-Adv, a set of adversarial tasks based on VisualWebArena, an environment for web-based multimodal agent tasks. Within an L-infinity norm of $16/256$ on a single image, the captioner attack can make a captioner-augmented GPT-4V agent execute the adversarial goals with a 75% success rate. When we remove the captioner or use GPT-4V to generate its own captions, the CLIP attack can achieve success rates of 21% and 43%, respectively. Experiments on agents based on other VLMs, such as Gemini-1.5, Claude-3, and GPT-4o, show interesting differences in their robustness. Further analysis reveals several key factors contributing to the attack's success, and we also discuss the implications for defenses as well. Project page: https://chenwu.io/attack-agent Code and data: https://github.com/ChenWu98/agent-attack
URLs: https://chenwu.io/attack-agent, https://github.com/ChenWu98/agent-attack
Authors: Nikolas Koutsoubis, Yasin Yilmaz, Ravi P. Ramachandran, Matthew Schabath, Ghulam Rasool
Abstract: Machine learning (ML) and Artificial Intelligence (AI) have fueled remarkable advancements, particularly in healthcare. Within medical imaging, ML models hold the promise of improving disease diagnoses, treatment planning, and post-treatment monitoring. Various computer vision tasks like image classification, object detection, and image segmentation are poised to become routine in clinical analysis. However, privacy concerns surrounding patient data hinder the assembly of large training datasets needed for developing and training accurate, robust, and generalizable models. Federated Learning (FL) emerges as a compelling solution, enabling organizations to collaborate on ML model training by sharing model training information (gradients) rather than data (e.g., medical images). FL's distributed learning framework facilitates inter-institutional collaboration while preserving patient privacy. However, FL, while robust in privacy preservation, faces several challenges. Sensitive information can still be gleaned from shared gradients that are passed on between organizations during model training. Additionally, in medical imaging, quantifying model confidence\uncertainty accurately is crucial due to the noise and artifacts present in the data. Uncertainty estimation in FL encounters unique hurdles due to data heterogeneity across organizations. This paper offers a comprehensive review of FL, privacy preservation, and uncertainty estimation, with a focus on medical imaging. Alongside a survey of current research, we identify gaps in the field and suggest future directions for FL research to enhance privacy and address noisy medical imaging data challenges.
Authors: Berthy T. Feng, Ricardo Baptista, Katherine L. Bouman
Abstract: Diffusion models excel at creating visually-convincing images, but they often struggle to meet subtle constraints inherent in the training data. Such constraints could be physics-based (e.g., satisfying a PDE), geometric (e.g., respecting symmetry), or semantic (e.g., including a particular number of objects). When the training data all satisfy a certain constraint, enforcing this constraint on a diffusion model not only improves its distribution-matching accuracy but also makes it more reliable for generating valid synthetic data and solving constrained inverse problems. However, existing methods for constrained diffusion models are inflexible with different types of constraints. Recent work proposed to learn mirror diffusion models (MDMs) in an unconstrained space defined by a mirror map and to impose the constraint with an inverse mirror map, but analytical mirror maps are challenging to derive for complex constraints. We propose neural approximate mirror maps (NAMMs) for general constraints. Our approach only requires a differentiable distance function from the constraint set. We learn an approximate mirror map that pushes data into an unconstrained space and a corresponding approximate inverse that maps data back to the constraint set. A generative model, such as an MDM, can then be trained in the learned mirror space and its samples restored to the constraint set by the inverse map. We validate our approach on a variety of constraints, showing that compared to an unconstrained diffusion model, a NAMM-based MDM substantially improves constraint satisfaction. We also demonstrate how existing diffusion-based inverse-problem solvers can be easily applied in the learned mirror space to solve constrained inverse problems.
Authors: Yuting Feng, Vincent Y. F. Tan, Bogdan Cautis
Abstract: We consider a ubiquitous scenario in the study of Influence Maximization (IM), in which there is limited knowledge about the topology of the diffusion network. We set the IM problem in a multi-round diffusion campaign, aiming to maximize the number of distinct users that are influenced. Leveraging the capability of bandit algorithms to effectively balance the objectives of exploration and exploitation, as well as the expressivity of neural networks, our study explores the application of neural bandit algorithms to the IM problem. We propose the framework IM-GNB (Influence Maximization with Graph Neural Bandits), where we provide an estimate of the users' probabilities of being influenced by influencers (also known as diffusion seeds). This initial estimate forms the basis for constructing both an exploitation graph and an exploration one. Subsequently, IM-GNB handles the exploration-exploitation tradeoff, by selecting seed nodes in real-time using Graph Convolutional Networks (GCN), in which the pre-estimated graphs are employed to refine the influencers' estimated rewards in each contextual setting. Through extensive experiments on two large real-world datasets, we demonstrate the effectiveness of IM-GNB compared with other baseline methods, significantly improving the spread outcome of such diffusion campaigns, when the underlying network is unknown.
Authors: Jinuk Kim, Marwa El Halabi, Mingi Ji, Hyun Oh Song
Abstract: Recent works show that reducing the number of layers in a convolutional neural network can enhance efficiency while maintaining the performance of the network. Existing depth compression methods remove redundant non-linear activation functions and merge the consecutive convolution layers into a single layer. However, these methods suffer from a critical drawback; the kernel size of the merged layers becomes larger, significantly undermining the latency reduction gained from reducing the depth of the network. We show that this problem can be addressed by jointly pruning convolution layers and activation functions. To this end, we propose LayerMerge, a novel depth compression method that selects which activation layers and convolution layers to remove, to achieve a desired inference speed-up while minimizing performance loss. Since the corresponding selection problem involves an exponential search space, we formulate a novel surrogate optimization problem and efficiently solve it via dynamic programming. Empirical results demonstrate that our method consistently outperforms existing depth compression and layer pruning methods on various network architectures, both on image classification and generation tasks. We release the code at https://github.com/snu-mllab/LayerMerge.
Authors: Yuqing Wang, Ye He, Molei Tao
Abstract: Most existing theoretical investigations of the accuracy of diffusion models, albeit significant, assume the score function has been approximated to a certain accuracy, and then use this a priori bound to control the error of generation. This article instead provides a first quantitative understanding of the whole generation process, i.e., both training and sampling. More precisely, it conducts a non-asymptotic convergence analysis of denoising score matching under gradient descent. In addition, a refined sampling error analysis for variance exploding models is also provided. The combination of these two results yields a full error analysis, which elucidates (again, but this time theoretically) how to design the training and sampling processes for effective generation. For instance, our theory implies a preference toward noise distribution and loss weighting that qualitatively agree with the ones used in [Karras et al. 2022]. It also provides some perspectives on why the time and variance schedule used in [Karras et al. 2022] could be better tuned than the pioneering version in [Song et al. 2020].
Authors: Maciej Besta, Florian Scheidl, Lukas Gianinazzi, Shachar Klaiman, J\"urgen M\"uller, Torsten Hoefler
Abstract: Higher-order graph neural networks (HOGNNs) are an important class of GNN models that harness polyadic relations between vertices beyond plain edges. They have been used to eliminate issues such as over-smoothing or over-squashing, to significantly enhance the accuracy of GNN predictions, to improve the expressiveness of GNN architectures, and for numerous other goals. A plethora of HOGNN models have been introduced, and they come with diverse neural architectures, and even with different notions of what the "higher-order" means. This richness makes it very challenging to appropriately analyze and compare HOGNN models, and to decide in what scenario to use specific ones. To alleviate this, we first design an in-depth taxonomy and a blueprint for HOGNNs. This facilitates designing models that maximize performance. Then, we use our taxonomy to analyze and compare the available HOGNN models. The outcomes of our analysis are synthesized in a set of insights that help to select the most beneficial GNN model in a given scenario, and a comprehensive list of challenges and opportunities for further research into more powerful HOGNNs.
Authors: Tom Tseng, Euan McLean, Kellin Pelrine, Tony T. Wang, Adam Gleave
Abstract: Prior work found that superhuman Go AIs like KataGo can be defeated by simple adversarial strategies. In this paper, we study if simple defenses can improve KataGo's worst-case performance. We test three natural defenses: adversarial training on hand-constructed positions, iterated adversarial training, and changing the network architecture. We find that some of these defenses are able to protect against previously discovered attacks. Unfortunately, we also find that none of these defenses are able to withstand adaptive attacks. In particular, we are able to train new adversaries that reliably defeat our defended agents by causing them to blunder in ways humans would not. Our results suggest that building robust AI systems is challenging even in narrow domains such as Go. For interactive examples of attacks and a link to our codebase, see https://goattack.far.ai.
URLs: https://goattack.far.ai.
Authors: Shenghui Li, Fanghua Ye, Meng Fang, Jiaxu Zhao, Yun-Hin Chan, Edith C. -H. Ngai, Thiemo Voigt
Abstract: The recent development of Foundation Models (FMs), represented by large language models, vision transformers, and multimodal models, has been making a significant impact on both academia and industry. Compared with small-scale models, FMs have a much stronger demand for high-volume data during the pre-training phase. Although general FMs can be pre-trained on data collected from open sources such as the Internet, domain-specific FMs need proprietary data, posing a practical challenge regarding the amount of data available due to privacy concerns. Federated Learning (FL) is a collaborative learning paradigm that breaks the barrier of data availability from different participants. Therefore, it provides a promising solution to customize and adapt FMs to a wide range of domain-specific tasks using distributed datasets whilst preserving privacy. This survey paper discusses the potentials and challenges of synergizing FL and FMs and summarizes core techniques, future directions, and applications. A periodically updated paper collection on FM-FL is available at https://github.com/lishenghui/awesome-fm-fl.
Authors: Haoxiang Wang, Wei Xiong, Tengyang Xie, Han Zhao, Tong Zhang
Abstract: Reinforcement learning from human feedback (RLHF) has emerged as the primary method for aligning large language models (LLMs) with human preferences. The RLHF process typically starts by training a reward model (RM) using human preference data. Conventional RMs are trained on pairwise responses to the same user request, with relative ratings indicating which response humans prefer. The trained RM serves as a proxy for human preferences. However, due to the black-box nature of RMs, their outputs lack interpretability, as humans cannot intuitively understand why an RM thinks a response is good or not. As RMs act as human preference proxies, we believe they should be human-interpretable to ensure that their internal decision processes are consistent with human preferences and to prevent reward hacking in LLM alignment. To build RMs with interpretable preferences, we propose a two-stage approach: i) train an Absolute-Rating Multi-Objective Reward Model (ArmoRM) with multi-dimensional absolute-rating data, each dimension corresponding to a human-interpretable objective (e.g., honesty, verbosity, safety); ii) employ a Mixture-of-Experts (MoE) strategy with a gating network that automatically selects the most suitable reward objectives based on the context. We efficiently trained an ArmoRM with Llama-3 8B and a gating network consisting of a shallow MLP on top of the ArmoRM. Our trained model, ArmoRM-Llama3-8B, obtains state-of-the-art performance on RewardBench, a benchmark evaluating RMs for language modeling. Notably, the performance of our model surpasses the LLM-as-a-judge method with GPT-4 judges by a margin, and approaches the performance of the much larger Nemotron-4 340B reward model.
Authors: Jin Yuan, Xuelan Qiu, Jinran Wu, Jiesi Guo, Weide Li, You-Gan Wang
Abstract: The interest in predicting online learning performance using ML algorithms has been steadily increasing. We first conducted a scientometric analysis to provide a systematic review of research in this area. The findings show that most existing studies apply the ML methods without considering learning behavior patterns, which may compromise the prediction accuracy and precision of the ML methods. This study proposes an integration framework that blends learning behavior analysis with ML algorithms to enhance the prediction accuracy of students' online learning performance. Specifically, the framework identifies distinct learning patterns among students by employing clustering analysis and implements various ML algorithms to predict performance within each pattern. For demonstration, the integration framework is applied to a real dataset from edX and distinguishes two learning patterns, as in, low autonomy students and motivated students. The results show that the framework yields nearly perfect prediction performance for autonomous students and satisfactory performance for motivated students. Additionally, this study compares the prediction performance of the integration framework to that of directly applying ML methods without learning behavior analysis using comprehensive evaluation metrics. The results consistently demonstrate the superiority of the integration framework over the direct approach, particularly when integrated with the best-performing XGBoosting method. Moreover, the framework significantly improves prediction accuracy for the motivated students and for the worst-performing random forest method. This study also evaluates the importance of various learning behaviors within each pattern using LightGBM with SHAP values. The implications of the integration framework and the results for online education practice and future research are discussed.
Authors: Raj Krishna Nepal, Bibek Khanal, Vibek Ghimire, Kismat Neupane, Atul Pokharel, Kshitij Niraula, Baburam Tiwari, Nawaraj Bhattarai, Khem N. Poudyal, Nawaraj Karki, Mohan B Dangi, John Biden
Abstract: Due to the unavailability of solar irradiance data for many potential sites of Nepal, the paper proposes predicting solar irradiance based on alternative meteorological parameters. The study focuses on five distinct regions in Nepal and utilizes a dataset spanning almost ten years, obtained from CERES SYN1deg and MERRA-2. Machine learning models such as Random Forest, XGBoost, K-Nearest Neighbors, and deep learning models like LSTM and ANN-MLP are employed and evaluated for their performance. The results indicate high accuracy in predicting solar irradiance, with R-squared(R2) scores close to unity for both train and test datasets. The impact of parameter integration on model performance is analyzed, revealing the significance of various parameters in enhancing predictive accuracy. Each model demonstrates strong performance across all parameters, consistently achieving MAE values below 6, RMSE values under 10, MBE within |2|, and nearly unity R2 values. Upon removal of various solar parameters such as "Solar_Irradiance_Clear_Sky", "UVA", etc. from the datasets, the model's performance is significantly affected. This exclusion leads to considerable increases in MAE, reaching up to 82, RMSE up to 135, and MBE up to |7|. Among the models, KNN displays the weakest performance, with an R2 of 0.7582546. Conversely, ANN exhibits the strongest performance, boasting an R2 value of 0.9245877. Hence, the study concludes that Artificial Neural Network (ANN) performs exceptionally well, showcasing its versatility even under sparse data parameter conditions.
Authors: Jason Martin, Kenneth Yeung
Abstract: Many Large Language Models (LLMs) and LLM-powered apps deployed today use some form of prompt filter or alignment to protect their integrity. However, these measures aren't foolproof. This paper introduces KROP, a prompt injection technique capable of obfuscating prompt injection attacks, rendering them virtually undetectable to most of these security measures.
Authors: Feini Huang, Shijie Jiang, Lu Li, Yongkun Zhang, Ye Zhang, Ruqing Zhang, Qingliang Li, Danxi Li, Wei Shangguan, Yongjiu Dai
Abstract: In recent years, artificial intelligence (AI) rapidly accelerated its influence and is expected to promote the development of Earth system science (ESS) if properly harnessed. In application of AI to ESS, a significant hurdle lies in the interpretability conundrum, an inherent problem of black-box nature arising from the complexity of AI algorithms. To address this, explainable AI (XAI) offers a set of powerful tools that make the models more transparent. The purpose of this review is twofold: First, to provide ESS scholars, especially newcomers, with a foundational understanding of XAI, serving as a primer to inspire future research advances; second, to encourage ESS professionals to embrace the benefits of AI, free from preconceived biases due to its lack of interpretability. We begin with elucidating the concept of XAI, along with typical methods. We then delve into a review of XAI applications in the ESS literature, highlighting the important role that XAI has played in facilitating communication with AI model decisions, improving model diagnosis, and uncovering scientific insights. We identify four significant challenges that XAI faces within the ESS, and propose solutions. Furthermore, we provide a comprehensive illustration of multifaceted perspectives. Given the unique challenges in ESS, an interpretable hybrid approach that seamlessly integrates AI with domain-specific knowledge appears to be a promising way to enhance the utility of AI in ESS. A visionary outlook for ESS envisions a harmonious blend where process-based models govern the known, AI models explore the unknown, and XAI bridges the gap by providing explanations.
Authors: Christian Anti\'c
Abstract: Neural-symbolic integration aims to combine the connectionist subsymbolic with the logical symbolic approach to artificial intelligence. In this paper, we first define the answer set semantics of (boolean) neural nets and then introduce from first principles a class of neural logic programs and show that nets and programs are equivalent.
Authors: Siwei Zhang, Xi Chen, Yun Xiong, Xixi Wu, Yao Zhang, Yongrui Fu, Yinglong Zhao, Jiawei Zhang
Abstract: Temporal Graph Networks (TGNs) have demonstrated their remarkable performance in modeling temporal interaction graphs. These works can generate temporal node representations by encoding the surrounding neighborhoods for the target node. However, an inherent limitation of existing TGNs is their reliance on fixed, hand-crafted rules for neighborhood encoding, overlooking the necessity for an adaptive and learnable neighborhood that can accommodate both personalization and temporal evolution across different timestamps. In this paper, we aim to enhance existing TGNs by introducing an adaptive neighborhood encoding mechanism. We present SEAN, a flexible plug-and-play model that can be seamlessly integrated with existing TGNs, effectively boosting their performance. To achieve this, we decompose the adaptive neighborhood encoding process into two phases: (i) representative neighbor selection, and (ii) temporal-aware neighborhood information aggregation. Specifically, we propose the Representative Neighbor Selector component, which automatically pinpoints the most important neighbors for the target node. It offers a tailored understanding of each node's unique surrounding context, facilitating personalization. Subsequently, we propose a Temporal-aware Aggregator, which synthesizes neighborhood aggregation by selectively determining the utilization of aggregation routes and decaying the outdated information, allowing our model to adaptively leverage both the contextually significant and current information during aggregation. We conduct extensive experiments by integrating SEAN into three representative TGNs, evaluating their performance on four public datasets and one financial benchmark dataset introduced in this paper. The results demonstrate that SEAN consistently leads to performance improvements across all models, achieving SOTA performance and exceptional robustness.
Authors: Ankur Nath, Alan Kuhnle
Abstract: Recently, there has been much work on the design of general heuristics for graph-based, combinatorial optimization problems via the incorporation of Graph Neural Networks (GNNs) to learn distribution-specific solution structures.However, there is a lack of consistency in the evaluation of these heuristics, in terms of the baselines and instances chosen, which makes it difficult to assess the relative performance of the algorithms. In this paper, we propose an open-source benchmark suite MaxCut-Bench dedicated to the NP-hard Maximum Cut problem in both its weighted and unweighted variants, based on a careful selection of instances curated from diverse graph datasets. The suite offers a unified interface to various heuristics, both traditional and machine learning-based. Next, we use the benchmark in an attempt to systematically corroborate or reproduce the results of several, popular learning-based approaches, including S2V-DQN [31], ECO-DQN [4], among others, in terms of three dimensions: objective value, generalization, and scalability. Our empirical results show that several of the learned heuristics fail to outperform a naive greedy algorithm, and that only one of them consistently outperforms Tabu Search, a simple, general heuristic based upon local search. Furthermore, we find that the performance of ECO-DQN remains the same or is improved if the GNN is replaced by a simple linear regression on a subset of the features that are related to Tabu Search. Code, data, and pretrained models are available at: \url{https://github.com/ankurnath/MaxCut-Bench}.
Authors: Harry Shomer, Jay Revolinsky, Jiliang Tang
Abstract: Knowledge Graph Completion (KGC) attempts to predict missing facts in a Knowledge Graph (KG). Recently, there's been an increased focus on designing KGC methods that can excel in the {\it inductive setting}, where a portion or all of the entities and relations seen in inference are unobserved during training. Numerous benchmark datasets have been proposed for inductive KGC, all of which are subsets of existing KGs used for transductive KGC. However, we find that the current procedure for constructing inductive KGC datasets inadvertently creates a shortcut that can be exploited even while disregarding the relational information. Specifically, we observe that the Personalized PageRank (PPR) score can achieve strong or near SOTA performance on most inductive datasets. In this paper, we study the root cause of this problem. Using these insights, we propose an alternative strategy for constructing inductive KGC datasets that helps mitigate the PPR shortcut. We then benchmark multiple popular methods using the newly constructed datasets and analyze their performance. The new benchmark datasets help promote a better understanding of the capabilities and challenges of inductive KGC by removing any shortcuts that obfuscate performance.
Authors: Qi-Jie Li, Qian Sun, Shao-Qun Zhang
Abstract: Identifying gene splicing is a core and significant task confronted in modern collaboration between artificial intelligence and bioinformatics. Past decades have witnessed great efforts on this concern, such as the bio-plausible splicing pattern AT-CG and the famous SpliceAI. In this paper, we propose a novel framework for the task of gene splicing identification, named Horizon-wise Gene Splicing Identification (H-GSI). The proposed H-GSI follows the horizon-wise identification paradigm and comprises four components: the pre-processing procedure transforming string data into tensors, the sliding window technique handling long sequences, the SeqLab model, and the predictor. In contrast to existing studies that process gene information with a truncated fixed-length sequence, H-GSI employs a horizon-wise identification paradigm in which all positions in a sequence are predicted with only one forward computation, improving accuracy and efficiency. The experiments conducted on the real-world Human dataset show that our proposed H-GSI outperforms SpliceAI and achieves the best accuracy of 97.20\%. The source code is available from this link.
Authors: Alireza Rashnu, Sadegh Aliakbary
Abstract: Modeling complex networks allows us to analyze the characteristics and discover the basic mechanisms governing phenomena such as disease outbreaks, information diffusion, transportation efficiency, social influence, and even human brain function. Consequently, various network generative models (called temporal network models) have been presented to model how the network topologies evolve dynamically over time. Temporal network models face the challenge of results evaluation because common evaluation methods are appropriate only for static networks. This paper proposes an automatic approach based on deep learning to handle this issue. In addition to an evaluation method, the proposed method can also be used for anomaly detection in evolving networks. The proposed method has been evaluated on five different datasets, and the evaluations show that it outperforms the alternative methods based on the error rate measure in different datasets.
Authors: Silvia Sapora, Gokul Swamy, Chris Lu, Yee Whye Teh, Jakob Nicolaus Foerster
Abstract: Often times in imitation learning (IL), the environment we collect expert demonstrations in and the environment we want to deploy our learned policy in aren't exactly the same (e.g. demonstrations collected in simulation but deployment in the real world). Compared to policy-centric approaches to IL like behavioural cloning, reward-centric approaches like inverse reinforcement learning (IRL) often better replicate expert behaviour in new environments. This transfer is usually performed by optimising the recovered reward under the dynamics of the target environment. However, (a) we find that modern deep IL algorithms frequently recover rewards which induce policies far weaker than the expert, even in the same environment the demonstrations were collected in. Furthermore, (b) these rewards are often quite poorly shaped, necessitating extensive environment interaction to optimise effectively. We provide simple and scalable fixes to both of these concerns. For (a), we find that reward model ensembles combined with a slightly different training objective significantly improves re-training and transfer performance. For (b), we propose a novel evolution-strategies based method EvIL to optimise for a reward-shaping term that speeds up re-training in the target environment, closing a gap left open by the classical theory of IRL. On a suite of continuous control tasks, we are able to re-train policies in target (and source) environments more interaction-efficiently than prior work.
Authors: X. Angelo Huang, Emanuele La Malfa, Samuele Marro, Andrea Asperti, Anthony Cohn, Michael Wooldridge
Abstract: Theory of Mind (ToM) can be used to assess the capabilities of Large Language Models (LLMs) in complex scenarios where social reasoning is required. While the research community has proposed many ToM benchmarks, their hardness varies greatly, and their complexity is not well defined. This work proposes a framework to measure the complexity of ToM tasks. We quantify a problem's complexity as the number of states necessary to solve it correctly. Our complexity measure also accounts for spurious states of a ToM problem designed to make it apparently harder. We use our method to assess the complexity of five widely adopted ToM benchmarks. On top of this framework, we design a prompting technique that augments the information available to a model with a description of how the environment changes with the agents' interactions. We name this technique Discrete World Models (DWM) and show how it elicits superior performance on ToM tasks.
Authors: Evan Lohn, Sean Welleck
Abstract: Neural networks have shown initial promise in automating mathematical theorem proving in proof assistants such as Lean. The same proof assistants can be used to verify the correctness of code by pairing code with specifications and proofs that the specifications hold. Automating the writing of code, specifications, and proofs could lower the cost of verification, or, ambitiously, enable a machine learning system to output provably correct code. However, it remains unclear whether current neural theorem provers can automatically verify even relatively simple programs. We present miniCodeProps, a benchmark of 177 program specifications in the Lean proof assistant, aimed at the subproblem of automatically generating a proof for a provided program and specification. miniCodeProps contains specifications about simple, self-contained programs (e.g., lists, natural numbers, binary trees) with varied proof difficulty. Despite its simplicity, miniCodeProps is challenging for current LLM-based provers, which succeed in proving about 25 percent of the specifications. We publicly release miniCodeProps as a benchmark for furthering automated theorem proving in the context of formally verified code.
Authors: Silvia Garc\'ia-M\'endez, Francisco de Arriba-P\'erez, Jaime Gonz\'alez-Gonz\'aleza, Francisco J. Gonz\'alez-Casta\~no
Abstract: Social media include diverse interaction metrics related to user popularity, the most evident example being the number of user followers. The latter has raised concerns about the credibility of the posts by the most popular creators. However, most existing approaches to assess credibility in social media strictly consider this problem a binary classification, often based on a priori information, without checking if actual real-world facts back the users' comments. In addition, they do not provide automatic explanations of their predictions to foster their trustworthiness. In this work, we propose a credibility assessment solution for financial creators in social media that combines Natural Language Processing and Machine Learning. The reputation of the contributors is assessed by automatically classifying their forecasts on asset values by type and verifying these predictions with actual market data to approximate their probability of success. The outcome of this verification is a continuous credibility score instead of a binary result, an entirely novel contribution by this work. Moreover, social media metrics (i.e., user context) are exploited by calculating their correlation with the credibility rankings, providing insights on the interest of the end-users in financial posts and their forecasts (i.e., drop or rise). Finally, the system provides natural language explanations of its decisions based on a model-agnostic analysis of relevant features.
Authors: DeepSeek-AI, Qihao Zhu, Daya Guo, Zhihong Shao, Dejian Yang, Peiyi Wang, Runxin Xu, Y. Wu, Yukun Li, Huazuo Gao, Shirong Ma, Wangding Zeng, Xiao Bi, Zihui Gu, Hanwei Xu, Damai Dai, Kai Dong, Liyue Zhang, Yishi Piao, Zhibin Gou, Zhenda Xie, Zhewen Hao, Bingxuan Wang, Junxiao Song, Deli Chen, Xin Xie, Kang Guan, Yuxiang You, Aixin Liu, Qiushi Du, Wenjun Gao, Xuan Lu, Qinyu Chen, Yaohui Wang, Chengqi Deng, Jiashi Li, Chenggang Zhao, Chong Ruan, Fuli Luo, Wenfeng Liang
Abstract: We present DeepSeek-Coder-V2, an open-source Mixture-of-Experts (MoE) code language model that achieves performance comparable to GPT4-Turbo in code-specific tasks. Specifically, DeepSeek-Coder-V2 is further pre-trained from an intermediate checkpoint of DeepSeek-V2 with additional 6 trillion tokens. Through this continued pre-training, DeepSeek-Coder-V2 substantially enhances the coding and mathematical reasoning capabilities of DeepSeek-V2, while maintaining comparable performance in general language tasks. Compared to DeepSeek-Coder-33B, DeepSeek-Coder-V2 demonstrates significant advancements in various aspects of code-related tasks, as well as reasoning and general capabilities. Additionally, DeepSeek-Coder-V2 expands its support for programming languages from 86 to 338, while extending the context length from 16K to 128K. In standard benchmark evaluations, DeepSeek-Coder-V2 achieves superior performance compared to closed-source models such as GPT4-Turbo, Claude 3 Opus, and Gemini 1.5 Pro in coding and math benchmarks.
Authors: Kenneth Li, Yiming Wang, Fernanda Vi\'egas, Martin Wattenberg
Abstract: We present an approach called Dialogue Action Tokens (DAT) that adapts language model agents to plan goal-directed dialogues. The core idea is to treat each utterance as an action, thereby converting dialogues into games where existing approaches such as reinforcement learning can be applied. Specifically, we freeze a pretrained language model and train a small planner model that predicts a continuous action vector, used for controlled generation in each round. This design avoids the problem of language degradation under reward optimization. When evaluated on the Sotopia platform for social simulations, the DAT-steered LLaMA model surpasses GPT-4's performance. We also apply DAT to steer an attacker language model in a novel multi-turn red-teaming setting, revealing a potential new attack surface.
Authors: Abhishek Sureddy, Dishant Padalia, Nandhinee Periyakaruppa, Oindrila Saha, Adina Williams, Adriana Romero-Soriano, Megan Richards, Polina Kirichenko, Melissa Hall
Abstract: Recent work has identified substantial disparities in generated images of different geographic regions, including stereotypical depictions of everyday objects like houses and cars. However, existing measures for these disparities have been limited to either human evaluations, which are time-consuming and costly, or automatic metrics evaluating full images, which are unable to attribute these disparities to specific parts of the generated images. In this work, we introduce a new set of metrics, Decomposed Indicators of Disparities in Image Generation (Decomposed-DIG), that allows us to separately measure geographic disparities in the depiction of objects and backgrounds in generated images. Using Decomposed-DIG, we audit a widely used latent diffusion model and find that generated images depict objects with better realism than backgrounds and that backgrounds in generated images tend to contain larger regional disparities than objects. We use Decomposed-DIG to pinpoint specific examples of disparities, such as stereotypical background generation in Africa, struggling to generate modern vehicles in Africa, and unrealistically placing some objects in outdoor settings. Informed by our metric, we use a new prompting structure that enables a 52% worst-region improvement and a 20% average improvement in generated background diversity.
Authors: Ning Jiang, Weiqi Chu, Yao Li
Abstract: Classical compartmental models in epidemiology often struggle to accurately capture real-world dynamics due to their inability to address the inherent heterogeneity of populations. In this paper, we introduce a novel approach that incorporates heterogeneity through a mobility variable, transforming the traditional ODE system into a system of integro-differential equations that describe the dynamics of population densities across different compartments. Our results show that, for the same basic reproduction number, our mobility-based model predicts a smaller final pandemic size compared to classic compartmental models, whose population densities are represented as Dirac delta functions in our density-based framework. This addresses the overestimation issue common in many classical models. Additionally, we demonstrate that the time series of the infected population is sufficient to uniquely identify the mobility distribution. We reconstruct this distribution using a machine-learning-based framework, providing both theoretical and algorithmic support to effectively constrain the mobility-based model with real-world data.
Authors: Romina Yalovetzky, Niran Kumar, Changhao Li, Marco Pistoia
Abstract: Random Forest (RF) is a popular tree-ensemble method for supervised learning, prized for its ease of use and flexibility. Online RF models require to account for new training data to maintain model accuracy. This is particularly important in applications were data is periodically and sequentially generated over time in data streams, such as auto-driving systems, and credit card payments. In this setting, performing periodic model retraining with the old and new data accumulated is beneficial as it fully captures possible drifts in the data distribution over time. However, this is unpractical with state-of-the-art classical algorithms for RF as they scale linearly with the accumulated number of samples. We propose QC-Forest, a classical-quantum algorithm designed to time-efficiently retrain RF models in the streaming setting for multi-class classification and regression, achieving a runtime poly-logarithmic in the total number of accumulated samples. QC-Forest leverages Des-q, a quantum algorithm for single tree construction and retraining proposed by Kumar et al. by expanding to multi-class classification, as the original proposal was limited to binary classes, and introducing an exact classical method to replace an underlying quantum subroutine incurring a finite error, while maintaining the same poly-logarithmic dependence. Finally, we showcase that QC-Forest achieves competitive accuracy in comparison to state-of-the-art RF methods on widely used benchmark datasets with up to 80,000 samples, while significantly speeding up the model retrain.
Authors: Zezhi Wang, Jin Zhu, Junxian Zhu, Borui Tang, Hongmei Lin, Xueqin Wang
Abstract: Sparsity-constraint optimization has wide applicability in signal processing, statistics, and machine learning. Existing fast algorithms must burdensomely tune parameters, such as the step size or the implementation of precise stop criteria, which may be challenging to determine in practice. To address this issue, we develop an algorithm named Sparsity-Constraint Optimization via sPlicing itEration (SCOPE) to optimize nonlinear differential objective functions with strong convexity and smoothness in low dimensional subspaces. Algorithmically, the SCOPE algorithm converges effectively without tuning parameters. Theoretically, SCOPE has a linear convergence rate and converges to a solution that recovers the true support set when it correctly specifies the sparsity. We also develop parallel theoretical results without restricted-isometry-property-type conditions. We apply SCOPE's versatility and power to solve sparse quadratic optimization, learn sparse classifiers, and recover sparse Markov networks for binary variables. The numerical results on these specific tasks reveal that SCOPE perfectly identifies the true support set with a 10--1000 speedup over the standard exact solver, confirming SCOPE's algorithmic and theoretical merits. Our open-source Python package skscope based on C++ implementation is publicly available on GitHub, reaching a ten-fold speedup on the competing convex relaxation methods implemented by the cvxpy library.
Authors: Christian Herold, Michael Kozielski, Leonid Ekimov, Pavel Petrushkov, Pierre-Yves Vandenbussche, Shahram Khadivi
Abstract: We introduce the LiLiuM series of large language models (LLMs): 1B, 7B, and 13B parameter models developed 100% in-house to fit eBay's specific needs in the e-commerce domain. This gives eBay full control over all aspects of the models including license, data, vocabulary, and architecture. We expect these models to be used as a foundation for fine-tuning and instruction-tuning, eliminating dependencies to external models. The LiLiuM LLMs have been trained on 3 trillion tokens of multilingual text from general and e-commerce domain. They perform similar to the popular LLaMA-2 models on English natural language understanding (NLU) benchmarks. At the same time, we outperform LLaMA-2 on non-English NLU tasks, machine translation and on e-commerce specific downstream tasks. As part of our data mixture, we utilize the newly released RedPajama-V2 dataset for training and share our insights regarding data filtering and deduplication. We also discuss in detail how to serialize structured data for use in autoregressive language modeling. We provide insights on the effects of including code and parallel machine translation data in pre-training. Furthermore, we develop our own tokenizer and model vocabulary, customized towards e-commerce. This way, we can achieve up to 34% speed-up in text generation on eBay-specific downstream tasks compared to LLaMA-2. Finally, in relation to LLM pretraining, we show that checkpoint averaging can further improve over the best individual model checkpoint.
Authors: Junmo Kang, Leonid Karlinsky, Hongyin Luo, Zhen Wang, Jacob Hansen, James Glass, David Cox, Rameswar Panda, Rogerio Feris, Alan Ritter
Abstract: We present Self-MoE, an approach that transforms a monolithic LLM into a compositional, modular system of self-specialized experts, named MiXSE (MiXture of Self-specialized Experts). Our approach leverages self-specialization, which constructs expert modules using self-generated synthetic data, each equipped with a shared base LLM and incorporating self-optimized routing. This allows for dynamic and capability-specific handling of various target tasks, enhancing overall capabilities, without extensive human-labeled data and added parameters. Our empirical results reveal that specializing LLMs may exhibit potential trade-offs in performances on non-specialized tasks. On the other hand, our Self-MoE demonstrates substantial improvements over the base LLM across diverse benchmarks such as knowledge, reasoning, math, and coding. It also consistently outperforms other methods, including instance merging and weight merging, while offering better flexibility and interpretability by design with semantic experts and routing. Our findings highlight the critical role of modularity and the potential of self-improvement in achieving efficient, scalable, and adaptable systems.
Authors: Karuna Bhaila, Minh-Hao Van, Xintao Wu
Abstract: The widespread popularity of Large Language Models (LLMs), partly due to their unique ability to perform in-context learning, has also brought to light the importance of ethical and safety considerations when deploying these pre-trained models. In this work, we focus on investigating machine unlearning for LLMs motivated by data protection regulations. In contrast to the growing literature on fine-tuning methods to achieve unlearning, we focus on a comparatively lightweight alternative called soft prompting to realize the unlearning of a subset of training data. With losses designed to enforce forgetting as well as utility preservation, our framework \textbf{S}oft \textbf{P}rompting for \textbf{U}n\textbf{l}earning (SPUL) learns prompt tokens that can be appended to an arbitrary query to induce unlearning of specific examples at inference time without updating LLM parameters. We conduct a rigorous evaluation of the proposed method and our results indicate that SPUL can significantly improve the trade-off between utility and forgetting in the context of text classification with LLMs. We further validate our method using multiple LLMs to highlight the scalability of our framework and provide detailed insights into the choice of hyperparameters and the influence of the size of unlearning data. Our implementation is available at \url{https://github.com/karuna-bhaila/llm_unlearning}.
Authors: Alireza Ganjdanesh, Reza Shirkavand, Shangqian Gao, Heng Huang
Abstract: Text-to-image (T2I) diffusion models have demonstrated impressive image generation capabilities. Still, their computational intensity prohibits resource-constrained organizations from deploying T2I models after fine-tuning them on their internal target data. While pruning techniques offer a potential solution to reduce the computational burden of T2I models, static pruning methods use the same pruned model for all input prompts, overlooking the varying capacity requirements of different prompts. Dynamic pruning addresses this issue by utilizing a separate sub-network for each prompt, but it prevents batch parallelism on GPUs. To overcome these limitations, we introduce Adaptive Prompt-Tailored Pruning (APTP), a novel prompt-based pruning method designed for T2I diffusion models. Central to our approach is a prompt router model, which learns to determine the required capacity for an input text prompt and routes it to an architecture code, given a total desired compute budget for prompts. Each architecture code represents a specialized model tailored to the prompts assigned to it, and the number of codes is a hyperparameter. We train the prompt router and architecture codes using contrastive learning, ensuring that similar prompts are mapped to nearby codes. Further, we employ optimal transport to prevent the codes from collapsing into a single one. We demonstrate APTP's effectiveness by pruning Stable Diffusion (SD) V2.1 using CC3M and COCO as target datasets. APTP outperforms the single-model pruning baselines in terms of FID, CLIP, and CMMD scores. Our analysis of the clusters learned by APTP reveals they are semantically meaningful. We also show that APTP can automatically discover previously empirically found challenging prompts for SD, e.g., prompts for generating text images, assigning them to higher capacity codes.
Authors: Yi Fang, Dongzhe Fan, Sirui Ding, Ninghao Liu, Qiaoyu Tan
Abstract: Representation learning on text-attributed graphs (TAGs), where nodes are represented by textual descriptions, is crucial for textual and relational knowledge systems and recommendation systems. Currently, state-of-the-art embedding methods for TAGs primarily focus on fine-tuning language models (e.g., BERT) using structure-aware training signals. While effective, these methods are tailored for individual TAG and cannot generalize across various graph scenarios. Given the shared textual space, leveraging multiple TAGs for joint fine-tuning, aligning text and graph structure from different aspects, would be more beneficial. Motivated by this, we introduce a novel Unified Graph Language Model (UniGLM) framework, the first graph embedding model that generalizes well to both in-domain and cross-domain TAGs. Specifically, UniGLM is trained over multiple TAGs with different domains and scales using self-supervised contrastive learning. UniGLM includes an adaptive positive sample selection technique for identifying structurally similar nodes and a lazy contrastive module that is devised to accelerate training by minimizing repetitive encoding calculations. Extensive empirical results across 9 benchmark TAGs demonstrate UniGLM's efficacy against leading embedding baselines in terms of generalization (various downstream tasks and backbones) and transfer learning (in and out of domain scenarios). The code is available at https://github.com/NYUSHCS/UniGLM.
Authors: Anna Sokolova, Anna Vorontsova, Bulat Gabdullin, Alexander Limonov
Abstract: Leveraging 3D semantics for direct 3D reconstruction has a great potential yet unleashed. For instance, by assuming that walls are vertical, and a floor is planar and horizontal, we can correct distorted room shapes and eliminate local artifacts such as holes, pits, and hills. In this paper, we propose FAWN, a modification of truncated signed distance function (TSDF) reconstruction methods, which considers scene structure by detecting walls and floor in a scene, and penalizing the corresponding surface normals for deviating from the horizontal and vertical directions. Implemented as a 3D sparse convolutional module, FAWN can be incorporated into any trainable pipeline that predicts TSDF. Since FAWN requires 3D semantics only for training, no additional limitations on further use are imposed. We demonstrate, that FAWN-modified methods use semantics more effectively, than existing semantic-based approaches. Besides, we apply our modification to state-of-the-art TSDF reconstruction methods, and demonstrate a quality gain in SCANNET, ICL-NUIM, TUM RGB-D, and 7SCENES benchmarks.
Authors: Ukyo Honda, Tatsushi Oka, Peinan Zhang, Masato Mita
Abstract: Recent models for natural language understanding are inclined to exploit simple patterns in datasets, commonly known as shortcuts. These shortcuts hinge on spurious correlations between labels and latent features existing in the training data. At inference time, shortcut-dependent models are likely to generate erroneous predictions under distribution shifts, particularly when some latent features are no longer correlated with the labels. To avoid this, previous studies have trained models to eliminate the reliance on shortcuts. In this study, we explore a different direction: pessimistically aggregating the predictions of a mixture-of-experts, assuming each expert captures relatively different latent features. The experimental results demonstrate that our post-hoc control over the experts significantly enhances the model's robustness to the distribution shift in shortcuts. Besides, we show that our approach has some practical advantages. We also analyze our model and provide results to support the assumption.
Authors: Christopher W. Curtis, Erik Bollt, Daniel Jay Alford-Lago
Abstract: In this work, we present a method which determines optimal multi-step dynamic mode decomposition (DMD) models via entropic regression, which is a nonlinear information flow detection algorithm. Motivated by the higher-order DMD (HODMD) method of \cite{clainche}, and the entropic regression (ER) technique for network detection and model construction found in \cite{bollt, bollt2}, we develop a method that we call ERDMD that produces high fidelity time-delay DMD models that allow for nonuniform time space, and the time spacing is discovered by consider most informativity based on ER. These models are shown to be highly efficient and robust. We test our method over several data sets generated by chaotic attractors and show that we are able to build excellent reconstructions using relatively minimal models. We likewise are able to better identify multiscale features via our models which enhances the utility of dynamic mode decomposition.
Authors: Jiasheng Zhang, Jialin Chen, Menglin Yang, Aosong Feng, Shuang Liang, Jie Shao, Rex Ying
Abstract: Dynamic text-attributed graphs (DyTAGs) are prevalent in various real-world scenarios, where each node and edge are associated with text descriptions, and both the graph structure and text descriptions evolve over time. Despite their broad applicability, there is a notable scarcity of benchmark datasets tailored to DyTAGs, which hinders the potential advancement in many research fields. To address this gap, we introduce Dynamic Text-attributed Graph Benchmark (DTGB), a collection of large-scale, time-evolving graphs from diverse domains, with nodes and edges enriched by dynamically changing text attributes and categories. To facilitate the use of DTGB, we design standardized evaluation procedures based on four real-world use cases: future link prediction, destination node retrieval, edge classification, and textual relation generation. These tasks require models to understand both dynamic graph structures and natural language, highlighting the unique challenges posed by DyTAGs. Moreover, we conduct extensive benchmark experiments on DTGB, evaluating 7 popular dynamic graph learning algorithms and their variants of adapting to text attributes with LLM embeddings, along with 6 powerful large language models (LLMs). Our results show the limitations of existing models in handling DyTAGs. Our analysis also demonstrates the utility of DTGB in investigating the incorporation of structural and textual dynamics. The proposed DTGB fosters research on DyTAGs and their broad applications. It offers a comprehensive benchmark for evaluating and advancing models to handle the interplay between dynamic graph structures and natural language. The dataset and source code are available at https://github.com/zjs123/DTGB.
Authors: Xinglong Sun, Barath Lakshmanan, Maying Shen, Shiyi Lan, Jingde Chen, Jose Alvarez
Abstract: As we push the boundaries of performance in various vision tasks, the models grow in size correspondingly. To keep up with this growth, we need very aggressive pruning techniques for efficient inference and deployment on edge devices. Existing pruning approaches are limited to channel pruning and struggle with aggressive parameter reductions. In this paper, we propose a novel multi-dimensional pruning framework that jointly optimizes pruning across channels, layers, and blocks while adhering to latency constraints. We develop a latency modeling technique that accurately captures model-wide latency variations during pruning, which is crucial for achieving an optimal latency-accuracy trade-offs at high pruning ratio. We reformulate pruning as a Mixed-Integer Nonlinear Program (MINLP) to efficiently determine the optimal pruned structure with only a single pass. Our extensive results demonstrate substantial improvements over previous methods, particularly at large pruning ratios. In classification, our method significantly outperforms prior art HALP with a Top-1 accuracy of 70.0(v.s. 68.6) and an FPS of 5262 im/s(v.s. 4101 im/s). In 3D object detection, we establish a new state-of-the-art by pruning StreamPETR at a 45% pruning ratio, achieving higher FPS (37.3 vs. 31.7) and mAP (0.451 vs. 0.449) than the dense baseline.
Authors: Anna Susmelj, Mael Macuglia, Nata\v{s}a Tagasovska, Reto Sutter, Sebastiano Caprara, Jean-Philippe Thiran, Ender Konukoglu
Abstract: Implicit functions such as Neural Radiance Fields (NeRFs), occupancy networks, and signed distance functions (SDFs) have become pivotal in computer vision for reconstructing detailed object shapes from sparse views. Achieving optimal performance with these models can be challenging due to the extreme sparsity of inputs and distribution shifts induced by data corruptions. To this end, large, noise-free synthetic datasets can serve as shape priors to help models fill in gaps, but the resulting reconstructions must be approached with caution. Uncertainty estimation is crucial for assessing the quality of these reconstructions, particularly in identifying areas where the model is uncertain about the parts it has inferred from the prior. In this paper, we introduce Dropsembles, a novel method for uncertainty estimation in tuned implicit functions. We demonstrate the efficacy of our approach through a series of experiments, starting with toy examples and progressing to a real-world scenario. Specifically, we train a Convolutional Occupancy Network on synthetic anatomical data and test it on low-resolution MRI segmentations of the lumbar spine. Our results show that Dropsembles achieve the accuracy and calibration levels of deep ensembles but with significantly less computational cost.
Authors: Ibrahim Can Yilmaz, Said Aldemir
Abstract: Click-Through Rate (CTR) prediction has become an essential task in digital industries, such as digital advertising or online shopping. Many deep learning-based methods have been implemented and have become state-of-the-art models in the domain. To further improve the performance of CTR models, Knowledge Distillation based approaches have been widely used. However, most of the current CTR prediction models do not have much complex architectures, so it's hard to call one of them 'cumbersome' and the other one 'tiny'. On the other hand, the performance gap is also not very large between complex and simple models. So, distilling knowledge from one model to the other could not be worth the effort. Under these considerations, Mutual Learning could be a better approach, since all the models could be improved mutually. In this paper, we showed how useful the mutual learning algorithm could be when it is between equals. In our experiments on the Criteo and Avazu datasets, the mutual learning algorithm improved the performance of the model by up to 0.66% relative improvement.
Authors: Karime Maamari, Amine Mhedhbi
Abstract: Recent advancements in Text-to-SQL have pushed database management systems towards greater democratization of data access. Today's language models are at the core of these advancements. They enable impressive Text-to-SQL generation as experienced in the development of Distyl AI's Analytics Insight Engine. Its early deployment with enterprise customers has highlighted three core challenges. First, data analysts expect support with authoring SQL queries of very high complexity. Second, requests are ad-hoc and, as such, require low latency. Finally, generation requires an understanding of domain-specific terminology and practices. The design and implementation of our Text-to-SQL generation pipeline, powered by large language models, tackles these challenges. The core tenants of our approach rely on external knowledge that we extract in a pre-processing phase, on retrieving the appropriate external knowledge at query generation time, and on decomposing SQL query generation following a hierarchical CTE-based structure. Finally, an adaptation framework leverages feedback to update the external knowledge, in turn improving query generation over time. We give an overview of our end-to-end approach and highlight the operators generating SQL during inference.
Authors: Emily Shinkle, Aleksandra Pachalieva, Riti Bahl, Sakib Matin, Brendan Gifford, Galen T. Craven, Nicholas Lubbers
Abstract: Coarse-graining is a molecular modeling technique in which an atomistic system is represented in a simplified fashion that retains the most significant system features that contribute to a target output, while removing the degrees of freedom that are less relevant. This reduction in model complexity allows coarse-grained molecular simulations to reach increased spatial and temporal scales compared to corresponding all-atom models. A core challenge in coarse-graining is to construct a force field that represents the interactions in the new representation in a way that preserves the atomistic-level properties. Many approaches to building coarse-grained force fields have limited transferability between different thermodynamic conditions as a result of averaging over internal fluctuations at a specific thermodynamic state point. Here, we use a graph-convolutional neural network architecture, the Hierarchically Interacting Particle Neural Network with Tensor Sensitivity (HIP-NN-TS), to develop a highly automated training pipeline for coarse grained force fields which allows for studying the transferability of coarse-grained models based on the force-matching approach. We show that this approach not only yields highly accurate force fields, but also that these force fields are more transferable through a variety of thermodynamic conditions. These results illustrate the potential of machine learning techniques such as graph neural networks to improve the construction of transferable coarse-grained force fields.
Authors: Hamidreza Rouzegar, Masoud Makrehchi
Abstract: In the context of text classification, the financial burden of annotation exercises for creating training data is a critical issue. Active learning techniques, particularly those rooted in uncertainty sampling, offer a cost-effective solution by pinpointing the most instructive samples for manual annotation. Similarly, Large Language Models (LLMs) such as GPT-3.5 provide an alternative for automated annotation but come with concerns regarding their reliability. This study introduces a novel methodology that integrates human annotators and LLMs within an Active Learning framework. We conducted evaluations on three public datasets. IMDB for sentiment analysis, a Fake News dataset for authenticity discernment, and a Movie Genres dataset for multi-label classification.The proposed framework integrates human annotation with the output of LLMs, depending on the model uncertainty levels. This strategy achieves an optimal balance between cost efficiency and classification performance. The empirical results show a substantial decrease in the costs associated with data annotation while either maintaining or improving model accuracy.
Authors: Jingxi Xu, Runsheng Wang, Siqi Shang, Ava Chen, Lauren Winterbottom, To-Liang Hsu, Wenxi Chen, Khondoker Ahmed, Pedro Leandro La Rotta, Xinyue Zhu, Dawn M. Nilsen, Joel Stein, Matei Ciocarlie
Abstract: Intent inferral on a hand orthosis for stroke patients is challenging due to the difficulty of data collection from impaired subjects. Additionally, EMG signals exhibit significant variations across different conditions, sessions, and subjects, making it hard for classifiers to generalize. Traditional approaches require a large labeled dataset from the new condition, session, or subject to train intent classifiers; however, this data collection process is burdensome and time-consuming. In this paper, we propose ChatEMG, an autoregressive generative model that can generate synthetic EMG signals conditioned on prompts (i.e., a given sequence of EMG signals). ChatEMG enables us to collect only a small dataset from the new condition, session, or subject and expand it with synthetic samples conditioned on prompts from this new context. ChatEMG leverages a vast repository of previous data via generative training while still remaining context-specific via prompting. Our experiments show that these synthetic samples are classifier-agnostic and can improve intent inferral accuracy for different types of classifiers. We demonstrate that our complete approach can be integrated into a single patient session, including the use of the classifier for functional orthosis-assisted tasks. To the best of our knowledge, this is the first time an intent classifier trained partially on synthetic data has been deployed for functional control of an orthosis by a stroke survivor. Videos and additional information can be found at https://jxu.ai/chatemg.
URLs: https://jxu.ai/chatemg.
Authors: Xinrui Zu, Qian Tao
Abstract: Diffusion models have demonstrated strong performance in sampling and editing multi-modal data with high generation quality, yet they suffer from the iterative generation process which is computationally expensive and slow. In addition, most methods are constrained to generate data from Gaussian noise, which limits their sampling and editing flexibility. To overcome both disadvantages, we present Contrastive Optimal Transport Flow (COT Flow), a new method that achieves fast and high-quality generation with improved zero-shot editing flexibility compared to previous diffusion models. Benefiting from optimal transport (OT), our method has no limitation on the prior distribution, enabling unpaired image-to-image (I2I) translation and doubling the editable space (at both the start and end of the trajectory) compared to other zero-shot editing methods. In terms of quality, COT Flow can generate competitive results in merely one step compared to previous state-of-the-art unpaired image-to-image (I2I) translation methods. To highlight the advantages of COT Flow through the introduction of OT, we introduce the COT Editor to perform user-guided editing with excellent flexibility and quality. The code will be released at https://github.com/zuxinrui/cot_flow.
Authors: Anna C. Marbut, John W. Chandler, Travis J. Wheeler
Abstract: It is generally thought that transformer-based large language models benefit from pre-training by learning generic linguistic knowledge that can be focused on a specific task during fine-tuning. However, we propose that much of the benefit from pre-training may be captured by geometric characteristics of the latent space representations, divorced from any specific linguistic knowledge. In this work we explore the relationship between GLUE benchmarking task performance and a variety of measures applied to the latent space resulting from BERT-type contextual language models. We find that there is a strong linear relationship between a measure of quantized cell density and average GLUE performance and that these measures may be predictive of otherwise surprising GLUE performance for several non-standard BERT-type models from the literature. These results may be suggestive of a strategy for decreasing pre-training requirements, wherein model initialization can be informed by the geometric characteristics of the model's latent space.
Authors: Alex Chen, Nathan Lay, Stephanie Harmon, Kutsev Ozyoruk, Enis Yilmaz, Brad J. Wood, Peter A. Pinto, Peter L. Choyke, Baris Turkbey
Abstract: Prostate cancer is one of the most prevalent malignancies in the world. While deep learning has potential to further improve computer-aided prostate cancer detection on MRI, its efficacy hinges on the exhaustive curation of manually annotated images. We propose a novel methodology of semisupervised learning (SSL) guided by automatically extracted clinical information, specifically the lesion locations in radiology reports, allowing for use of unannotated images to reduce the annotation burden. By leveraging lesion locations, we refined pseudo labels, which were then used to train our location-based SSL model. We show that our SSL method can improve prostate lesion detection by utilizing unannotated images, with more substantial impacts being observed when larger proportions of unannotated images are used.
Authors: Z. T. Wang, Qiuhao Chen, Yuxuan Du, Z. H. Yang, Xiaoxia Cai, Kaixuan Huang, Jingning Zhang, Kai Xu, Jun Du, Yinan Li, Yuling Jiao, Xingyao Wu, Wu Liu, Xiliang Lu, Huikai Xu, Yirong Jin, Ruixia Wang, Haifeng Yu, S. P. Zhao
Abstract: To effectively implement quantum algorithms on noisy intermediate-scale quantum (NISQ) processors is a central task in modern quantum technology. NISQ processors feature tens to a few hundreds of noisy qubits with limited coherence times and gate operations with errors, so NISQ algorithms naturally require employing circuits of short lengths via quantum compilation. Here, we develop a reinforcement learning (RL)-based quantum compiler for a superconducting processor and demonstrate its capability of discovering novel and hardware-amenable circuits with short lengths. We show that for the three-qubit quantum Fourier transformation, a compiled circuit using only seven CZ gates with unity circuit fidelity can be achieved. The compiler is also able to find optimal circuits under device topological constraints, with lengths considerably shorter than those by the conventional method. Our study exemplifies the codesign of the software with hardware for efficient quantum compilation, offering valuable insights for the advancement of RL-based compilers.
Authors: Alex G. Zalles, Kai M. Hung, Ann E. Finneran, Lydia Beaudrot, C\'esar A. Uribe
Abstract: We study the problem of network regression, where one is interested in how the topology of a network changes as a function of Euclidean covariates. We build upon recent developments in generalized regression models on metric spaces based on Fr\'echet means and propose a network regression method using the Wasserstein metric. We show that when representing graphs as multivariate Gaussian distributions, the network regression problem requires the computation of a Riemannian center of mass (i.e., Fr\'echet means). Fr\'echet means with non-negative weights translates into a barycenter problem and can be efficiently computed using fixed point iterations. Although the convergence guarantees of fixed-point iterations for the computation of Wasserstein affine averages remain an open problem, we provide evidence of convergence in a large number of synthetic and real-data scenarios. Extensive numerical results show that the proposed approach improves existing procedures by accurately accounting for graph size, topology, and sparsity in synthetic experiments. Additionally, real-world experiments using the proposed approach result in higher Coefficient of Determination ($R^{2}$) values and lower mean squared prediction error (MSPE), cementing improved prediction capabilities in practice.
Authors: Yi Liu, Cong Wang, Xingliang Yuan
Abstract: Federated Learning (FL) is susceptible to poisoning attacks, wherein compromised clients manipulate the global model by modifying local datasets or sending manipulated model updates. Experienced defenders can readily detect and mitigate the poisoning effects of malicious behaviors using Byzantine-robust aggregation rules. However, the exploration of poisoning attacks in scenarios where such behaviors are absent remains largely unexplored for Byzantine-robust FL. This paper addresses the challenging problem of poisoning Byzantine-robust FL by introducing catastrophic forgetting. To fill this gap, we first formally define generalization error and establish its connection to catastrophic forgetting, paving the way for the development of a clean-label data poisoning attack named BadSampler. This attack leverages only clean-label data (i.e., without poisoned data) to poison Byzantine-robust FL and requires the adversary to selectively sample training data with high loss to feed model training and maximize the model's generalization error. We formulate the attack as an optimization problem and present two elegant adversarial sampling strategies, Top-$\kappa$ sampling, and meta-sampling, to approximately solve it. Additionally, our formal error upper bound and time complexity analysis demonstrate that our design can preserve attack utility with high efficiency. Extensive evaluations on two real-world datasets illustrate the effectiveness and performance of our proposed attacks.
Authors: Changxi Chi, Hang Shi, Qi Zhu, Daoqiang Zhang, Wei Shao
Abstract: The rapid development of spatial transcriptomics(ST) enables the measurement of gene expression at spatial resolution, making it possible to simultaneously profile the gene expression, spatial locations of spots, and the matched histopathological images. However, the cost for collecting ST data is much higher than acquiring histopathological images, and thus several studies attempt to predict the gene expression on ST by leveraging their corresponding histopathological images. Most of the existing image-based gene prediction models treat the prediction task on each spot of ST data independently, which ignores the spatial dependency among spots. In addition, while the histology images share phenotypic characteristics with the ST data, it is still challenge to extract such common information to help align paired image and expression representations. To address the above issues, we propose a Multi-view Graph Contrastive Learning framework with HSIC-bottleneck Regularization(ST-GCHB) aiming at learning shared representation to help impute the gene expression of the queried imagingspots by considering their spatial dependency.
Authors: Chenyuan Wu, Gangwei Jiang, Defu Lian
Abstract: Lifelong prompt tuning has significantly advanced parameter-efficient lifelong learning with its efficiency and minimal storage demands on various tasks. Our empirical studies, however, highlights certain transferability constraints in the current methodologies: a universal algorithm that guarantees consistent positive transfer across all tasks is currently unattainable, especially when dealing dissimilar tasks that may engender negative transfer. Identifying the misalignment between algorithm selection and task specificity as the primary cause of negative transfer, we present the Similarity Heuristic Lifelong Prompt Tuning (SHLPT) framework. This innovative strategy partitions tasks into two distinct subsets by harnessing a learnable similarity metric, thereby facilitating fruitful transfer from tasks regardless of their similarity or dissimilarity. Additionally, SHLPT incorporates a parameter pool to combat catastrophic forgetting effectively. Our experiments shows that SHLPT outperforms state-of-the-art techniques in lifelong learning benchmarks and demonstrates robustness against negative transfer in diverse task sequences.
Authors: Lijie Hu, Liang Liu, Shu Yang, Xin Chen, Hongru Xiao, Mengdi Li, Pan Zhou, Muhammad Asif Ali, Di Wang
Abstract: Chain-of-Thought (CoT) holds a significant place in augmenting the reasoning performance for large language models (LLMs). While some studies focus on improving CoT accuracy through methods like retrieval enhancement, yet a rigorous explanation for why CoT achieves such success remains unclear. In this paper, we analyze CoT methods under two different settings by asking the following questions: (1) For zero-shot CoT, why does prompting the model with "let's think step by step" significantly impact its outputs? (2) For few-shot CoT, why does providing examples before questioning the model could substantially improve its reasoning ability? To answer these questions, we conduct a top-down explainable analysis from the Hopfieldian view and propose a Read-and-Control approach for controlling the accuracy of CoT. Through extensive experiments on seven datasets for three different tasks, we demonstrate that our framework can decipher the inner workings of CoT, provide reasoning error localization, and control to come up with the correct reasoning path.
Authors: Emanuele Zappala
Abstract: We obtain a new universal approximation theorem for continuous operators on arbitrary Banach spaces using the Leray-Schauder mapping. Moreover, we introduce and study a method for operator learning in Banach spaces $L^p$ of functions with multiple variables, based on orthogonal projections on polynomial bases. We derive a universal approximation result for operators where we learn a linear projection and a finite dimensional mapping under some additional assumptions. For the case of $p=2$, we give some sufficient conditions for the approximation results to hold. This article serves as the theoretical framework for a deep learning methodology whose implementation will be provided in subsequent work.
Authors: Sebastian Neumayer, Fabian Altekr\"uger
Abstract: Theoretical guarantees for the robust solution of inverse problems have important implications for applications. To achieve both guarantees and high reconstruction quality, we propose to learn a pixel-based ridge regularizer with a data-dependent and spatially-varying regularization strength. For this architecture, we establish the existence of solutions to the associated variational problem and the stability of its solution operator. Further, we prove that the reconstruction forms a maximum-a-posteriori approach. Simulations for biomedical imaging and material sciences demonstrate that the approach yields high-quality reconstructions even if only a small instance-specific training set is available.
Authors: Viet Vo, Thusitha Dayaratne, Blake Haydon, Xingliang Yuan, Shangqi Lai, Sharif Abuadbba, Hajime Suzuki, Carsten Rudolph
Abstract: Spectrum sharing is increasingly vital in 6G wireless communication, facilitating dynamic access to unused spectrum holes. Recently, there has been a significant shift towards employing machine learning (ML) techniques for sensing spectrum holes. In this context, federated learning (FL)-enabled spectrum sensing technology has garnered wide attention, allowing for the construction of an aggregated ML model without disclosing the private spectrum sensing information of wireless user devices. However, the integrity of collaborative training and the privacy of spectrum information from local users have remained largely unexplored. This article first examines the latest developments in FL-enabled spectrum sharing for prospective 6G scenarios. It then identifies practical attack vectors in 6G to illustrate potential AI-powered security and privacy threats in these contexts. Finally, the study outlines future directions, including practical defense challenges and guidelines.
Authors: Zhiyu Guo, Hidetaka Kamigaito, Taro Watanabe
Abstract: Scaling the context size of large language models (LLMs) enables them to perform various new tasks, e.g., book summarization. However, the memory cost of the Key and Value (KV) cache in attention significantly limits the practical applications of LLMs. Recent works have explored token pruning for KV cache reduction in LLMs, relying solely on attention scores as a token importance indicator. However, our investigation into value vector norms revealed a notably non-uniform pattern questioning their reliance only on attention scores. Inspired by this, we propose a new method: Value-Aware Token Pruning (VATP) which uses both attention scores and the $ \ell_{1} $ norm of value vectors to evaluate token importance. Extensive experiments on LLaMA2-7B-chat and Vicuna-v1.5-7B across 16 LongBench tasks demonstrate VATP's superior performance.
Authors: Sujoy Roychowdhury, Sumit Soman, H. G. Ranjani, Vansh Chhabra, Neeraj Gunda, Subhadip Bandyopadhyay, Sai Krishna Bala
Abstract: A plethora of sentence embedding models makes it challenging to choose one, especially for domains such as telecom, rich with specialized vocabulary. We evaluate multiple embeddings obtained from publicly available models and their domain-adapted variants, on both point retrieval accuracies as well as their (95\%) confidence intervals. We establish a systematic method to obtain thresholds for similarity scores for different embeddings. We observe that fine-tuning improves mean bootstrapped accuracies as well as tightens confidence intervals. The pre-training combined with fine-tuning makes confidence intervals even tighter. To understand these variations, we analyse and report significant correlations between the distributional overlap between top-$K$, correct and random sentence similarities with retrieval accuracies and similarity thresholds. Following current literature, we analyze if retrieval accuracy variations can be attributed to isotropy of embeddings. Our conclusions are that isotropy of embeddings (as measured by two independent state-of-the-art isotropy metric definitions) cannot be attributed to better retrieval performance. However, domain adaptation which improves retrieval accuracies also improves isotropy. We establish that domain adaptation moves domain specific embeddings further away from general domain embeddings.
Authors: Hao Zeng, Jiaqi Wang, Avirup Das, Junying He, Kunpeng Han, Haoyuan Hu, Mingfei Sun
Abstract: Feasible solutions are crucial for Integer Programming (IP) since they can substantially speed up the solving process. In many applications, similar IP instances often exhibit similar structures and shared solution distributions, which can be potentially modeled by deep learning methods. Unfortunately, existing deep-learning-based algorithms, such as Neural Diving and Predict-and-search framework, are limited to generating only partial feasible solutions, and they must rely on solvers like SCIP and Gurobi to complete the solutions for a given IP problem. In this paper, we propose a novel framework that generates complete feasible solutions end-to-end. Our framework leverages contrastive learning to characterize the relationship between IP instances and solutions, and learns latent embeddings for both IP instances and their solutions. Further, the framework employs diffusion models to learn the distribution of solution embeddings conditioned on IP representations, with a dedicated guided sampling strategy that accounts for both constraints and objectives. We empirically evaluate our framework on four typical datasets of IP problems, and show that it effectively generates complete feasible solutions with a high probability (> 89.7 \%) without the reliance of Solvers and the quality of solutions is comparable to the best heuristic solutions from Gurobi. Furthermore, by integrating our method's sampled partial solutions with the CompleteSol heuristic from SCIP, the resulting feasible solutions outperform those from state-of-the-art methods across all datasets, exhibiting a 3.7 to 33.7\% improvement in the gap to optimal values, and maintaining a feasible ratio of over 99.7\% for all datasets.
Authors: Soma Yokoi, Issei Sato
Abstract: Sum-product networks (SPNs) are probabilistic models characterized by exact and fast evaluation of fundamental probabilistic operations. Its superior computational tractability has led to applications in many fields, such as machine learning with time constraints or accuracy requirements and real-time systems. The structural constraints of SPNs supporting fast inference, however, lead to increased learning-time complexity and can be an obstacle to building highly expressive SPNs. This study aimed to develop a Bayesian learning approach that can be efficiently implemented on large-scale SPNs. We derived a new full conditional probability of Gibbs sampling by marginalizing multiple random variables to expeditiously obtain the posterior distribution. The complexity analysis revealed that our sampling algorithm works efficiently even for the largest possible SPN. Furthermore, we proposed a hyperparameter tuning method that balances the diversity of the prior distribution and optimization efficiency in large-scale SPNs. Our method has improved learning-time complexity and demonstrated computational speed tens to more than one hundred times faster and superior predictive performance in numerical experiments on more than 20 datasets.
Authors: Honglei Zhang, Jukka I. Ahonen, Nam Le, Ruiying Yang, Francesco Cricri
Abstract: This paper investigates the efficacy of jointly optimizing content-specific post-processing filters to adapt a human oriented video/image codec into a codec suitable for machine vision tasks. By observing that artifacts produced by video/image codecs are content-dependent, we propose a novel training strategy based on competitive learning principles. This strategy assigns training samples to filters dynamically, in a fuzzy manner, which further optimizes the winning filter on the given sample. Inspired by simulated annealing optimization techniques, we employ a softmax function with a temperature variable as the weight allocation function to mitigate the effects of random initialization. Our evaluation, conducted on a system utilizing multiple post-processing filters within a Versatile Video Coding (VVC) codec framework, demonstrates the superiority of content-specific filters trained with our proposed strategies, specifically, when images are processed in blocks. Using VVC reference software VTM 12.0 as the anchor, experiments on the OpenImages dataset show an improvement in the BD-rate reduction from -41.3% and -44.6% to -42.3% and -44.7% for object detection and instance segmentation tasks, respectively, compared to independently trained filters. The statistics of the filter usage align with our hypothesis and underscore the importance of jointly optimizing filters for both content and reconstruction quality. Our findings pave the way for further improving the performance of video/image codecs.
Authors: Yichen Pan, Dehan Kong, Sida Zhou, Cheng Cui, Yifei Leng, Bing Jiang, Hangyu Liu, Yanyi Shang, Shuyan Zhou, Tongshuang Wu, Zhengyang Wu
Abstract: For web agents to be practically useful, they must adapt to the continuously evolving web environment characterized by frequent updates to user interfaces and content. However, most existing benchmarks only capture the static aspects of the web. To bridge this gap, we introduce WebCanvas, an innovative online evaluation framework for web agents that effectively addresses the dynamic nature of web interactions. WebCanvas contains three main components to facilitate realistic assessments: (1) A novel evaluation metric which reliably capture critical intermediate actions or states necessary for task completions while disregarding noise caused by insignificant events or changed web-elements. (2) A benchmark dataset called Mind2Web-Live, a refined version of original Mind2Web static dataset containing 542 tasks with 2439 intermediate evaluation states; (3) Lightweight and generalizable annotation tools and testing pipelines that enables the community to collect and maintain the high-quality, up-to-date dataset. Building on WebCanvas, we open-source an agent framework with extensible modules for reasoning, providing a foundation for the community to conduct online inference and evaluations. Our best-performing agent achieves a task success rate of 23.1% and a task completion rate of 48.8% on the Mind2Web-Live test set. Additionally, we analyze the performance discrepancies across various websites, domains, and experimental environments. We encourage the community to contribute further insights on online agent evaluation, thereby advancing this field of research.
Authors: Maximilian Schebek, Michele Invernizzi, Frank No\'e, Jutta Rogal
Abstract: The accurate prediction of phase diagrams is of central importance for both the fundamental understanding of materials as well as for technological applications in material sciences. However, the computational prediction of the relative stability between phases based on their free energy is a daunting task, as traditional free energy estimators require a large amount of simulation data to obtain uncorrelated equilibrium samples over a grid of thermodynamic states. In this work, we develop deep generative machine learning models for entire phase diagrams, employing normalizing flows conditioned on the thermodynamic states, e.g., temperature and pressure, that they map to. By training a single normalizing flow to transform the equilibrium distribution sampled at only one reference thermodynamic state to a wide range of target temperatures and pressures, we can efficiently generate equilibrium samples across the entire phase diagram. Using a permutation-equivariant architecture allows us, thereby, to treat solid and liquid phases on the same footing. We demonstrate our approach by predicting the solid-liquid coexistence line for a Lennard-Jones system in excellent agreement with state-of-the-art free energy methods while significantly reducing the number of energy evaluations needed.
Authors: Jincheng Zhou
Abstract: Question-Options Generation (QOG) is a task that involves generating a set of question-options pairs given context. This task has various applications, including fine-tuning large models, information retrieval, and automated multiple-choice question generation for education. In this paper, we develop QOG models using three different methods based on fine-tuning sequence-to-sequence language models (LMs). Experiments demonstrate that the end-to-end QOG model is computationally efficient and stable during both training and inference, outperforming other methods. Furthermore, our analysis indicates that our QOG models are competitive on the QOG task compared to the large language model Llama 3-8B.
Authors: Matthew Ashman, Cristiana Diaconu, Junhyuck Kim, Lakee Sivaraya, Stratis Markou, James Requeima, Wessel P. Bruinsma, Richard E. Turner
Abstract: The effectiveness of neural processes (NPs) in modelling posterior prediction maps -- the mapping from data to posterior predictive distributions -- has significantly improved since their inception. This improvement can be attributed to two principal factors: (1) advancements in the architecture of permutation invariant set functions, which are intrinsic to all NPs; and (2) leveraging symmetries present in the true posterior predictive map, which are problem dependent. Transformers are a notable development in permutation invariant set functions, and their utility within NPs has been demonstrated through the family of models we refer to as TNPs. Despite significant interest in TNPs, little attention has been given to incorporating symmetries. Notably, the posterior prediction maps for data that are stationary -- a common assumption in spatio-temporal modelling -- exhibit translation equivariance. In this paper, we introduce of a new family of translation equivariant TNPs that incorporate translation equivariance. Through an extensive range of experiments on synthetic and real-world spatio-temporal data, we demonstrate the effectiveness of TE-TNPs relative to their non-translation-equivariant counterparts and other NP baselines.
Authors: Mattia Litrico, Francesco Guarnera, Valerio Giuffirda, Daniele Rav\`i, Sebastiano Battiato
Abstract: Generating realistic images to accurately predict changes in the structure of brain MRI is a crucial tool for clinicians. Such applications help assess patients' outcomes and analyze how diseases progress at the individual level. However, existing methods for this task present some limitations. Some approaches attempt to model the distribution of MRI scans directly by conditioning the model on patients' ages, but they fail to explicitly capture the relationship between structural changes in the brain and time intervals, especially on age-unbalanced datasets. Other approaches simply rely on interpolation between scans, which limits their clinical application as they do not predict future MRIs. To address these challenges, we propose a Temporally-Aware Diffusion Model (TADM), which introduces a novel approach to accurately infer progression in brain MRIs. TADM learns the distribution of structural changes in terms of intensity differences between scans and combines the prediction of these changes with the initial baseline scans to generate future MRIs. Furthermore, during training, we propose to leverage a pre-trained Brain-Age Estimator (BAE) to refine the model's training process, enhancing its ability to produce accurate MRIs that match the expected age gap between baseline and generated scans. Our assessment, conducted on the OASIS-3 dataset, uses similarity metrics and region sizes computed by comparing predicted and real follow-up scans on 3 relevant brain regions. TADM achieves large improvements over existing approaches, with an average decrease of 24% in region size error and an improvement of 4% in similarity metrics. These evaluations demonstrate the improvement of our model in mimicking temporal brain neurodegenerative progression compared to existing methods. Our approach will benefit applications, such as predicting patient outcomes or improving treatments for patients.
Authors: Philipp Seeberger, Dominik Wagner, Korbinian Riedhammer
Abstract: With the advancement of multimedia technologies, news documents and user-generated content are often represented as multiple modalities, making Multimedia Event Extraction (MEE) an increasingly important challenge. However, recent MEE methods employ weak alignment strategies and data augmentation with simple classification models, which ignore the capabilities of natural language-formulated event templates for the challenging Event Argument Extraction (EAE) task. In this work, we focus on EAE and address this issue by introducing a unified template filling model that connects the textual and visual modalities via textual prompts. This approach enables the exploitation of cross-ontology transfer and the incorporation of event-specific semantics. Experiments on the M2E2 benchmark demonstrate the effectiveness of our approach. Our system surpasses the current SOTA on textual EAE by +7% F1, and performs generally better than the second-best systems for multimedia EAE.
Authors: Kentaro Mitsui, Koh Mitsuda, Toshiaki Wakatsuki, Yukiya Hono, Kei Sawada
Abstract: Multimodal language models that process both text and speech have a potential for applications in spoken dialogue systems. However, current models face two major challenges in response generation latency: (1) generating a spoken response requires the prior generation of a written response, and (2) speech sequences are significantly longer than text sequences. This study addresses these issues by extending the input and output sequences of the language model to support the parallel generation of text and speech. Our experiments on spoken question answering tasks demonstrate that our approach improves latency while maintaining the quality of response content. Additionally, we show that latency can be further reduced by generating speech in multiple sequences. Demo samples are available at https://rinnakk.github.io/research/publications/PSLM.
Authors: Myeonghwa Lee, Seonho An, Min-Soo Kim
Abstract: In this paper, we conduct a study to utilize LLMs as a solution for decision making that requires complex data analysis. We define Decision QA as the task of answering the best decision, $d_{best}$, for a decision-making question $Q$, business rules $R$ and a database $D$. Since there is no benchmark that can examine Decision QA, we propose Decision QA benchmark, DQA. It has two scenarios, Locating and Building, constructed from two video games (Europa Universalis IV and Victoria 3) that have almost the same goal as Decision QA. To address Decision QA effectively, we also propose a new RAG technique called the iterative plan-then-retrieval augmented generation (PlanRAG). Our PlanRAG-based LM generates the plan for decision making as the first step, and the retriever generates the queries for data analysis as the second step. The proposed method outperforms the state-of-the-art iterative RAG method by 15.8% in the Locating scenario and by 7.4% in the Building scenario, respectively. We release our code and benchmark at https://github.com/myeon9h/PlanRAG.
Authors: Yasaman Torabi, Shahram Shirani, James P. Reilly
Abstract: This paper presents a comprehensive review of cardiorespiratory auscultation sensing devices which is useful for understanding the theoretical aspects of sensing devices, as well as practical notes to design novel sensing devices. One of the methods to design a stethoscope is using electret condenser microphones (ECM). In this paper, we first introduce the acoustic properties of the heart and lungs, as well as a brief history of stethoscope evolution. Then, we discuss the basic concept of ECM sensors and a recent stethoscope based on this technology. In response to the limitations of ECM-based systems, we explore the potential of microelectromechanical systems (MEMS), particularly focusing on piezoelectric transducer (PZT) sensors. This paper comprehensively reviews sensing technologies, emphasizing innovative MEMS-based designs for wearable cardiopulmonary auscultation in the past decade. To our knowledge, this is the first paper to summarize ECM and MEMS applications for heart and lung sound analysis. Keywords: Micro-electro-mechanical Systems (MEMS); Electret Condenser Microphone (ECM); Wearable Sensing Devices; Cardiorespiratory Auscultation; Phonocardiography (PCG); Heart Sound; Lung Sound
Authors: Jia Qi Yip, Shengkui Zhao, Dianwen Ng, Eng Siong Chng, Bin Ma
Abstract: Recent improvements in neural audio codec (NAC) models have generated interest in adopting pre-trained codecs for a variety of speech processing applications to take advantage of the efficiencies gained from high compression, but these have yet been applied to the speech separation (SS) task. SS can benefit from high compression because the compute required for traditional SS models makes them impractical for many edge computing use cases. However, SS is a waveform-masking task where compression tends to introduce distortions that severely impact performance. Here we propose a novel task of Audio Codec-based SS, where SS is performed within the embedding space of a NAC, and propose a new model, Codecformer, to address this task. At inference, Codecformer achieves a 52x reduction in MAC while producing separation performance comparable to a cloud deployment of Sepformer. This method charts a new direction for performing efficient SS in practical scenarios.
Authors: Aditya Jain, Fagner Cunha, Michael James Bunsen, Juan Sebasti\'an Ca\~nas, L\'eonard Pasi, Nathan Pinoy, Flemming Helsing, JoAnne Russo, Marc Botham, Michael Sabourin, Jonathan Fr\'echette, Alexandre Anctil, Yacksecari Lopez, Eduardo Navarro, Filonila Perez Pimentel, Ana Cecilia Zamora, Jos\'e Alejandro Ramirez Silva, Jonathan Gagnon, Tom August, Kim Bjerge, Alba Gomez Segura, Marc B\'elisle, Yves Basset, Kent P. McFarland, David Roy, Toke Thomas H{\o}ye, Maxim Larriv\'ee, David Rolnick
Abstract: Insects represent half of all global biodiversity, yet many of the world's insects are disappearing, with severe implications for ecosystems and agriculture. Despite this crisis, data on insect diversity and abundance remain woefully inadequate, due to the scarcity of human experts and the lack of scalable tools for monitoring. Ecologists have started to adopt camera traps to record and study insects, and have proposed computer vision algorithms as an answer for scalable data processing. However, insect monitoring in the wild poses unique challenges that have not yet been addressed within computer vision, including the combination of long-tailed data, extremely similar classes, and significant distribution shifts. We provide the first large-scale machine learning benchmarks for fine-grained insect recognition, designed to match real-world tasks faced by ecologists. Our contributions include a curated dataset of images from citizen science platforms and museums, and an expert-annotated dataset drawn from automated camera traps across multiple continents, designed to test out-of-distribution generalization under field conditions. We train and evaluate a variety of baseline algorithms and introduce a combination of data augmentation techniques that enhance generalization across geographies and hardware setups. Code and datasets are made publicly available.
Authors: Stefan Sylvius Wagner, Maike Behrendt, Marc Ziegele, Stefan Harmeling
Abstract: Stance detection holds great potential for enhancing the quality of online political discussions, as it has shown to be useful for summarizing discussions, detecting misinformation, and evaluating opinion distributions. Usually, transformer-based models are used directly for stance detection, which require large amounts of data. However, the broad range of debate questions in online political discussion creates a variety of possible scenarios that the model is faced with and thus makes data acquisition for model training difficult. In this work, we show how to leverage LLM-generated synthetic data to train and improve stance detection agents for online political discussions:(i) We generate synthetic data for specific debate questions by prompting a Mistral-7B model and show that fine-tuning with the generated synthetic data can substantially improve the performance of stance detection. (ii) We examine the impact of combining synthetic data with the most informative samples from an unlabelled dataset. First, we use the synthetic data to select the most informative samples, second, we combine both these samples and the synthetic data for fine-tuning. This approach reduces labelling effort and consistently surpasses the performance of the baseline model that is trained with fully labeled data. Overall, we show in comprehensive experiments that LLM-generated data greatly improves stance detection performance for online political discussions.
Authors: Heng Xu, Tianqing Zhu, Lefeng Zhang, Wanlei Zhou, Philip S. Yu
Abstract: Federated learning is a promising privacy-preserving paradigm for distributed machine learning. In this context, there is sometimes a need for a specialized process called machine unlearning, which is required when the effect of some specific training samples needs to be removed from a learning model due to privacy, security, usability, and/or legislative factors. However, problems arise when current centralized unlearning methods are applied to existing federated learning, in which the server aims to remove all information about a class from the global model. Centralized unlearning usually focuses on simple models or is premised on the ability to access all training data at a central node. However, training data cannot be accessed on the server under the federated learning paradigm, conflicting with the requirements of the centralized unlearning process. Additionally, there are high computation and communication costs associated with accessing clients' data, especially in scenarios involving numerous clients or complex global models. To address these concerns, we propose a more effective and efficient federated unlearning scheme based on the concept of model explanation. Model explanation involves understanding deep networks and individual channel importance, so that this understanding can be used to determine which model channels are critical for classes that need to be unlearned. We select the most influential channels within an already-trained model for the data that need to be unlearned and fine-tune only influential channels to remove the contribution made by those data. In this way, we can simultaneously avoid huge consumption costs and ensure that the unlearned model maintains good performance. Experiments with different training models on various datasets demonstrate the effectiveness of the proposed approach.
Authors: Julian Rodemann
Abstract: A wide range of machine learning algorithms iteratively add data to the training sample. Examples include semi-supervised learning, active learning, multi-armed bandits, and Bayesian optimization. We embed this kind of data addition into decision theory by framing data selection as a decision problem. This paves the way for finding Bayes-optimal selections of data. For the illustrative case of self-training in semi-supervised learning, we derive the respective Bayes criterion. We further show that deploying this criterion mitigates the issue of confirmation bias by empirically assessing our method for generalized linear models, semi-parametric generalized additive models, and Bayesian neural networks on simulated and real-world data.
Authors: Viktor Moskvoretskii, Nazarii Tupitsa, Chris Biemann, Samuel Horv\'ath, Eduard Gorbunov, Irina Nikishina
Abstract: We present a new approach based on the Personalized Federated Learning algorithm MeritFed that can be applied to Natural Language Tasks with heterogeneous data. We evaluate it on the Low-Resource Machine Translation task, using the dataset from the Large-Scale Multilingual Machine Translation Shared Task (Small Track #2) and the subset of Sami languages from the multilingual benchmark for Finno-Ugric languages. In addition to its effectiveness, MeritFed is also highly interpretable, as it can be applied to track the impact of each language used for training. Our analysis reveals that target dataset size affects weight distribution across auxiliary languages, that unrelated languages do not interfere with the training, and auxiliary optimizer parameters have minimal impact. Our approach is easy to apply with a few lines of code, and we provide scripts for reproducing the experiments at https://github.com/VityaVitalich/MeritFed
Authors: Eldar Kurtic, Amir Moeini, Dan Alistarh
Abstract: We introduce Mathador-LM, a new benchmark for evaluating the mathematical reasoning on large language models (LLMs), combining ruleset interpretation, planning, and problem-solving. This benchmark is inspired by the Mathador game, where the objective is to reach a target number using basic arithmetic operations on a given set of base numbers, following a simple set of rules. We show that, across leading LLMs, we obtain stable average performance while generating benchmark instances dynamically, following a target difficulty level. Thus, our benchmark alleviates concerns about test-set leakage into training data, an issue that often undermines popular benchmarks. Additionally, we conduct a comprehensive evaluation of both open and closed-source state-of-the-art LLMs on Mathador-LM. Our findings reveal that contemporary models struggle with Mathador-LM, scoring significantly lower than average 5th graders. This stands in stark contrast to their strong performance on popular mathematical reasoning benchmarks.
Authors: Tuan-Luc Huynh, Thuy-Trang Vu, Weiqing Wang, Yinwei Wei, Trung Le, Dragan Gasevic, Yuan-Fang Li, Thanh-Toan Do
Abstract: Differentiable Search Index (DSI) utilizes Pre-trained Language Models (PLMs) for efficient document retrieval without relying on external indexes. However, DSIs need full re-training to handle updates in dynamic corpora, causing significant computational inefficiencies. We introduce PromptDSI, a rehearsal-free, prompt-based approach for instance-wise incremental learning in document retrieval. PromptDSI attaches prompts to the frozen PLM's encoder of DSI, leveraging its powerful representation to efficiently index new corpora while maintaining a balance between stability and plasticity. We eliminate the initial forward pass of prompt-based continual learning methods that doubles training and inference time. Moreover, we propose a topic-aware prompt pool that employs neural topic embeddings as fixed keys. This strategy ensures diverse and effective prompt usage, addressing the challenge of parameter underutilization caused by the collapse of the query-key matching mechanism. Our empirical evaluations demonstrate that PromptDSI matches IncDSI in managing forgetting while significantly enhancing recall by over 4% on new corpora.
Authors: A. L. Garc\'ia Navarro, Nataliia Koneva, Alfonso S\'anchez-Maci\'an, Jos\'e Alberto Hern\'andez, \'Oscar Gonz\'alez de Dios, J. M. Rivas-Moscoso
Abstract: This article provides a methodology and open-source implementation of Reinforcement Learning algorithms for finding optimal routes in a packet-optical network scenario. The algorithm uses measurements provided by the physical layer (pre-FEC bit error rate and propagation delay) and the link layer (link load) to configure a set of latency-based rewards and penalties based on such measurements. Then, the algorithm executes Q-learning based on this set of rewards for finding the optimal routing strategies. It is further shown that the algorithm dynamically adapts to changing network conditions by re-calculating optimal policies upon either link load changes or link degradation as measured by pre-FEC BER.
Authors: Jo\~ao A. Leite, Olesya Razuvayevskaya, Kalina Bontcheva, Carolina Scarton
Abstract: This work introduces EUvsDisinfo, a multilingual dataset of trustworthy and disinformation articles related to pro-Kremlin themes. It is sourced directly from the debunk articles written by experts leading the EUvsDisinfo project. Our dataset is the largest to-date resource in terms of the overall number of articles and distinct languages. It also provides the largest topical and temporal coverage. Using this dataset, we investigate the dissemination of pro-Kremlin disinformation across different languages, uncovering language-specific patterns targeting specific disinformation topics. We further analyse the evolution of topic distribution over an eight-year period, noting a significant surge in disinformation content before the full-scale invasion of Ukraine in 2022. Lastly, we demonstrate the dataset's applicability in training models to effectively distinguish between disinformation and trustworthy content in multilingual settings.
Authors: Jiang-Xin Shi, Chi Zhang, Tong Wei, Yu-Feng Li
Abstract: Pre-trained vision-language models like CLIP have shown powerful zero-shot inference ability via image-text matching and prove to be strong few-shot learners in various downstream tasks. However, in real-world scenarios, adapting CLIP to downstream tasks may encounter the following challenges: 1) data may exhibit long-tailed data distributions and might not have abundant samples for all the classes; 2) There might be emerging tasks with new classes that contain no samples at all. To overcome them, we propose a novel framework to achieve efficient and long-tailed generalization, which can be termed as Candle. During the training process, we propose compensating logit-adjusted loss to encourage large margins of prototypes and alleviate imbalance both within the base classes and between the base and new classes. For efficient adaptation, we treat the CLIP model as a black box and leverage the extracted features to obtain visual and textual prototypes for prediction. To make full use of multi-modal information, we also propose cross-modal attention to enrich the features from both modalities. For effective generalization, we introduce virtual prototypes for new classes to make up for their lack of training images. Candle achieves state-of-the-art performance over extensive experiments on 11 diverse datasets while substantially reducing the training time, demonstrating the superiority of our approach. The source code is available at https://github.com/shijxcs/Candle.
Authors: Sunny Soni, Aaqib Saeed, Yuki M. Asano
Abstract: Federated Learning (FL) enables multiple machines to collaboratively train a machine learning model without sharing of private training data. Yet, especially for heterogeneous models, a key bottleneck remains the transfer of knowledge gained from each client model with the server. One popular method, FedDF, uses distillation to tackle this task with the use of a common, shared dataset on which predictions are exchanged. However, in many contexts such a dataset might be difficult to acquire due to privacy and the clients might not allow for storage of a large shared dataset. To this end, in this paper, we introduce a new method that improves this knowledge distillation method to only rely on a single shared image between clients and server. In particular, we propose a novel adaptive dataset pruning algorithm that selects the most informative crops generated from only a single image. With this, we show that federated learning with distillation under a limited shared dataset budget works better by using a single image compared to multiple individual ones. Finally, we extend our approach to allow for training heterogeneous client architectures by incorporating a non-uniform distillation schedule and client-model mirroring on the server side.
Authors: Isma\"el Castillo, Alice L'Huillier, Kolyan Ray, Luke Travis
Abstract: We propose a scalable variational Bayes method for statistical inference for a single or low-dimensional subset of the coordinates of a high-dimensional parameter in sparse linear regression. Our approach relies on assigning a mean-field approximation to the nuisance coordinates and carefully modelling the conditional distribution of the target given the nuisance. This requires only a preprocessing step and preserves the computational advantages of mean-field variational Bayes, while ensuring accurate and reliable inference for the target parameter, including for uncertainty quantification. We investigate the numerical performance of our algorithm, showing that it performs competitively with existing methods. We further establish accompanying theoretical guarantees for estimation and uncertainty quantification in the form of a Bernstein--von Mises theorem.
Authors: Oliver J. Sutton, Qinghua Zhou, Wei Wang, Desmond J. Higham, Alexander N. Gorban, Alexander Bastounis, Ivan Y. Tyukin
Abstract: We reveal new methods and the theoretical foundations of techniques for editing large language models. We also show how the new theory can be used to assess the editability of models and to expose their susceptibility to previously unknown malicious attacks. Our theoretical approach shows that a single metric (a specific measure of the intrinsic dimensionality of the model's features) is fundamental to predicting the success of popular editing approaches, and reveals new bridges between disparate families of editing methods. We collectively refer to these approaches as stealth editing methods, because they aim to directly and inexpensively update a model's weights to correct the model's responses to known hallucinating prompts without otherwise affecting the model's behaviour, without requiring retraining. By carefully applying the insight gleaned from our theoretical investigation, we are able to introduce a new network block -- named a jet-pack block -- which is optimised for highly selective model editing, uses only standard network operations, and can be inserted into existing networks. The intrinsic dimensionality metric also determines the vulnerability of a language model to a stealth attack: a small change to a model's weights which changes its response to a single attacker-chosen prompt. Stealth attacks do not require access to or knowledge of the model's training data, therefore representing a potent yet previously unrecognised threat to redistributed foundation models. They are computationally simple enough to be implemented in malware in many cases. Extensive experimental results illustrate and support the method and its theoretical underpinnings. Demos and source code for editing language models are available at https://github.com/qinghua-zhou/stealth-edits.
Authors: Bernhard Stankewitz, Botond Szabo
Abstract: Due to their flexibility and theoretical tractability Gaussian process (GP) regression models have become a central topic in modern statistics and machine learning. While the true posterior in these models is given explicitly, numerical evaluations depend on the inversion of the augmented kernel matrix $ K + \sigma^2 I $, which requires up to $ O(n^3) $ operations. For large sample sizes n, which are typically given in modern applications, this is computationally infeasible and necessitates the use of an approximate version of the posterior. Although such methods are widely used in practice, they typically have very limtied theoretical underpinning. In this context, we analyze a class of recently proposed approximation algorithms from the field of Probabilistic numerics. They can be interpreted in terms of Lanczos approximate eigenvectors of the kernel matrix or a conjugate gradient approximation of the posterior mean, which are particularly advantageous in truly large scale applications, as they are fundamentally only based on matrix vector multiplications amenable to the GPU acceleration of modern software frameworks. We combine result from the numerical analysis literature with state of the art concentration results for spectra of kernel matrices to obtain minimax contraction rates. Our theoretical findings are illustrated by numerical experiments.
Authors: Nagur Shareef Shaik, Teja Krishna Cherukuri, Vince Calhoun, Dong Hye Ye
Abstract: Schizophrenia is a debilitating, chronic mental disorder that significantly impacts an individual's cognitive abilities, behavior, and social interactions. It is characterized by subtle morphological changes in the brain, particularly in the gray matter. These changes are often imperceptible through manual observation, demanding an automated approach to diagnosis. This study introduces a deep learning methodology for the classification of individuals with Schizophrenia. We achieve this by implementing a diversified attention mechanism known as Spatial Sequence Attention (SSA) which is designed to extract and emphasize significant feature representations from structural MRI (sMRI). Initially, we employ the transfer learning paradigm by leveraging pre-trained DenseNet to extract initial feature maps from the final convolutional block which contains morphological alterations associated with Schizophrenia. These features are further processed by the proposed SSA to capture and emphasize intricate spatial interactions and relationships across volumes within the brain. Our experimental studies conducted on a clinical dataset have revealed that the proposed attention mechanism outperforms the existing Squeeze & Excitation Network for Schizophrenia classification.
Authors: Polina Karpikova, Andrei Spiridonov, Anna Vorontsova, Anastasia Yaschenko, Ekaterina Radionova, Igor Medvedev, Alexander Limonov
Abstract: Self-portraits captured from a short distance might look unnatural or even unattractive due to heavy distortions making facial features malformed, and ill-placed head poses. In this paper, we propose SUPER, a novel method of eliminating distortions and adjusting head pose in a close-up face crop. We perform 3D GAN inversion for a facial image by optimizing camera parameters and face latent code, which gives a generated image. Besides, we estimate depth from the obtained latent code, create a depth-induced 3D mesh, and render it with updated camera parameters to obtain a warped portrait. Finally, we apply the visibility-based blending so that visible regions are reprojected, and occluded parts are restored with a generative model. Experiments on face undistortion benchmarks and on our self-collected Head Rotation dataset (HeRo), show that SUPER outperforms previous approaches both qualitatively and quantitatively, opening new possibilities for photorealistic selfie editing.
Authors: Mirko Stumpo, Monica Laurenza, Simone Benella, Maria Federica Marcucci
Abstract: The need of real-time of monitoring and alerting systems for Space Weather hazards has grown significantly in the last two decades. One of the most important challenge for space mission operations and planning is the prediction of solar proton events (SPEs). In this context, artificial intelligence and machine learning techniques have opened a new frontier, providing a new paradigm for statistical forecasting algorithms. The great majority of these models aim to predict the occurrence of a SPE, i.e., they are based on the classification approach. In this work we present a simple and efficient machine learning regression algorithm which is able to forecast the energetic proton flux up to 1 hour ahead by exploiting features derived from the electron flux only. This approach could be helpful to improve monitoring systems of the radiation risk in both deep space and near-Earth environments. The model is very relevant for mission operations and planning, especially when flare characteristics and source location are not available in real time, as at Mars distance.
Authors: Francisco de Arriba-P\'erez, Silvia Garc\'ia-M\'endez, Javier Otero-Mosquera, Francisco J. Gonz\'alez-Casta\~no, Felipe Gil-Casti\~neira
Abstract: New technologies such as Machine Learning (ML) gave great potential for evaluating industry workflows and automatically generating key performance indicators (KPIs). However, despite established standards for measuring the efficiency of industrial machinery, there is no precise equivalent for workers' productivity, which would be highly desirable given the lack of a skilled workforce for the next generation of industry workflows. Therefore, an ML solution combining data from manufacturing processes and workers' performance for that goal is required. Additionally, in recent times intense effort has been devoted to explainable ML approaches that can automatically explain their decisions to a human operator, thus increasing their trustworthiness. We propose to apply explainable ML solutions to differentiate between expert and inexpert workers in industrial workflows, which we validate at a quality assessment industrial workstation. Regarding the methodology used, input data are captured by a manufacturing machine and stored in a NoSQL database. Data are processed to engineer features used in automatic classification and to compute workers' KPIs to predict their level of expertise (with all classification metrics exceeding 90 %). These KPIs, and the relevant features in the decisions are textually explained by natural language expansion on an explainability dashboard. These automatic explanations made it possible to infer knowledge from expert workers for inexpert workers. The latter illustrates the interest of research in self-explainable ML for automatically generating insights to improve productivity in industrial workflows.
Authors: Yunhao Chen, Xingjun Ma, Difan Zou, Yu-Gang Jiang
Abstract: As diffusion probabilistic models (DPMs) are being employed as mainstream models for generative artificial intelligence (AI), the study of their memorization of the raw training data has attracted growing attention. Existing works in this direction aim to establish an understanding of whether or to what extent DPMs learn by memorization. Such an understanding is crucial for identifying potential risks of data leakage and copyright infringement in diffusion models and, more importantly, for more controllable generation and trustworthy application of Artificial Intelligence Generated Content (AIGC). While previous works have made important observations of when DPMs are prone to memorization, these findings are mostly empirical, and the developed data extraction methods only work for conditional diffusion models. In this work, we aim to establish a theoretical understanding of memorization in DPMs with 1) a memorization metric for theoretical analysis, 2) an analysis of conditional memorization with informative and random labels, and 3) two better evaluation metrics for measuring memorization. Based on the theoretical analysis, we further propose a novel data extraction method called \textbf{Surrogate condItional Data Extraction (SIDE)} that leverages a classifier trained on generated data as a surrogate condition to extract training data directly from unconditional diffusion models. Our empirical results demonstrate that SIDE can extract training data from diffusion models where previous methods fail, and it is on average over 50\% more effective across different scales of the CelebA dataset.
Authors: Scott Pesme, Radu-Alexandru Dragomir, Nicolas Flammarion
Abstract: We examine the continuous-time counterpart of mirror descent, namely mirror flow, on classification problems which are linearly separable. Such problems are minimised `at infinity' and have many possible solutions; we study which solution is preferred by the algorithm depending on the mirror potential. For exponential tailed losses and under mild assumptions on the potential, we show that the iterates converge in direction towards a $\phi_\infty$-maximum margin classifier. The function $\phi_\infty$ is the $\textit{horizon function}$ of the mirror potential and characterises its shape `at infinity'. When the potential is separable, a simple formula allows to compute this function. We analyse several examples of potentials and provide numerical experiments highlighting our results.
Authors: David Huk, Yuanhe Zhang, Mark Steel, Ritabrata Dutta
Abstract: Recently proposed quasi-Bayesian (QB) methods initiated a new era in Bayesian computation by directly constructing the Bayesian predictive distribution through recursion, removing the need for expensive computations involved in sampling the Bayesian posterior distribution. This has proved to be data-efficient for univariate predictions, but extensions to multiple dimensions rely on a conditional decomposition resulting from predefined assumptions on the kernel of the Dirichlet Process Mixture Model, which is the implicit nonparametric model used. Here, we propose a different way to extend Quasi-Bayesian prediction to high dimensions through the use of Sklar's theorem by decomposing the predictive distribution into one-dimensional predictive marginals and a high-dimensional copula. Thus, we use the efficient recursive QB construction for the one-dimensional marginals and model the dependence using highly expressive vine copulas. Further, we tune hyperparameters using robust divergences (eg. energy score) and show that our proposed Quasi-Bayesian Vine (QB-Vine) is a fully non-parametric density estimator with \emph{an analytical form} and convergence rate independent of the dimension of data in some situations. Our experiments illustrate that the QB-Vine is appropriate for high dimensional distributions ($\sim$64), needs very few samples to train ($\sim$200) and outperforms state-of-the-art methods with analytical forms for density estimation and supervised tasks by a considerable margin.
Authors: Guy Kornowski, Swati Padmanabhan, Kai Wang, Zhe Zhang, Suvrit Sra
Abstract: Algorithms for bilevel optimization often encounter Hessian computations, which are prohibitive in high dimensions. While recent works offer first-order methods for unconstrained bilevel problems, the constrained setting remains relatively underexplored. We present first-order linearly constrained optimization methods with finite-time hypergradient stationarity guarantees. For linear equality constraints, we attain $\epsilon$-stationarity in $\widetilde{O}(\epsilon^{-2})$ gradient oracle calls, which is nearly-optimal. For linear inequality constraints, we attain $(\delta,\epsilon)$-Goldstein stationarity in $\widetilde{O}(d{\delta^{-1} \epsilon^{-3}})$ gradient oracle calls, where $d$ is the upper-level dimension. Finally, we obtain for the linear inequality setting dimension-free rates of $\widetilde{O}({\delta^{-1} \epsilon^{-4}})$ oracle complexity under the additional assumption of oracle access to the optimal dual variable. Along the way, we develop new nonsmooth nonconvex optimization methods with inexact oracles. We verify these guarantees with preliminary numerical experiments.
Authors: Joshua Durso-Finley, Berardino Barile, Jean-Pierre Falet, Douglas L. Arnold, Nick Pawlowski, Tal Arbel
Abstract: Personalized medicine based on medical images, including predicting future individualized clinical disease progression and treatment response, would have an enormous impact on healthcare and drug development, particularly for diseases (e.g. multiple sclerosis (MS)) with long term, complex, heterogeneous evolutions and no cure. In this work, we present the first stochastic causal temporal framework to model the continuous temporal evolution of disease progression via Neural Stochastic Differential Equations (NSDE). The proposed causal inference model takes as input the patient's high dimensional images (MRI) and tabular data, and predicts both factual and counterfactual progression trajectories on different treatments in latent space. The NSDE permits the estimation of high-confidence personalized trajectories and treatment effects. Extensive experiments were performed on a large, multi-centre, proprietary dataset of patient 3D MRI and clinical data acquired during several randomized clinical trials for MS treatments. Our results present the first successful uncertainty-based causal Deep Learning (DL) model to: (a) accurately predict future patient MS disability evolution (e.g. EDSS) and treatment effects leveraging baseline MRI, and (b) permit the discovery of subgroups of patients for which the model has high confidence in their response to treatment even in clinical trials which did not reach their clinical endpoints.
Authors: Siemen Brussee, Giorgio Buzzanca, Anne M. R. Schrader, Jesper Kers
Abstract: Histopathological analysis of Whole Slide Images (WSIs) has seen a surge in the utilization of deep learning methods, particularly Convolutional Neural Networks (CNNs). However, CNNs often fall short in capturing the intricate spatial dependencies inherent in WSIs. Graph Neural Networks (GNNs) present a promising alternative, adept at directly modeling pairwise interactions and effectively discerning the topological tissue and cellular structures within WSIs. Recognizing the pressing need for deep learning techniques that harness the topological structure of WSIs, the application of GNNs in histopathology has experienced rapid growth. In this comprehensive review, we survey GNNs in histopathology, discuss their applications, and exploring emerging trends that pave the way for future advancements in the field. We begin by elucidating the fundamentals of GNNs and their potential applications in histopathology. Leveraging quantitative literature analysis, we identify four emerging trends: Hierarchical GNNs, Adaptive Graph Structure Learning, Multimodal GNNs, and Higher-order GNNs. Through an in-depth exploration of these trends, we offer insights into the evolving landscape of GNNs in histopathological analysis. Based on our findings, we propose future directions to propel the field forward. Our analysis serves to guide researchers and practitioners towards innovative approaches and methodologies, fostering advancements in histopathological analysis through the lens of graph neural networks.
Authors: Seyedarmin Azizi, Souvik Kundu, Massoud Pedram
Abstract: Low-rank adaptation (LoRA) has become the default approach to fine-tune large language models (LLMs) due to its significant reduction in trainable parameters. However, trainable parameter demand for LoRA increases with increasing model embedding dimensions, leading to high compute costs. Additionally, its backward updates require storing high-dimensional intermediate activations and optimizer states, demanding high peak GPU memory. In this paper, we introduce large model fine-tuning via spectrally decomposed low-dimensional adaptation (LaMDA), a novel approach to fine-tuning large language models, which leverages low-dimensional adaptation to achieve significant reductions in trainable parameters and peak GPU memory footprint. LaMDA freezes a first projection matrix (PMA) in the adaptation path while introducing a low-dimensional trainable square matrix, resulting in substantial reductions in trainable parameters and peak GPU memory usage. LaMDA gradually freezes a second projection matrix (PMB) during the early fine-tuning stages, reducing the compute cost associated with weight updates to enhance parameter efficiency further. We also present an enhancement, LaMDA++, incorporating a ``lite-weight" adaptive rank allocation for the LoRA path via normalized spectrum analysis of pre-trained model weights. We evaluate LaMDA/LaMDA++ across various tasks, including natural language understanding with the GLUE benchmark, text summarization, natural language generation, and complex reasoning on different LLMs. Results show that LaMDA matches or surpasses the performance of existing alternatives while requiring up to 17.7x fewer parameter updates and up to 1.32x lower peak GPU memory usage during fine-tuning. Code will be publicly available.
Authors: Liyuan Xu, Heishiro Kanagawa, Arthur Gretton
Abstract: Proxy causal learning (PCL) is a method for estimating the causal effect of treatments on outcomes in the presence of unobserved confounding, using proxies (structured side information) for the confounder. This is achieved via two-stage regression: in the first stage, we model relations among the treatment and proxies; in the second stage, we use this model to learn the effect of treatment on the outcome, given the context provided by the proxies. PCL guarantees recovery of the true causal effect, subject to identifiability conditions. We propose a novel method for PCL, the deep feature proxy variable method (DFPV), to address the case where the proxies, treatments, and outcomes are high-dimensional and have nonlinear complex relationships, as represented by deep neural network features. We show that DFPV outperforms recent state-of-the-art PCL methods on challenging synthetic benchmarks, including settings involving high dimensional image data. Furthermore, we show that PCL can be applied to off-policy evaluation for the confounded bandit problem, in which DFPV also exhibits competitive performance.
Authors: Alex Shtoff
Abstract: Model training algorithms which observe a small portion of the training set in each computational step are ubiquitous in practical machine learning, and include both stochastic and online optimization methods. In the vast majority of cases, such algorithms typically observe the training samples via the gradients of the cost functions the samples incur. Thus, these methods exploit are the slope of the cost functions via their first-order approximations. To address limitations of gradient-based methods, such as sensitivity to step-size choice in the stochastic setting, or inability to use small function variability in the online setting, several streams of research attempt to exploit more information about the cost functions than just their gradients via the well-known proximal operators. However, implementing such methods in practice poses a challenge, since each iteration step boils down to computing the proximal operator, which may not be easy. In this work we devise a novel algorithmic framework, which exploits convex duality theory to achieve both algorithmic efficiency and software modularity of proximal operator implementations, in order to make experimentation with incremental proximal optimization algorithms accessible to a larger audience of researchers and practitioners, by reducing the gap between their theoretical description in research papers and their use in practice. We provide a reference Python implementation for the framework developed in this paper as an open source library at on https://github.com/alexshtf/inc_prox_pt/releases/tag/prox_pt_paper, along with examples which demonstrate our implementation on a variety of problems, and reproduce the numerical experiments in this paper. The pure Python reference implementation is not necessarily the most efficient, but is a basis for creating efficient implementations by combining Python with a native backend.
URLs: https://github.com/alexshtf/inc_prox_pt/releases/tag/prox_pt_paper,
Authors: Nate Gruver, Marc Finzi, Micah Goldblum, Andrew Gordon Wilson
Abstract: Equivariance guarantees that a model's predictions capture key symmetries in data. When an image is translated or rotated, an equivariant model's representation of that image will translate or rotate accordingly. The success of convolutional neural networks has historically been tied to translation equivariance directly encoded in their architecture. The rising success of vision transformers, which have no explicit architectural bias towards equivariance, challenges this narrative and suggests that augmentations and training data might also play a significant role in their performance. In order to better understand the role of equivariance in recent vision models, we introduce the Lie derivative, a method for measuring equivariance with strong mathematical foundations and minimal hyperparameters. Using the Lie derivative, we study the equivariance properties of hundreds of pretrained models, spanning CNNs, transformers, and Mixer architectures. The scale of our analysis allows us to separate the impact of architecture from other factors like model size or training method. Surprisingly, we find that many violations of equivariance can be linked to spatial aliasing in ubiquitous network layers, such as pointwise non-linearities, and that as models get larger and more accurate they tend to display more equivariance, regardless of architecture. For example, transformers can be more equivariant than convolutional neural networks after training.
Authors: Stefanos Laskaridis, Stylianos I. Venieris, Alexandros Kouris, Rui Li, Nicholas D. Lane
Abstract: In the last decade, Deep Learning has rapidly infiltrated the consumer end, mainly thanks to hardware acceleration across devices. However, as we look towards the future, it is evident that isolated hardware will be insufficient. Increasingly complex AI tasks demand shared resources, cross-device collaboration, and multiple data types, all without compromising user privacy or quality of experience. To address this, we introduce a novel paradigm centered around EdgeAI-Hub devices, designed to reorganise and optimise compute resources and data access at the consumer edge. To this end, we lay a holistic foundation for the transition from on-device to Edge-AI serving systems in consumer environments, detailing their components, structure, challenges and opportunities.
Authors: Philipp Becker, Sebastian Mossburger, Fabian Otto, Gerhard Neumann
Abstract: Learning self-supervised representations using reconstruction or contrastive losses improves performance and sample complexity of image-based and multimodal reinforcement learning (RL). Here, different self-supervised loss functions have distinct advantages and limitations depending on the information density of the underlying sensor modality. Reconstruction provides strong learning signals but is susceptible to distractions and spurious information. While contrastive approaches can ignore those, they may fail to capture all relevant details and can lead to representation collapse. For multimodal RL, this suggests that different modalities should be treated differently based on the amount of distractions in the signal. We propose Contrastive Reconstructive Aggregated representation Learning (CoRAL), a unified framework enabling us to choose the most appropriate self-supervised loss for each sensor modality and allowing the representation to better focus on relevant aspects. We evaluate CoRAL's benefits on a wide range of tasks with images containing distractions or occlusions, a new locomotion suite, and a challenging manipulation suite with visually realistic distractions. Our results show that learning a multimodal representation by combining contrastive and reconstruction-based losses can significantly improve performance and solve tasks that are out of reach for more naive representation learning approaches and other recent baselines.
Authors: Ziyu Cao, Kailai Li
Abstract: We present a principled study on defining Gaussian processes (GPs) with inputs on the product of directional manifolds. A circular kernel is first presented according to the von Mises distribution. Based thereon, the hypertoroidal von Mises (HvM) kernel is proposed to establish GPs on hypertori with consideration of correlated circular components. The proposed HvM kernel is demonstrated with multi-output GP regression for learning vector-valued functions on hypertori using the intrinsic coregionalization model. Analytic derivatives for hyperparameter optimization are provided for runtime-critical applications. For evaluation, we synthesize a ranging-based sensor network and employ the HvM-based GPs for data-driven recursive localization. Numerical results show that the HvM-based GP achieves superior tracking accuracy compared to parametric model and GPs of conventional kernel designs.
Authors: Haoran Li, Jingfeng Wu, Vladimir Braverman
Abstract: We consider a continual learning (CL) problem with two linear regression tasks in the fixed design setting, where the feature vectors are assumed fixed and the labels are assumed to be random variables. We consider an $\ell_2$-regularized CL algorithm, which computes an Ordinary Least Squares parameter to fit the first dataset, then computes another parameter that fits the second dataset under an $\ell_2$-regularization penalizing its deviation from the first parameter, and outputs the second parameter. For this algorithm, we provide tight bounds on the average risk over the two tasks. Our risk bounds reveal a provable trade-off between forgetting and intransigence of the $\ell_2$-regularized CL algorithm: with a large regularization parameter, the algorithm output forgets less information about the first task but is intransigent to extract new information from the second task; and vice versa. Our results suggest that catastrophic forgetting could happen for CL with dissimilar tasks (under a precise similarity measurement) and that a well-tuned $\ell_2$-regularization can partially mitigate this issue by introducing intransigence.
Authors: Kostis Gourgoulias, Najah Ghalyan, Maxime Labonne, Yash Satsangi, Sean Moran, Joseph Sabelja
Abstract: This paper introduces an unsupervised method to estimate the class separability of text datasets from a topological point of view. Using persistent homology, we demonstrate how tracking the evolution of embedding manifolds during training can inform about class separability. More specifically, we show how this technique can be applied to detect when the training process stops improving the separability of the embeddings. Our results, validated across binary and multi-class text classification tasks, show that the proposed method's estimates of class separability align with those obtained from supervised methods. This approach offers a novel perspective on monitoring and improving the fine-tuning of sentence transformers for classification tasks, particularly in scenarios where labeled data is scarce. We also discuss how tracking these quantities can provide additional insights into the properties of the trained classifier.
Authors: Tuo Zhang, Tiantian Feng, Samiul Alam, Dimitrios Dimitriadis, Sunwoo Lee, Mi Zhang, Shrikanth S. Narayanan, Salman Avestimehr
Abstract: In this work, we propose GPT-FL, a generative pre-trained model-assisted federated learning (FL) framework. At its core, GPT-FL leverages generative pre-trained models to generate diversified synthetic data. These generated data are used to train a downstream model on the server, which is then fine-tuned with private client data under the standard FL framework. We show that GPT-FL consistently outperforms state-of-the-art FL methods in terms of model test accuracy, communication efficiency, and client sampling efficiency. Through comprehensive ablation analysis across various data modalities, we discover that the downstream model generated by synthetic data plays a crucial role in controlling the direction of gradient diversity during FL training, which enhances convergence speed and contributes to the notable accuracy boost observed with GPT-FL. Also, regardless of whether the target data falls within or outside the domain of the pre-trained generative model, GPT-FL consistently achieves significant performance gains, surpassing the results obtained by models trained solely with FL or synthetic data. The code is available at https://github.com/AvestimehrResearchGroup/GPT-FL.
Authors: Weiye Zhao, Rui Chen, Yifan Sun, Tianhao Wei, Changliu Liu
Abstract: Reinforcement Learning (RL) algorithms have shown tremendous success in simulation environments, but their application to real-world problems faces significant challenges, with safety being a major concern. In particular, enforcing state-wise constraints is essential for many challenging tasks such as autonomous driving and robot manipulation. However, existing safe RL algorithms under the framework of Constrained Markov Decision Process (CMDP) do not consider state-wise constraints. To address this gap, we propose State-wise Constrained Policy Optimization (SCPO), the first general-purpose policy search algorithm for state-wise constrained reinforcement learning. SCPO provides guarantees for state-wise constraint satisfaction in expectation. In particular, we introduce the framework of Maximum Markov Decision Process, and prove that the worst-case safety violation is bounded under SCPO. We demonstrate the effectiveness of our approach on training neural network policies for extensive robot locomotion tasks, where the agent must satisfy a variety of state-wise safety constraints. Our results show that SCPO significantly outperforms existing methods and can handle state-wise constraints in high-dimensional robotics tasks.
Authors: Siqi Deng, Emily Diana, Michael Kearns, Aaron Roth
Abstract: We study the problem of collecting a cohort or set that is balanced with respect to sensitive groups when group membership is unavailable or prohibited from use at deployment time. Specifically, our deployment-time collection mechanism does not reveal significantly more about the group membership of any individual sample than can be ascertained from base rates alone. To do this, we study a learner that can use a small set of labeled data to train a proxy function that can later be used for this filtering or selection task. We then associate the range of the proxy function with sampling probabilities; given a new example, we classify it using our proxy function and then select it with probability corresponding to its proxy classification. Importantly, we require that the proxy classification does not reveal significantly more information about the sensitive group membership of any individual example compared to population base rates alone (i.e., the level of disclosure should be controlled) and show that we can find such a proxy in a sample- and oracle-efficient manner. Finally, we experimentally evaluate our algorithm and analyze its generalization properties.
Authors: Xunlian Luo, Chunjiang Zhu, Detian Zhang, Qing Li
Abstract: Traffic prediction has been an active research topic in the domain of spatial-temporal data mining. Accurate real-time traffic prediction is essential to improve the safety, stability, and versatility of smart city systems, i.e., traffic control and optimal routing. The complex and highly dynamic spatial-temporal dependencies make effective predictions still face many challenges. Recent studies have shown that spatial-temporal graph neural networks exhibit great potential applied to traffic prediction, which combines sequential models with graph convolutional networks to jointly model temporal and spatial correlations. However, a survey study of graph learning, spatial-temporal graph models for traffic, as well as a fair comparison of baseline models are pending and unavoidable issues. In this paper, we first provide a systematic review of graph learning strategies and commonly used graph convolution algorithms. Then we conduct a comprehensive analysis of the strengths and weaknesses of recently proposed spatial-temporal graph network models. Furthermore, we build a study called STG4Traffic using the deep learning framework PyTorch to establish a standardized and scalable benchmark on two types of traffic datasets. We can evaluate their performance by personalizing the model settings with uniform metrics. Finally, we point out some problems in the current study and discuss future directions. Source codes are available at https://github.com/trainingl/STG4Traffic.
Authors: Qi Yan, Zhengyang Liang, Yang Song, Renjie Liao, Lele Wang
Abstract: Diffusion models based on permutation-equivariant networks can learn permutation-invariant distributions for graph data. However, in comparison to their non-invariant counterparts, we have found that these invariant models encounter greater learning challenges since 1) their effective target distributions exhibit more modes; 2) their optimal one-step denoising scores are the score functions of Gaussian mixtures with more components. Motivated by this analysis, we propose a non-invariant diffusion model, called $\textit{SwinGNN}$, which employs an efficient edge-to-edge 2-WL message passing network and utilizes shifted window based self-attention inspired by SwinTransformers. Further, through systematic ablations, we identify several critical training and sampling techniques that significantly improve the sample quality of graph generation. At last, we introduce a simple post-processing trick, $\textit{i.e.}$, randomly permuting the generated graphs, which provably converts any graph generative model to a permutation-invariant one. Extensive experiments on synthetic and real-world protein and molecule datasets show that our SwinGNN achieves state-of-the-art performances. Our code is released at https://github.com/qiyan98/SwinGNN.
Authors: Guihong Li, Duc Hoang, Kartikeya Bhardwaj, Ming Lin, Zhangyang Wang, Radu Marculescu
Abstract: Recently, zero-shot (or training-free) Neural Architecture Search (NAS) approaches have been proposed to liberate NAS from the expensive training process. The key idea behind zero-shot NAS approaches is to design proxies that can predict the accuracy of some given networks without training the network parameters. The proxies proposed so far are usually inspired by recent progress in theoretical understanding of deep learning and have shown great potential on several datasets and NAS benchmarks. This paper aims to comprehensively review and compare the state-of-the-art (SOTA) zero-shot NAS approaches, with an emphasis on their hardware awareness. To this end, we first review the mainstream zero-shot proxies and discuss their theoretical underpinnings. We then compare these zero-shot proxies through large-scale experiments and demonstrate their effectiveness in both hardware-aware and hardware-oblivious NAS scenarios. Finally, we point out several promising ideas to design better proxies. Our source code and the list of related papers are available on https://github.com/SLDGroup/survey-zero-shot-nas.
Authors: Banghee So
Abstract: The property and casualty (P&C) insurance industry faces challenges in developing claim predictive models due to the highly right-skewed distribution of positive claims with excess zeros. To address this, actuarial science researchers have employed "zero-inflated" models that combine a traditional count model and a binary model. This paper investigates the use of boosting algorithms to process insurance claim data, including zero-inflated telematics data, to construct claim frequency models. Three popular gradient boosting libraries - XGBoost, LightGBM, and CatBoost - are evaluated and compared to determine the most suitable library for training insurance claim data and fitting actuarial frequency models. Through a comprehensive analysis of two distinct datasets, it is determined that CatBoost is the best for developing auto claim frequency models based on predictive performance. Furthermore, we propose a new zero-inflated Poisson boosted tree model, with variation in the assumption about the relationship between inflation probability $p$ and distribution mean $\mu$, and find that it outperforms others depending on data characteristics. This model enables us to take advantage of particular CatBoost tools, which makes it easier and more convenient to investigate the effects and interactions of various risk features on the frequency model when using telematics data.
Authors: Pranshu Malviya, Gon\c{c}alo Mordido, Aristide Baratin, Reza Babanezhad Harikandeh, Jerry Huang, Simon Lacoste-Julien, Razvan Pascanu, Sarath Chandar
Abstract: Adaptive gradient-based optimizers, notably Adam, have left their mark in training large-scale deep learning models, offering fast convergence and robustness to hyperparameter settings. However, they often struggle with generalization, attributed to their tendency to converge to sharp minima in the loss landscape. To address this, we propose a new memory-augmented version of Adam that encourages exploration towards flatter minima by incorporating a buffer of critical momentum terms during training. This buffer prompts the optimizer to overshoot beyond narrow minima, promoting exploration. Through comprehensive analysis in simple settings, we illustrate the efficacy of our approach in increasing exploration and bias towards flatter minima. We empirically demonstrate that it can improve model performance for image classification on ImageNet and CIFAR10/100, language modelling on Penn Treebank, and online learning tasks on TinyImageNet and 5-dataset. Our code is available at \url{https://github.com/chandar-lab/CMOptimizer}.
Authors: Matthew Zurek, Yudong Chen
Abstract: We study graph clustering in the Stochastic Block Model (SBM) in the presence of both large clusters and small, unrecoverable clusters. Previous convex relaxation approaches achieving exact recovery do not allow any small clusters of size $o(\sqrt{n})$, or require a size gap between the smallest recovered cluster and the largest non-recovered cluster. We provide an algorithm based on semidefinite programming (SDP) which removes these requirements and provably recovers large clusters regardless of the remaining cluster sizes. Mid-sized clusters pose unique challenges to the analysis, since their proximity to the recovery threshold makes them highly sensitive to small noise perturbations and precludes a closed-form candidate solution. We develop novel techniques, including a leave-one-out-style argument which controls the correlation between SDP solutions and noise vectors even when the removal of one row of noise can drastically change the SDP solution. We also develop improved eigenvalue perturbation bounds of potential independent interest. Our results are robust to certain semirandom settings that are challenging for alternative algorithms. Using our gap-free clustering procedure, we obtain efficient algorithms for the problem of clustering with a faulty oracle with superior query complexities, notably achieving $o(n^2)$ sample complexity even in the presence of a large number of small clusters. Our gap-free clustering procedure also leads to improved algorithms for recursive clustering.
Authors: Leyang Zhang, Yaoyu Zhang, Tao Luo
Abstract: Under mild assumptions, we investigate the structure of loss landscape of two-layer neural networks near global minima, determine the set of parameters which recovers the target function, and characterize the gradient flows around it. With novel techniques, our work uncovers some simple aspects of the complicated loss landscape and reveals how model, target function, samples and initialization affect the training dynamics differently. These results concludes that two-layer neural networks can be recovered locally at overparameterization.
Authors: Chenxiao Yang, Qitian Wu, David Wipf, Ruoyu Sun, Junchi Yan
Abstract: A long-standing goal in deep learning has been to characterize the learning behavior of black-box models in a more interpretable manner. For graph neural networks (GNNs), considerable advances have been made in formalizing what functions they can represent, but whether GNNs will learn desired functions during the optimization process remains less clear. To fill this gap, we study their training dynamics in function space. In particular, we find that the gradient descent optimization of GNNs implicitly leverages the graph structure to update the learned function, as can be quantified by a phenomenon which we call \emph{kernel-graph alignment}. We provide theoretical explanations for the emergence of this phenomenon in the overparameterized regime and empirically validate it on real-world GNNs. This finding offers new interpretable insights into when and why the learned GNN functions generalize, highlighting their limitations in heterophilic graphs. Practically, we propose a parameter-free algorithm that directly uses a sparse matrix (i.e. graph adjacency) to update the learned function. We demonstrate that this embarrassingly simple approach can be as effective as GNNs while being orders-of-magnitude faster.
Authors: Nate Gruver, Marc Finzi, Shikai Qiu, Andrew Gordon Wilson
Abstract: By encoding time series as a string of numerical digits, we can frame time series forecasting as next-token prediction in text. Developing this approach, we find that large language models (LLMs) such as GPT-3 and LLaMA-2 can surprisingly zero-shot extrapolate time series at a level comparable to or exceeding the performance of purpose-built time series models trained on the downstream tasks. To facilitate this performance, we propose procedures for effectively tokenizing time series data and converting discrete distributions over tokens into highly flexible densities over continuous values. We argue the success of LLMs for time series stems from their ability to naturally represent multimodal distributions, in conjunction with biases for simplicity, and repetition, which align with the salient features in many time series, such as repeated seasonal trends. We also show how LLMs can naturally handle missing data without imputation through non-numerical text, accommodate textual side information, and answer questions to help explain predictions. While we find that increasing model size generally improves performance on time series, we show GPT-4 can perform worse than GPT-3 because of how it tokenizes numbers, and poor uncertainty calibration, which is likely the result of alignment interventions such as RLHF.
Authors: James K Ruffle, Henry Watkins, Robert J Gray, Harpreet Hyare, Michel Thiebaut de Schotten, Parashkev Nachev
Abstract: The architecture of the brain is too complex to be intuitively surveyable without the use of compressed representations that project its variation into a compact, navigable space. The task is especially challenging with high-dimensional data, such as gene expression, where the joint complexity of anatomical and transcriptional patterns demands maximum compression. Established practice is to use standard principal component analysis (PCA), whose computational felicity is offset by limited expressivity, especially at great compression ratios. Employing whole-brain, voxel-wise Allen Brain Atlas transcription data, here we systematically compare compressed representations based on the most widely supported linear and non-linear methods-PCA, kernel PCA, non-negative matrix factorization (NMF), t-stochastic neighbour embedding (t-SNE), uniform manifold approximation and projection (UMAP), and deep auto-encoding-quantifying reconstruction fidelity, anatomical coherence, and predictive utility with respect to signalling, microstructural, and metabolic targets. We show that deep auto-encoders yield superior representations across all metrics of performance and target domains, supporting their use as the reference standard for representing transcription patterns in the human brain.
Authors: Wassim Tenachi, Rodrigo Ibata, Thibaut L. Fran\c{c}ois, Foivos I. Diakogiannis
Abstract: We introduce 'Class Symbolic Regression' (Class SR) a first framework for automatically finding a single analytical functional form that accurately fits multiple datasets - each realization being governed by its own (possibly) unique set of fitting parameters. This hierarchical framework leverages the common constraint that all the members of a single class of physical phenomena follow a common governing law. Our approach extends the capabilities of our earlier Physical Symbolic Optimization ($\Phi$-SO) framework for Symbolic Regression, which integrates dimensional analysis constraints and deep reinforcement learning for unsupervised symbolic analytical function discovery from data. Additionally, we introduce the first Class SR benchmark, comprising a series of synthetic physical challenges specifically designed to evaluate such algorithms. We demonstrate the efficacy of our novel approach by applying it to these benchmark challenges and showcase its practical utility for astrophysics by successfully extracting an analytic galaxy potential from a set of simulated orbits approximating stellar streams.
Authors: Wenzhao Jiang, Hao Liu, Hui Xiong
Abstract: Graph Neural Networks (GNNs) have emerged as powerful representation learning tools for capturing complex dependencies within diverse graph-structured data. Despite their success in a wide range of graph mining tasks, GNNs have raised serious concerns regarding their trustworthiness, including susceptibility to distribution shift, biases towards certain populations, and lack of explainability. Recently, integrating causal learning techniques into GNNs has sparked numerous ground-breaking studies since many GNN trustworthiness issues can be alleviated by capturing the underlying data causality rather than superficial correlations. In this survey, we comprehensively review recent research efforts on Causality-Inspired GNNs (CIGNNs). Specifically, we first employ causal tools to analyze the primary trustworthiness risks of existing GNNs, underscoring the necessity for GNNs to comprehend the causal mechanisms within graph data. Moreover, we introduce a taxonomy of CIGNNs based on the type of causal learning capability they are equipped with, i.e., causal reasoning and causal representation learning. Besides, we systematically introduce typical methods within each category and discuss how they mitigate trustworthiness risks. Finally, we summarize useful resources and discuss several future directions, hoping to shed light on new research opportunities in this emerging field. The representative papers, along with open-source data and codes, are available in https://github.com/usail-hkust/Causality-Inspired-GNNs.
URLs: https://github.com/usail-hkust/Causality-Inspired-GNNs.
Authors: Tianyi Chen, Yingzhou Lu, Nan Hao, Capucine Van Rechem, Jintai Chen, Tianfan Fu
Abstract: The importance of uncertainty quantification is increasingly recognized in the diverse field of machine learning. Accurately assessing model prediction uncertainty can help provide deeper understanding and confidence for researchers and practitioners. This is especially critical in medical diagnosis and drug discovery areas, where reliable predictions directly impact research quality and patient health. In this paper, we proposed incorporating uncertainty quantification into clinical trial outcome predictions. Our main goal is to enhance the model's ability to discern nuanced differences, thereby significantly improving its overall performance. We have adopted a selective classification approach to fulfill our objective, integrating it seamlessly with the Hierarchical Interaction Network (HINT), which is at the forefront of clinical trial prediction modeling. Selective classification, encompassing a spectrum of methods for uncertainty quantification, empowers the model to withhold decision-making in the face of samples marked by ambiguity or low confidence, thereby amplifying the accuracy of predictions for the instances it chooses to classify. A series of comprehensive experiments demonstrate that incorporating selective classification into clinical trial predictions markedly enhances the model's performance, as evidenced by significant upticks in pivotal metrics such as PR-AUC, F1, ROC-AUC, and overall accuracy. Specifically, the proposed method achieved 32.37\%, 21.43\%, and 13.27\% relative improvement on PR-AUC over the base model (HINT) in phase I, II, and III trial outcome prediction, respectively. When predicting phase III, our method reaches 0.9022 PR-AUC scores. These findings illustrate the robustness and prospective utility of this strategy within the area of clinical trial predictions, potentially setting a new benchmark in the field.
Authors: Chenxi Liu, Sun Yang, Qianxiong Xu, Zhishuai Li, Cheng Long, Ziyue Li, Rui Zhao
Abstract: Traffic prediction, an essential component for intelligent transportation systems, endeavours to use historical data to foresee future traffic features at specific locations. Although existing traffic prediction models often emphasize developing complex neural network structures, their accuracy has not improved. Recently, large language models have shown outstanding capabilities in time series analysis. Differing from existing models, LLMs progress mainly through parameter expansion and extensive pretraining while maintaining their fundamental structures. Motivated by these developments, we propose a Spatial-Temporal Large Language Model (ST-LLM) for traffic prediction. In the ST-LLM, we define timesteps at each location as tokens and design a spatial-temporal embedding to learn the spatial location and global temporal patterns of these tokens. Additionally, we integrate these embeddings by a fusion convolution to each token for a unified spatial-temporal representation. Furthermore, we innovate a partially frozen attention strategy to adapt the LLM to capture global spatial-temporal dependencies for traffic prediction. Comprehensive experiments on real traffic datasets offer evidence that ST-LLM is a powerful spatial-temporal learner that outperforms state-of-the-art models. Notably, the ST-LLM also exhibits robust performance in both few-shot and zero-shot prediction scenarios. The code is publicly available at https://github.com/ChenxiLiu-HNU/ST-LLM.
Authors: Wenhao Wang, Muhammad Ahmad Kaleem, Adam Dziedzic, Michael Backes, Nicolas Papernot, Franziska Boenisch
Abstract: Self-supervised learning (SSL) has recently received significant attention due to its ability to train high-performance encoders purely on unlabeled data-often scraped from the internet. This data can still be sensitive and empirical evidence suggests that SSL encoders memorize private information of their training data and can disclose them at inference time. Since existing theoretical definitions of memorization from supervised learning rely on labels, they do not transfer to SSL. To address this gap, we propose SSLMem, a framework for defining memorization within SSL. Our definition compares the difference in alignment of representations for data points and their augmented views returned by both encoders that were trained on these data points and encoders that were not. Through comprehensive empirical analysis on diverse encoder architectures and datasets we highlight that even though SSL relies on large datasets and strong augmentations-both known in supervised learning as regularization techniques that reduce overfitting-still significant fractions of training data points experience high memorization. Through our empirical results, we show that this memorization is essential for encoders to achieve higher generalization performance on different downstream tasks.
Authors: Mathew Huerta-Enochian, Seung Yong Ko
Abstract: We present a novel study analyzing the effects of various prompt loss token weights (PLW) for supervised instruction fine-tuning (SIFT). While prompt-masking (PLW = 0) is common for SIFT, some fine-tuning APIs support fractional PLWs and suggest that using a small non-zero PLW can help stabilize learning when fine-tuning on short-completion data. However, there has never been a study confirming this claim, and OpenAI, a major cloud-based SIFT provider, recently removed this parameter from their fine-tuning API. We found that performance of models fine-tuned on short-completion data had a statistically-significant negative quadratic relationship with PLW. Using small values (0.01 - 0.5) of PLW produced better results on multiple-choice and short-generation benchmarks (outperforming models fine-tuned on long-completion data) while large values (~ 1.0) of PLW produced better results on long-generation benchmarks. We explained this effect and verified its importance through additional experiments. This research serves as a warning to API providers about the importance of providing a PLW parameter for SIFT.
Authors: Viktor Moskvoretskii, Dmitry Osin, Egor Shvetsov, Igor Udovichenko, Maxim Zhelnin, Andrey Dukhovny, Anna Zhimerikina, Evgeny Burnaev
Abstract: This study explores the application of self-supervised learning techniques for event sequences. It is a key modality in various applications such as banking, e-commerce, and healthcare. However, there is limited research on self-supervised learning for event sequences, and methods from other domains like images, texts, and speech may not easily transfer. To determine the most suitable approach, we conduct a detailed comparative analysis of previously identified best-performing methods. We find that neither the contrastive nor generative method is superior. Our assessment includes classifying event sequences, predicting the next event, and evaluating embedding quality. These results further highlight the potential benefits of combining both methods. Given the lack of research on hybrid models in this domain, we initially adapt the baseline model from another domain. However, upon observing its underperformance, we develop a novel method called the Multimodal-Learning Event Model (MLEM). MLEM treats contrastive learning and generative modeling as distinct yet complementary modalities, aligning their embeddings. The results of our study demonstrate that combining contrastive and generative approaches into one procedure with MLEM achieves superior performance across multiple metrics.
Authors: Kotaro Yoshida, Hiroki Naganuma
Abstract: Machine learning models traditionally assume that training and test data are independently and identically distributed. However, in real-world applications, the test distribution often differs from training. This problem, known as out-of-distribution (OOD) generalization, challenges conventional models. Invariant Risk Minimization (IRM) emerges as a solution that aims to identify invariant features across different environments to enhance OOD robustness. However, IRM's complexity, particularly its bi-level optimization, has led to the development of various approximate methods. Our study investigates these approximate IRM techniques, using the consistency and variance of calibration across environments as metrics to measure the invariance aimed for by IRM. Calibration, which measures the reliability of model prediction, serves as an indicator of whether models effectively capture environment-invariant features by showing how uniformly over-confident the model remains across varied environments. Through a comparative analysis of datasets with distributional shifts, we observe that Information Bottleneck-based IRM achieves consistent calibration across different environments. This observation suggests that information compression techniques, such as IB, are potentially effective in achieving model invariance. Furthermore, our empirical evidence indicates that models exhibiting consistent calibration across environments are also well-calibrated. This demonstrates that invariance and cross-environment calibration are empirically equivalent. Additionally, we underscore the necessity for a systematic approach to evaluating OOD generalization. This approach should move beyond traditional metrics, such as accuracy and F1 scores, which fail to account for the model's degree of over-confidence, and instead focus on the nuanced interplay between accuracy, calibration, and model invariance.
Authors: Takashi Morita
Abstract: This study reports an unintuitive finding that positional encoding enhances learning of recurrent neural networks (RNNs). Positional encoding is a high-dimensional representation of time indices on input data. Most famously, positional encoding complements the capabilities of Transformer neural networks, which lack an inherent mechanism for representing the data order. By contrast, RNNs can encode the temporal information of data points on their own, rendering their use of positional encoding seemingly redundant/unnecessary. Nonetheless, investigations through synthetic benchmarks reveal an advantage of coupling positional encoding and RNNs, especially for handling a large vocabulary that yields low-frequency tokens. Further scrutinization unveils that these low-frequency tokens destabilizes the gradients of vanilla RNNs, and the positional encoding resolves this instability. These results shed a new light on the utility of positional encoding beyond its canonical role as a timekeeper for Transformers.
Authors: Yue Xing, Xiaofeng Lin, Chenheng Xu, Namjoon Suh, Qifan Song, Guang Cheng
Abstract: Large language models (LLMs) are powerful models that can learn concepts at the inference stage via in-context learning (ICL). While theoretical studies, e.g., \cite{zhang2023trained}, attempt to explain the mechanism of ICL, they assume the input $x_i$ and the output $y_i$ of each demonstration example are in the same token (i.e., structured data). However, in real practice, the examples are usually text input, and all words, regardless of their logic relationship, are stored in different tokens (i.e., unstructured data \cite{wibisono2023role}). To understand how LLMs learn from the unstructured data in ICL, this paper studies the role of each component in the transformer architecture and provides a theoretical understanding to explain the success of the architecture. In particular, we consider a simple transformer with one/two attention layers and linear regression tasks for the ICL prediction. We observe that (1) a transformer with two layers of (self-)attentions with a look-ahead attention mask can learn from the prompt in the unstructured data, and (2) positional encoding can match the $x_i$ and $y_i$ tokens to achieve a better ICL performance.
Authors: Zelong Li, Wenyue Hua, Hao Wang, He Zhu, Yongfeng Zhang
Abstract: Recent advancements on Large Language Models (LLMs) enable AI Agents to automatically generate and execute multi-step plans to solve complex tasks. However, since LLM's content generation process is hardly controllable, current LLM-based agents frequently generate invalid or non-executable plans, which jeopardizes the performance of the generated plans and corrupts users' trust in LLM-based agents. In response, this paper proposes a novel ``Formal-LLM'' framework for LLM-based agents by integrating the expressiveness of natural language and the precision of formal language. Specifically, the framework allows human users to express their requirements or constraints for the planning process as an automaton. A stack-based LLM plan generation process is then conducted under the supervision of the automaton to ensure that the generated plan satisfies the constraints, making the planning process controllable. We conduct experiments on both benchmark tasks and practical real-life tasks, and our framework achieves over 50% overall performance increase, which validates the feasibility and effectiveness of employing Formal-LLM to guide the plan generation of agents, preventing the agents from generating invalid and unsuccessful plans. Further, more controllable LLM-based agents can facilitate the broader utilization of LLM in application scenarios where high validity of planning is essential. The work is open-sourced at https://github.com/agiresearch/Formal-LLM.
Authors: Yongshuo Zong, Ondrej Bohdal, Tingyang Yu, Yongxin Yang, Timothy Hospedales
Abstract: Current vision large language models (VLLMs) exhibit remarkable capabilities yet are prone to generate harmful content and are vulnerable to even the simplest jailbreaking attacks. Our initial analysis finds that this is due to the presence of harmful data during vision-language instruction fine-tuning, and that VLLM fine-tuning can cause forgetting of safety alignment previously learned by the underpinning LLM. To address this issue, we first curate a vision-language safe instruction-following dataset VLGuard covering various harmful categories. Our experiments demonstrate that integrating this dataset into standard vision-language fine-tuning or utilizing it for post-hoc fine-tuning effectively safety aligns VLLMs. This alignment is achieved with minimal impact on, or even enhancement of, the models' helpfulness. The versatility of our safety fine-tuning dataset makes it a valuable resource for safety-testing existing VLLMs, training new models or safeguarding pre-trained VLLMs. Empirical results demonstrate that fine-tuned VLLMs effectively reject unsafe instructions and substantially reduce the success rates of several black-box adversarial attacks, which approach zero in many cases. The code and dataset are available at https://github.com/ys-zong/VLGuard.
Authors: Ravi Patel, Angus Brayne, Rogier Hintzen, Daniel Jaroslawicz, Georgiana Neculae, Dane Corneil
Abstract: Language models hold incredible promise for enabling scientific discovery by synthesizing massive research corpora. Many complex scientific research questions have multiple plausible answers, each supported by evidence of varying strength. However, existing language models lack the capability to quantitatively and faithfully compare answer plausibility in terms of supporting evidence. To address this issue, we introduce Retrieve to Explain (R2E), a retrieval-based language model. R2E scores and ranks all possible answers to a research question based on evidence retrieved from a document corpus. The architecture represents each answer only in terms of its supporting evidence, with the answer itself masked. This allows us to extend feature attribution methods, such as Shapley values, to transparently attribute each answer's score back to its supporting evidence at inference time. The architecture also allows R2E to incorporate new evidence without retraining, including non-textual data modalities templated into natural language. We assess on the challenging task of drug target identification from scientific literature, a human-in-the-loop process where failures are extremely costly and explainability is paramount. When predicting whether drug targets will subsequently be confirmed as efficacious in clinical trials, R2E not only matches non-explainable literature-based models but also surpasses a genetics-based target identification approach used throughout the pharmaceutical industry.
Authors: Lunyiu Nie, Zhimin Ding, Erdong Hu, Christopher Jermaine, Swarat Chaudhuri
Abstract: Large Language Models (LLMs) have a natural role in answering complex queries about data streams, but the high computational cost of LLM inference makes them infeasible in many such tasks. We propose online cascade learning, the first approach to address this challenge. The objective here is to learn a "cascade" of models, starting with lower-capacity models (such as logistic regression) and ending with a powerful LLM, along with a deferral policy that determines the model to be used on a given input. We formulate the task of learning cascades online as an imitation-learning problem, where smaller models are updated over time imitating the collected LLM demonstrations, and give a no-regret algorithm for the problem. Experimental results across four benchmarks show that our method parallels LLMs in accuracy while cutting down inference costs by as much as 90% with strong robustness against input distribution shifts, underscoring its efficacy and adaptability in stream processing.
Authors: Yuchen Zhang, Tianle Zhang, Kai Wang, Ziyao Guo, Yuxuan Liang, Xavier Bresson, Wei Jin, Yang You
Abstract: Graph condensation aims to reduce the size of a large-scale graph dataset by synthesizing a compact counterpart without sacrificing the performance of Graph Neural Networks (GNNs) trained on it, which has shed light on reducing the computational cost for training GNNs. Nevertheless, existing methods often fall short of accurately replicating the original graph for certain datasets, thereby failing to achieve the objective of lossless condensation. To understand this phenomenon, we investigate the potential reasons and reveal that the previous state-of-the-art trajectory matching method provides biased and restricted supervision signals from the original graph when optimizing the condensed one. This significantly limits both the scale and efficacy of the condensed graph. In this paper, we make the first attempt toward \textit{lossless graph condensation} by bridging the previously neglected supervision signals. Specifically, we employ a curriculum learning strategy to train expert trajectories with more diverse supervision signals from the original graph, and then effectively transfer the information into the condensed graph with expanding window matching. Moreover, we design a loss function to further extract knowledge from the expert trajectories. Theoretical analysis justifies the design of our method and extensive experiments verify its superiority across different datasets. Code is released at https://github.com/NUS-HPC-AI-Lab/GEOM.
Authors: Zhenlong Liu, Lei Feng, Huiping Zhuang, Xiaofeng Cao, Hongxin Wei
Abstract: Machine learning models are susceptible to membership inference attacks (MIAs), which aim to infer whether a sample is in the training set. Existing work utilizes gradient ascent to enlarge the loss variance of training data, alleviating the privacy risk. However, optimizing toward a reverse direction may cause the model parameters to oscillate near local minima, leading to instability and suboptimal performance. In this work, we propose a novel method -- Convex-Concave Loss, which enables a high variance of training loss distribution by gradient descent. Our method is motivated by the theoretical analysis that convex losses tend to decrease the loss variance during training. Thus, our key idea behind CCL is to reduce the convexity of loss functions with a concave term. Trained with CCL, neural networks produce losses with high variance for training data, reinforcing the defense against MIAs. Extensive experiments demonstrate the superiority of CCL, achieving state-of-the-art balance in the privacy-utility trade-off.
Authors: Eliad Tsfadia
Abstract: Private data analysis faces a significant challenge known as the curse of dimensionality, leading to increased costs. However, many datasets possess an inherent low-dimensional structure. For instance, during optimization via gradient descent, the gradients frequently reside near a low-dimensional subspace. If the low-dimensional structure could be privately identified using a small amount of points, we could avoid paying for the high ambient dimension. On the negative side, Dwork, Talwar, Thakurta, and Zhang (STOC 2014) proved that privately estimating subspaces, in general, requires an amount of points that has a polynomial dependency on the dimension. However, their bound do not rule out the possibility to reduce the number of points for "easy'' instances. Yet, providing a measure that captures how much a given dataset is "easy'' for this task turns out to be challenging, and was not properly addressed in prior works. Inspired by the work of Singhal and Steinke (NeurIPS 2021), we provide the first measures that quantify easiness as a function of multiplicative singular-value gaps in the input dataset, and support them with new upper and lower bounds. In particular, our results determine the first type of gap that is sufficient and necessary for estimating a subspace with an amount of points that is independent of the dimension. Furthermore, we realize our upper bounds using a practical algorithm and demonstrate its advantage in high-dimensional regimes compared to prior approaches.
Authors: Hongyun Zhou, Xiangyu Lu, Wang Xu, Conghui Zhu, Tiejun Zhao, Muyun Yang
Abstract: Low-Rank Adaptation (LoRA) is currently the most commonly used Parameter-efficient fine-tuning (PEFT) method, it introduces auxiliary parameters for each layer to fine-tune the pre-trained model under limited computing resources. However, it still faces resource consumption challenges during training when scaling up to larger models. Most previous studies have tackled this issue by using pruning techniques, which involve removing LoRA parameters deemed unimportant. Nonetheless, these efforts only analyze LoRA parameter features to evaluate their importance, such as parameter count, size, and gradient. In fact, the output of LoRA (product of LoRA parameter and hidden state), directly impacts the final results. Preliminary experiments indicate that a fraction of LoRA elements possesses significantly high output values, substantially influencing the layer output. Motivated by the observation, we propose LoRA-drop. Concretely, LoRA-drop evaluates the importance of LoRA based on the LoRA output. Then we retain LoRA for important layers and the other layers share the same LoRA. We conduct abundant experiments with models of different scales on NLU and NLG tasks. Results demonstrate that LoRA-drop can achieve performance comparable to full fine-tuning and LoRA, while retaining 50\% of the LoRA parameters on average.
Authors: Junhao Zheng, Shengjie Qiu, Qianli Ma
Abstract: Large Language Models (LLMs) have achieved remarkable success across various tasks, yet their ability to learn incrementally without forgetting remains underexplored. Incremental learning (IL) is crucial as it enables models to acquire new knowledge while retaining previously learned information, akin to human learning. Existing benchmarks for IL are insufficient due to data leakage issues and the overqualification of LLMs. To address these challenges, we introduce Concept-1K, a novel dataset comprising 1,023 recently emerged concepts across diverse domains. The concepts in Concept-1K are discrete, interpretable units of knowledge that allow for fine-grained analysis of learning and forgetting processes. Using Concept-1K as a testbed, we aim to answer the question: ``Can LLMs learn new concepts incrementally without forgetting like humans?'' Our investigation reveals that LLMs still suffer from catastrophic forgetting and that LoRA, despite fine-tuning fewer parameters, may lead to more forgetting on training data. Additionally, we explore the roles of in-context learning, model scale, buffer size, and pretraining in IL performance. These findings highlight the strengths and limitations of LLMs in IL scenarios and provide a robust benchmark for future research.
Authors: Nadav Dym, Hannah Lawrence, Jonathan W. Siegel
Abstract: Canonicalization provides an architecture-agnostic method for enforcing equivariance, with generalizations such as frame-averaging recently gaining prominence as a lightweight and flexible alternative to equivariant architectures. Recent works have found an empirical benefit to using probabilistic frames instead, which learn weighted distributions over group elements. In this work, we provide strong theoretical justification for this phenomenon: for commonly-used groups, there is no efficiently computable choice of frame that preserves continuity of the function being averaged. In other words, unweighted frame-averaging can turn a smooth, non-symmetric function into a discontinuous, symmetric function. To address this fundamental robustness problem, we formally define and construct \emph{weighted} frames, which provably preserve continuity, and demonstrate their utility by constructing efficient and continuous weighted frames for the actions of $SO(2)$, $SO(3)$, and $S_n$ on point clouds.
Authors: Zhiding Liu, Jiqian Yang, Mingyue Cheng, Yucong Luo, Zhi Li
Abstract: Recent efforts have been dedicated to enhancing time series forecasting accuracy by introducing advanced network architectures and self-supervised pretraining strategies. Nevertheless, existing approaches still exhibit two critical drawbacks. Firstly, these methods often rely on a single dataset for training, limiting the model's generalizability due to the restricted scale of the training data. Secondly, the one-step generation schema is widely followed, which necessitates a customized forecasting head and overlooks the temporal dependencies in the output series, and also leads to increased training costs under different horizon length settings. To address these issues, we propose a novel generative pretrained hierarchical transformer architecture for forecasting, named \textbf{GPHT}. There are two aspects of key designs in GPHT. On the one hand, we advocate for constructing a mixed dataset under the channel-independent assumption for pretraining our model, comprising various datasets from diverse data scenarios. This approach significantly expands the scale of training data, allowing our model to uncover commonalities in time series data and facilitating improved transfer to specific datasets. On the other hand, GPHT employs an auto-regressive forecasting approach, effectively modeling temporal dependencies in the output series. Importantly, no customized forecasting head is required, enabling \textit{a single model to forecast at arbitrary horizon settings.} We conduct sufficient experiments on eight datasets with mainstream self-supervised pretraining models and supervised models. The results demonstrated that GPHT surpasses the baseline models across various fine-tuning and zero/few-shot learning settings in the traditional long-term forecasting task. We make our codes publicly available\footnote{https://github.com/icantnamemyself/GPHT}.
Authors: Amey Agrawal, Nitin Kedia, Ashish Panwar, Jayashree Mohan, Nipun Kwatra, Bhargav S. Gulavani, Alexey Tumanov, Ramachandran Ramjee
Abstract: Each LLM serving request goes through two phases. The first is prefill which processes the entire input prompt and produces the first output token and the second is decode which generates the rest of output tokens, one-at-a-time. Prefill iterations have high latency but saturate GPU compute due to parallel processing of the input prompt. In contrast, decode iterations have low latency but also low compute utilization because a decode iteration processes only a single token per request. This makes batching highly effective for decodes and consequently for overall throughput. However, batching multiple requests leads to an interleaving of prefill and decode iterations which makes it challenging to achieve both high throughput and low latency. We introduce an efficient LLM inference scheduler, Sarathi-Serve, to address this throughput-latency tradeoff. Sarathi-Serve introduces chunked-prefills which splits a prefill request into near equal sized chunks and creates stall-free schedules that adds new requests in a batch without pausing ongoing decodes. Stall-free scheduling unlocks the opportunity to improve throughput with large batch sizes while minimizing the effect of batching on latency. Furthermore, uniform batches in Sarathi-Serve ameliorate the imbalance between iterations resulting in minimal pipeline bubbles. Our techniques yield significant improvements in inference performance across models and hardware under tail latency constraints. For Mistral-7B on single A100 GPUs, we achieve 2.6x higher serving capacity and up to 3.7x higher serving capacity for the Yi-34B model on two A100 GPUs as compared to vLLM. When used with pipeline parallelism on Falcon-180B, Sarathi-Serve provides up to 5.6x gain in the end-to-end serving capacity. The source code for Sarathi-Serve is available at https://github.com/microsoft/sarathi-serve.
Authors: Benjamin Eysenbach, Vivek Myers, Ruslan Salakhutdinov, Sergey Levine
Abstract: Given time series data, how can we answer questions like "what will happen in the future?" and "how did we get here?" These sorts of probabilistic inference questions are challenging when observations are high-dimensional. In this paper, we show how these questions can have compact, closed form solutions in terms of learned representations. The key idea is to apply a variant of contrastive learning to time series data. Prior work already shows that the representations learned by contrastive learning encode a probability ratio. By extending prior work to show that the marginal distribution over representations is Gaussian, we can then prove that joint distribution of representations is also Gaussian. Taken together, these results show that representations learned via temporal contrastive learning follow a Gauss-Markov chain, a graphical model where inference (e.g., prediction, planning) over representations corresponds to inverting a low-dimensional matrix. In one special case, inferring intermediate representations will be equivalent to interpolating between the learned representations. We validate our theory using numerical simulations on tasks up to 46-dimensions.
Authors: Xiaowei Qian, Zhimeng Guo, Jialiang Li, Haitao Mao, Bingheng Li, Suhang Wang, Yao Ma
Abstract: Fair graph learning plays a pivotal role in numerous practical applications. Recently, many fair graph learning methods have been proposed; however, their evaluation often relies on poorly constructed semi-synthetic datasets or substandard real-world datasets. In such cases, even a basic Multilayer Perceptron (MLP) can outperform Graph Neural Networks (GNNs) in both utility and fairness. In this work, we illustrate that many datasets fail to provide meaningful information in the edges, which may challenge the necessity of using graph structures in these problems. To address these issues, we develop and introduce a collection of synthetic, semi-synthetic, and real-world datasets that fulfill a broad spectrum of requirements. These datasets are thoughtfully designed to include relevant graph structures and bias information crucial for the fair evaluation of models. The proposed synthetic and semi-synthetic datasets offer the flexibility to create data with controllable bias parameters, thereby enabling the generation of desired datasets with user-defined bias values with ease. Moreover, we conduct systematic evaluations of these proposed datasets and establish a unified evaluation approach for fair graph learning models. Our extensive experimental results with fair graph learning methods across our datasets demonstrate their effectiveness in benchmarking the performance of these methods. Our datasets and the code for reproducing our experiments are available at https://github.com/XweiQ/Benchmark-GraphFairness.
Authors: Narim Jeong, Donghwan Lee
Abstract: Soft Q-learning is a variation of Q-learning designed to solve entropy regularized Markov decision problems where an agent aims to maximize the entropy regularized value function. Despite its empirical success, there have been limited theoretical studies of soft Q-learning to date. This paper aims to offer a novel and unified finite-time, control-theoretic analysis of soft Q-learning algorithms. We focus on two types of soft Q-learning algorithms: one utilizing the log-sum-exp operator and the other employing the Boltzmann operator. By using dynamical switching system models, we derive novel finite-time error bounds for both soft Q-learning algorithms. We hope that our analysis will deepen the current understanding of soft Q-learning by establishing connections with switching system models and may even pave the way for new frameworks in the finite-time analysis of other reinforcement learning algorithms.
Authors: Vincent Herrmann, Francesco Faccio, J\"urgen Schmidhuber
Abstract: Recurrent Neural Networks (RNNs) are general-purpose parallel-sequential computers. The program of an RNN is its weight matrix. How to learn useful representations of RNN weights that facilitate RNN analysis as well as downstream tasks? While the mechanistic approach directly looks at some RNN's weights to predict its behavior, the functionalist approach analyzes its overall functionality-specifically, its input-output mapping. We consider several mechanistic approaches for RNN weights and adapt the permutation equivariant Deep Weight Space layer for RNNs. Our two novel functionalist approaches extract information from RNN weights by 'interrogating' the RNN through probing inputs. We develop a theoretical framework that demonstrates conditions under which the functionalist approach can generate rich representations that help determine RNN behavior. We release the first two 'model zoo' datasets for RNN weight representation learning. One consists of generative models of a class of formal languages, and the other one of classifiers of sequentially processed MNIST digits.With the help of an emulation-based self-supervised learning technique we compare and evaluate the different RNN weight encoding techniques on multiple downstream applications. On the most challenging one, namely predicting which exact task the RNN was trained on, functionalist approaches show clear superiority.
Authors: Yuxuan Liang, Haomin Wen, Yuqi Nie, Yushan Jiang, Ming Jin, Dongjin Song, Shirui Pan, Qingsong Wen
Abstract: Time series analysis stands as a focal point within the data mining community, serving as a cornerstone for extracting valuable insights crucial to a myriad of real-world applications. Recent advances in Foundation Models (FMs) have fundamentally reshaped the paradigm of model design for time series analysis, boosting various downstream tasks in practice. These innovative approaches often leverage pre-trained or fine-tuned FMs to harness generalized knowledge tailored for time series analysis. This survey aims to furnish a comprehensive and up-to-date overview of FMs for time series analysis. While prior surveys have predominantly focused on either application or pipeline aspects of FMs in time series analysis, they have often lacked an in-depth understanding of the underlying mechanisms that elucidate why and how FMs benefit time series analysis. To address this gap, our survey adopts a methodology-centric classification, delineating various pivotal elements of time-series FMs, including model architectures, pre-training techniques, adaptation methods, and data modalities. Overall, this survey serves to consolidate the latest advancements in FMs pertinent to time series analysis, accentuating their theoretical underpinnings, recent strides in development, and avenues for future exploration.
Authors: Michael Chertkov
Abstract: The paper reflects on the future role of AI in scientific research, with a special focus on turbulence studies, and examines the evolution of AI, particularly through Diffusion Models rooted in non-equilibrium statistical mechanics. It underscores the significant impact of AI on advancing reduced, Lagrangian models of turbulence through innovative use of deep neural networks. Additionally, the paper reviews various other AI applications in turbulence research and outlines potential challenges and opportunities in the concurrent advancement of AI and statistical hydrodynamics. This discussion sets the stage for a future where AI and turbulence research are intricately intertwined, leading to more profound insights and advancements in both fields.
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: 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: Afzal Ahmad, Linfeng Du, Zhiyao Xie, Wei Zhang
Abstract: One of the primary challenges impeding the progress of Neural Architecture Search (NAS) is its extensive reliance on exorbitant computational resources. NAS benchmarks aim to simulate runs of NAS experiments at zero cost, remediating the need for extensive compute. However, existing NAS benchmarks use synthetic datasets and model proxies that make simplified assumptions about the characteristics of these datasets and models, leading to unrealistic evaluations. We present a technique that allows searching for training proxies that reduce the cost of benchmark construction by significant margins, making it possible to construct realistic NAS benchmarks for large-scale datasets. Using this technique, we construct an open-source bi-objective NAS benchmark for the ImageNet2012 dataset combined with the on-device performance of accelerators, including GPUs, TPUs, and FPGAs. Through extensive experimentation with various NAS optimizers and hardware platforms, we show that the benchmark is accurate and allows searching for state-of-the-art hardware-aware models at zero cost.
Authors: Ray Zirui Zhang, Xiaohui Xie, John Lowengrub
Abstract: We propose a new neural network based method for solving inverse problems for partial differential equations (PDEs) by formulating the PDE inverse problem as a bilevel optimization problem. At the upper level, we minimize the data loss with respect to the PDE parameters. At the lower level, we train a neural network to locally approximate the PDE solution operator in the neighborhood of a given set of PDE parameters, which enables an accurate approximation of the descent direction for the upper level optimization problem. The lower level loss function includes the L2 norms of both the residual and its derivative with respect to the PDE parameters. We apply gradient descent simultaneously on both the upper and lower level optimization problems, leading to an effective and fast algorithm. The method, which we refer to as BiLO (Bilevel Local Operator learning), is also able to efficiently infer unknown functions in the PDEs through the introduction of an auxiliary variable. Through extensive experiments over multiple PDE systems, we demonstrate that our method enforces strong PDE constraints, is robust to sparse and noisy data, and eliminates the need to balance the residual and the data loss, which is inherent to the soft PDE constraints in many existing methods.
Authors: Marco Bressan, Emmanuel Esposito, Maximilian Thiessen
Abstract: We study the problem of learning a binary classifier on the vertices of a graph. In particular, we consider classifiers given by monophonic halfspaces, partitions of the vertices that are convex in a certain abstract sense. Monophonic halfspaces, and related notions such as geodesic halfspaces,have recently attracted interest, and several connections have been drawn between their properties(e.g., their VC dimension) and the structure of the underlying graph $G$. We prove several novel results for learning monophonic halfspaces in the supervised, online, and active settings. Our main result is that a monophonic halfspace can be learned with near-optimal passive sample complexity in time polynomial in $n = |V(G)|$. This requires us to devise a polynomial-time algorithm for consistent hypothesis checking, based on several structural insights on monophonic halfspaces and on a reduction to $2$-satisfiability. We prove similar results for the online and active settings. We also show that the concept class can be enumerated with delay $\operatorname{poly}(n)$, and that empirical risk minimization can be performed in time $2^{\omega(G)}\operatorname{poly}(n)$ where $\omega(G)$ is the clique number of $G$. These results answer open questions from the literature (Gonz\'alez et al., 2020), and show a contrast with geodesic halfspaces, for which some of the said problems are NP-hard (Seiffarth et al., 2023).
Authors: Tao Wang, Bo Zhao, Sicun Gao, Rose Yu
Abstract: Physics-Informed Neural Networks (PINNs) have gained popularity in scientific computing in recent years. However, they often fail to achieve the same level of accuracy as classical methods in solving differential equations. In this paper, we identify two sources of this issue in the case of Cauchy problems: the use of $L^2$ residuals as objective functions and the approximation gap of neural networks. We show that minimizing the sum of $L^2$ residual and initial condition error is not sufficient to guarantee the true solution, as this loss function does not capture the underlying dynamics. Additionally, neural networks are not capable of capturing singularities in the solutions due to the non-compactness of their image sets. This, in turn, influences the existence of global minima and the regularity of the network. We demonstrate that when the global minimum does not exist, machine precision becomes the predominant source of achievable error in practice. We also present numerical experiments in support of our theoretical claims.
Authors: Shi Yin, Xinyang Pan, Fengyan Wang, Feng Wu, Lixin He
Abstract: We present both a theoretical and a methodological framework that addresses a critical challenge in applying deep learning to physical systems: the reconciliation of non-linear expressiveness with SO(3)-equivariance in predictions of SO(3)-equivariant quantities. Inspired by covariant theory in physics, we address this problem by exploring the mathematical relationships between SO(3)-invariant and SO(3)-equivariant quantities and their representations. We first construct theoretical SO(3)-invariant quantities derived from the SO(3)-equivariant regression targets, and use these invariant quantities as supervisory labels to guide the learning of high-quality SO(3)-invariant features. Given that SO(3)-invariance is preserved under non-linear operations, the encoding process for invariant features can extensively utilize non-linear mappings, thereby fully capturing the non-linear patterns inherent in physical systems. Building on this foundation, we propose a gradient-based mechanism to induce SO(3)-equivariant encodings of various degrees from the learned SO(3)-invariant features. This mechanism can incorporate non-linear expressive capabilities into SO(3)-equivariant representations, while theoretically preserving their equivariant properties as we prove. We apply our theory and method to the electronic-structure Hamiltonian prediction tasks, experimental results on eight benchmark databases covering multiple types of elements and challenging scenarios show dramatic breakthroughs on the state-of-the-art prediction accuracy, with improvements of up to 40% in predicting Hamiltonians and up to 76% in predicting downstream physical quantities such as occupied orbital energy. Our approach goes beyond handling physical systems and offers a promising general solution to the critical dilemma between equivariance and non-linear expressiveness for the deep learning paradigm.
Authors: Zhenqiao Song, Yunlong Zhao, Wenxian Shi, Wengong Jin, Yang Yang, Lei Li
Abstract: Enzymes are genetically encoded biocatalysts capable of accelerating chemical reactions. How can we automatically design functional enzymes? In this paper, we propose EnzyGen, an approach to learn a unified model to design enzymes across all functional families. Our key idea is to generate an enzyme's amino acid sequence and their three-dimensional (3D) coordinates based on functionally important sites and substrates corresponding to a desired catalytic function. These sites are automatically mined from enzyme databases. EnzyGen consists of a novel interleaving network of attention and neighborhood equivariant layers, which captures both long-range correlation in an entire protein sequence and local influence from nearest amino acids in 3D space. To learn the generative model, we devise a joint training objective, including a sequence generation loss, a position prediction loss and an enzyme-substrate interaction loss. We further construct EnzyBench, a dataset with 3157 enzyme families, covering all available enzymes within the protein data bank (PDB). Experimental results show that our EnzyGen consistently achieves the best performance across all 323 testing families, surpassing the best baseline by 10.79% in terms of substrate binding affinity. These findings demonstrate EnzyGen's superior capability in designing well-folded and effective enzymes binding to specific substrates with high affinities.
Authors: Michael Zhou, Ramin Ramezani
Abstract: This paper explores various socioeconomic factors that contribute to individual financial success using machine learning algorithms and approaches. Financial success, a critical aspect of all individual's well-being, is a complex concept influenced by a plethora of different factors. This study aims to understand the true determinants of financial success. It examines the survey data from the National Longitudinal Survey of Youth 1997 by the Bureau of Labor Statistics [1], consisting of a sample of 8,984 individuals's longitudinal data over years. The dataset comprises income variables and a large set of socioeconomic variables of individuals. An in-depth analysis demonstrates the effectiveness of machine learning algorithms in financial success research, highlights the potential of leveraging longitudinal data to enhance prediction accuracy, and provides valuable insights into how various socioeconomic factors influence financial success. The findings underscore the significant influence of highest education degree, occupation and gender as the top three determinants of individual income among socioeconomic factors examined. Yearly working hours, age and work tenure emerge as three secondary influencing factors, and all other factors including parental household income, industry, parents' highest grade and others are identified as tertiary factors. These insights allow researchers to better understand the complex nature of financial success and enable policymakers to grasp the underlying dynamics shaping aspirations, decision-making, and the broader socio-economic fabric of society. This comprehension is crucial for fostering financial success among individuals and advancing broader societal well-being.
Authors: Tolulope Fadina, Thorsten Schmidt
Abstract: Fairness in decision-making processes is often quantified using probabilistic metrics. However, these metrics may not fully capture the real-world consequences of unfairness. In this article, we adopt a utility-based approach to more accurately measure the real-world impacts of decision-making process. In particular, we show that if the concept of $\varepsilon$-fairness is employed, it can possibly lead to outcomes that are maximally unfair in the real-world context. Additionally, we address the common issue of unavailable data on false negatives by proposing a reduced setting that still captures essential fairness considerations. We illustrate our findings with two real-world examples: college admissions and credit risk assessment. Our analysis reveals that while traditional probability-based evaluations might suggest fairness, a utility-based approach uncovers the necessary actions to truly achieve equality. For instance, in the college admission case, we find that enhancing completion rates is crucial for ensuring fairness. Summarizing, this paper highlights the importance of considering the real-world context when evaluating fairness.
Authors: Daniel T. Chang
Abstract: The conventional definition of hypergraph has two major issues: (1) there is not a standard definition of directed hypergraph and (2) there is not a formal definition of nested hypergraph. To resolve these issues, we propose a new definition of hypergraph that unifies the concepts of undirected, directed and nested hypergraphs, and that is uniform in using hyperedge as a single construct for representing high-order correlations among things, i.e., nodes and hyperedges. Specifically, we define a hyperedge to be a simple hyperedge, a nesting hyperedge, or a directed hyperedge. With this new definition, a hypergraph is nested if it has nesting hyperedge(s), and is directed if it has directed hyperedge(s). Otherwise, a hypergraph is a simple hypergraph. The uniformity and power of this new definition, with visualization, should facilitate the use of hypergraph for representing (hierarchical) high-order correlations in general and chemical systems in particular. Graph has been widely used as a mathematical structure for machine learning on molecular structures and 3D molecular geometries. However, graph has a major limitation: it can represent only pairwise correlations between nodes. Hypergraph extends graph with high-order correlations among nodes. This extension is significant or essential for machine learning on chemical systems. For molecules, this is significant as it allows the direct, explicit representation of multicenter bonds and molecular substructures. For chemical reactions, this is essential since most chemical reactions involve multiple participants. We propose the use of chemical hypergraph, a multilevel hypergraph with simple, nesting and directed hyperedges, as a single mathematical structure for representing chemical systems. We apply the new definition of hypergraph to chemical hypergraph and, as simplified versions, molecular hypergraph and chemical reaction hypergraph.
Authors: B. N. Kausik
Abstract: Deep learning neural network models must be large enough to adapt to their problem domain, while small enough to avoid overfitting training data during gradient descent. To balance these competing demands, overprovisioned deep learning models such as transformers are trained for a single epoch on large data sets, and hence inefficient with both computing resources and training data. In response to these inefficiencies, we exploit learning theory to derive Occam Gradient Descent, an algorithm that interleaves adaptive reduction of model size to minimize generalization error, with gradient descent on model weights to minimize fitting error. In contrast, traditional gradient descent greedily minimizes fitting error without regard to generalization error. Our algorithm simultaneously descends the space of weights and topological size of any neural network without modification, and is effective in our experiments in outperforming traditional gradient descent with or without post-train pruning in accuracy, compute and model compression.
Authors: Arun D. Kulkarni
Abstract: Recently, convolution neural networks (CNNs) have attracted a great deal of attention due to their remarkable performance in various domains, particularly in image and text classification tasks. However, their application to tabular data classification remains underexplored. There are many fields such as bioinformatics, finance, medicine where nonimage data are prevalent. Adaption of CNNs to classify nonimage data remains highly challenging. This paper investigates the efficacy of CNNs for tabular data classification, aiming to bridge the gap between traditional machine learning approaches and deep learning techniques. We propose a novel framework fuzzy convolution neural network (FCNN) tailored specifically for tabular data to capture local patterns within feature vectors. In our approach, we map feature values to fuzzy memberships. The fuzzy membership vectors are converted into images that are used to train the CNN model. The trained CNN model is used to classify unknown feature vectors. To validate our approach, we generated six complex noisy data sets. We used randomly selected seventy percent samples from each data set for training and thirty percent for testing. The data sets were also classified using the state-of-the-art machine learning algorithms such as the decision tree (DT), support vector machine (SVM), fuzzy neural network (FNN), Bayes classifier, and Random Forest (RF). Experimental results demonstrate that our proposed model can effectively learn meaningful representations from tabular data, achieving competitive or superior performance compared to existing methods. Overall, our finding suggests that the proposed FCNN model holds promise as a viable alternative for tabular data classification tasks, offering a fresh prospective and potentially unlocking new opportunities for leveraging deep learning in structured data analysis.
Authors: Wei Shao, Yufan Kang, Ziyan Peng, Xiao Xiao, Lei Wang, Yuhui Yang, Flora D Salim
Abstract: Accuracy and timeliness are indeed often conflicting goals in prediction tasks. Premature predictions may yield a higher rate of false alarms, whereas delaying predictions to gather more information can render them too late to be useful. In applications such as wildfires, crimes, and traffic jams, timely forecasting are vital for safeguarding human life and property. Consequently, finding a balance between accuracy and timeliness is crucial. In this paper, we propose an early spatio-temporal forecasting model based on Multi-Objective reinforcement learning that can either implement an optimal policy given a preference or infer the preference based on a small number of samples. The model addresses two primary challenges: 1) enhancing the accuracy of early forecasting and 2) providing the optimal policy for determining the most suitable prediction time for each area. Our method demonstrates superior performance on three large-scale real-world datasets, surpassing existing methods in early spatio-temporal forecasting tasks.
Authors: Binchuan Qi, Li Li, Wei Gong
Abstract: There remain numerous unanswered research questions on deep learning (DL) within the classical learning theory framework. These include the remarkable generalization capabilities of overparametrized neural networks (NNs), the efficient optimization performance despite non-convexity of objectives, the mechanism of flat minima for generalization, and the exceptional performance of deep architectures in solving physical problems. This paper introduces General Distribution Learning (GD Learning), a novel theoretical learning framework designed to address a comprehensive range of machine learning and statistical tasks, including classification, regression and parameter estimation. Departing from traditional statistical machine learning, GD Learning focuses on the true underlying distribution. In GD Learning, learning error, corresponding to the expected error in classical statistical learning framework, is divided into fitting errors due to models and algorithms, as well as sampling errors introduced by limited sampling data. The framework significantly incorporates prior knowledge, especially in scenarios characterized by data scarcity, thereby enhancing performance. Within the GD Learning framework, we demonstrate that the global optimal solutions in non-convex optimization can be approached by minimizing the gradient norm and the non-uniformity of the eigenvalues of the model's Jacobian matrix. This insight leads to the development of the gradient structure control algorithm. GD Learning also offers fresh insights into the questions on deep learning, including overparameterization and non-convex optimization, bias-variance trade-off, and the mechanism of flat minima.
Authors: Sean Lovett, Miha Zgubic, Sofia Liguori, Sephora Madjiheurem, Hamish Tomlinson, Sophie Elster, Chris Apps, Sims Witherspoon, Luis Piloto
Abstract: Solving the AC optimal power flow problem (AC-OPF) is critical to the efficient and safe planning and operation of power grids. Small efficiency improvements in this domain have the potential to lead to billions of dollars of cost savings, and significant reductions in emissions from fossil fuel generators. Recent work on data-driven solution methods for AC-OPF shows the potential for large speed improvements compared to traditional solvers; however, no large-scale open datasets for this problem exist. We present the largest readily-available collection of solved AC-OPF problems to date. This collection is orders of magnitude larger than existing readily-available datasets, allowing training of high-capacity data-driven models. Uniquely, it includes topological perturbations - a critical requirement for usage in realistic power grid operations. We hope this resource will spur the community to scale research to larger grid sizes with variable topology.
Authors: Yuxuan Shu, Vasileios Lampos
Abstract: In multivariate time series (MTS) forecasting, existing state-of-the-art deep learning approaches tend to focus on autoregressive formulations and overlook the information within exogenous indicators. To address this limitation, we present DeformTime, a neural network architecture that attempts to capture correlated temporal patterns from the input space, and hence, improve forecasting accuracy. It deploys two core operations performed by deformable attention blocks (DABs): learning dependencies across variables from different time steps (variable DAB), and preserving temporal dependencies in data from previous time steps (temporal DAB). Input data transformation is explicitly designed to enhance learning from the deformed series of information while passing through a DAB. We conduct extensive experiments on 6 MTS data sets, using previously established benchmarks as well as challenging infectious disease modelling tasks with more exogenous variables. The results demonstrate that DeformTime improves accuracy against previous competitive methods across the vast majority of MTS forecasting tasks, reducing the mean absolute error by 10% on average. Notably, performance gains remain consistent across longer forecasting horizons.
Authors: Lu Li, Tianyu Zhang, Zhiqi Bu, Suyuchen Wang, Huan He, Jie Fu, Yonghui Wu, Jiang Bian, Yong Chen, Yoshua Bengio
Abstract: Model merging has emerged as an effective approach to combine multiple single-task models, fine-tuned from the same pre-trained model, into a multitask model. This process typically involves computing a weighted average of the model parameters without any additional training. Existing model-merging methods focus on enhancing average task accuracy. However, interference and conflicts between the objectives of different tasks can lead to trade-offs during model merging. In real-world applications, a set of solutions with various trade-offs can be more informative, helping practitioners make decisions based on diverse preferences. In this paper, we introduce a novel low-compute algorithm, Model Merging with Amortized Pareto Front (MAP). MAP identifies a Pareto set of scaling coefficients for merging multiple models to reflect the trade-offs. The core component of MAP is approximating the evaluation metrics of the various tasks using a quadratic approximation surrogate model derived from a pre-selected set of scaling coefficients, enabling amortized inference. Experimental results on vision and natural language processing tasks show that MAP can accurately identify the Pareto front. To further reduce the required computation of MAP, we propose (1) a Bayesian adaptive sampling algorithm and (2) a nested merging scheme with multiple stages.
Authors: Zhining Liu, Ruizhong Qiu, Zhichen Zeng, Yada Zhu, Hendrik Hamann, Hanghang Tong
Abstract: Data collected in the real world often encapsulates historical discrimination against disadvantaged groups and individuals. Existing fair machine learning (FairML) research has predominantly focused on mitigating discriminative bias in the model prediction, with far less effort dedicated towards exploring how to trace biases present in the data, despite its importance for the transparency and interpretability of FairML. To fill this gap, we investigate a novel research problem: discovering samples that reflect biases/prejudices from the training data. Grounding on the existing fairness notions, we lay out a sample bias criterion and propose practical algorithms for measuring and countering sample bias. The derived bias score provides intuitive sample-level attribution and explanation of historical bias in data. On this basis, we further design two FairML strategies via sample-bias-informed minimal data editing. They can mitigate both group and individual unfairness at the cost of minimal or zero predictive utility loss. Extensive experiments and analyses on multiple real-world datasets demonstrate the effectiveness of our methods in explaining and mitigating unfairness. Code is available at https://github.com/ZhiningLiu1998/AIM.
Authors: Arian Prabowo, Xiachong Lin, Imran Razzak, Hao Xue, Emily W. Yap, Matthew Amos, Flora D. Salim
Abstract: Buildings play a crucial role in human well-being, influencing occupant comfort, health, and safety. Additionally, they contribute significantly to global energy consumption, accounting for one-third of total energy usage, and carbon emissions. Optimizing building performance presents a vital opportunity to combat climate change and promote human flourishing. However, research in building analytics has been hampered by the lack of accessible, available, and comprehensive real-world datasets on multiple building operations. In this paper, we introduce the Building TimeSeries (BTS) dataset. Our dataset covers three buildings over a three-year period, comprising more than ten thousand timeseries data points with hundreds of unique ontologies. Moreover, the metadata is standardized using the Brick schema. To demonstrate the utility of this dataset, we performed benchmarks on two tasks: timeseries ontology classification and zero-shot forecasting. These tasks represent an essential initial step in addressing challenges related to interoperability in building analytics. Access to the dataset and the code used for benchmarking are available here: https://github.com/cruiseresearchgroup/DIEF_BTS .
Authors: Steven H. Berguin
Abstract: Jacobian-Enhanced Neural Networks (JENN) are densely connected multi-layer perceptrons, whose training process is modified to predict partial derivatives accurately. Their main benefit is better accuracy with fewer training points compared to standard neural networks. These attributes are particularly desirable in the field of computer-aided design, where there is often the need to replace computationally expensive, physics-based models with fast running approximations, known as surrogate models or meta-models. Since a surrogate emulates the original model accurately in near-real time, it yields a speed benefit that can be used to carry out orders of magnitude more function calls quickly. However, in the special case of gradient-enhanced methods, there is the additional value proposition that partial derivatives are accurate, which is a critical property for one important use-case: surrogate-based optimization. This work derives the complete theory and exemplifies its superiority over standard neural nets for surrogate-based optimization.
Authors: M\'arton Kardos, Jan Kostkan, Arnault-Quentin Vermillet, Kristoffer Nielbo, Kenneth Enevoldsen, Roberta Rocca
Abstract: Topic models are useful tools for discovering latent semantic structures in large textual corpora. Topic modeling historically relied on bag-of-words representations of language. This approach makes models sensitive to the presence of stop words and noise, and does not utilize potentially useful contextual information. Recent efforts have been oriented at incorporating contextual neural representations in topic modeling and have been shown to outperform classical topic models. These approaches are, however, typically slow, volatile and still require preprocessing for optimal results. We present Semantic Signal Separation ($S^3$), a theory-driven topic modeling approach in neural embedding spaces. $S^3$ conceptualizes topics as independent axes of semantic space, and uncovers these with blind-source separation. Our approach provides the most diverse, highly coherent topics, requires no preprocessing, and is demonstrated to be the fastest contextually sensitive topic model to date. We offer an implementation of $S^3$, among other approaches, in the Turftopic Python package.
Authors: Quan M. Tran, Suong N. Hoang, Lam M. Nguyen, Dzung Phan, Hoang Thanh Lam
Abstract: Foundational models (FMs), pretrained on extensive datasets using self-supervised techniques, are capable of learning generalized patterns from large amounts of data. This reduces the need for extensive labeled datasets for each new task, saving both time and resources by leveraging the broad knowledge base established during pretraining. Most research on FMs has primarily focused on unstructured data, such as text and images, or semi-structured data, like time-series. However, there has been limited attention to structured data, such as tabular data, which, despite its prevalence, remains under-studied due to a lack of clean datasets and insufficient research on the transferability of FMs for various tabular data tasks. In response to this gap, we introduce a framework called TabularFM, which incorporates state-of-the-art methods for developing FMs specifically for tabular data. This includes variations of neural architectures such as GANs, VAEs, and Transformers. We have curated a million of tabular datasets and released cleaned versions to facilitate the development of tabular FMs. We pretrained FMs on this curated data, benchmarked various learning methods on these datasets, and released the pretrained models along with leaderboards for future comparative studies. Our fully open-sourced system provides a comprehensive analysis of the transferability of tabular FMs. By releasing these datasets, pretrained models, and leaderboards, we aim to enhance the validity and usability of tabular FMs in the near future.
Authors: Ifigeneia Apostolopoulou, Benjamin Eysenbach, Frank Nielsen, Artur Dubrawski
Abstract: In supervised learning, understanding an input's proximity to the training data can help a model decide whether it has sufficient evidence for reaching a reliable prediction. While powerful probabilistic models such as Gaussian Processes naturally have this property, deep neural networks often lack it. In this paper, we introduce Distance Aware Bottleneck (DAB), i.e., a new method for enriching deep neural networks with this property. Building on prior information bottleneck approaches, our method learns a codebook that stores a compressed representation of all inputs seen during training. The distance of a new example from this codebook can serve as an uncertainty estimate for the example. The resulting model is simple to train and provides deterministic uncertainty estimates by a single forward pass. Finally, our method achieves better out-of-distribution (OOD) detection and misclassification prediction than prior methods, including expensive ensemble methods, deep kernel Gaussian Processes, and approaches based on the standard information bottleneck.
Authors: Bhrij Patel, Vishnu Sashank Dorbala, Dinesh Manocha, Amrit Singh Bedi
Abstract: Embodied Question Answering (EQA) is an important problem, which involves an agent exploring the environment to answer user queries. In the existing literature, EQA has exclusively been studied in single-agent scenarios, where exploration can be time-consuming and costly. In this work, we consider EQA in a multi-agent framework involving multiple large language models (LLM) based agents independently answering queries about a household environment. To generate one answer for each query, we use the individual responses to train a Central Answer Model (CAM) that aggregates responses for a robust answer. Using CAM, we observe a $50\%$ higher EQA accuracy when compared against aggregation methods for ensemble LLM, such as voting schemes and debates. CAM does not require any form of agent communication, alleviating it from the associated costs. We ablate CAM with various nonlinear (neural network, random forest, decision tree, XGBoost) and linear (logistic regression classifier, SVM) algorithms. Finally, we present a feature importance analysis for CAM via permutation feature importance (PFI), quantifying CAMs reliance on each independent agent and query context.
Authors: Shirley Wu, Shiyu Zhao, Qian Huang, Kexin Huang, Michihiro Yasunaga, Kaidi Cao, Vassilis N. Ioannidis, Karthik Subbian, Jure Leskovec, James Zou
Abstract: Large language model (LLM) agents have demonstrated impressive capability in utilizing external tools and knowledge to boost accuracy and reduce hallucinations. However, developing the prompting techniques that make LLM agents able to effectively use external tools and knowledge is a heuristic and laborious task. Here, we introduce AvaTaR, a novel and automatic framework that optimizes an LLM agent to effectively use the provided tools and improve its performance on a given task/domain. During optimization, we design a comparator module to iteratively provide insightful and holistic prompts to the LLM agent via reasoning between positive and negative examples sampled from training data. We demonstrate AvaTaR on four complex multimodal retrieval datasets featuring textual, visual, and relational information. We find AvaTaR consistently outperforms state-of-the-art approaches across all four challenging tasks and exhibits strong generalization ability when applied to novel cases, achieving an average relative improvement of 14% on the Hit@1 metric. Code and dataset are available at https://github.com/zou-group/avatar.
Authors: Jiangmeng Li, Bin Qin, Qirui Ji, Yi Li, Wenwen Qiang, Jianwen Cao, Fanjiang Xu
Abstract: Leveraging the development of structural causal model (SCM), researchers can establish graphical models for exploring the causal mechanisms behind machine learning techniques. As the complexity of machine learning applications rises, single-world interventionism causal analysis encounters theoretical adaptation limitations. Accordingly, cross-world counterfactual approach extends our understanding of causality beyond observed data, enabling hypothetical reasoning about alternative scenarios. However, the joint involvement of cross-world variables, encompassing counterfactual variables and real-world variables, challenges the construction of the graphical model. Twin network is a subtle attempt, establishing a symbiotic relationship, to bridge the gap between graphical modeling and the introduction of counterfactuals albeit with room for improvement in generalization. In this regard, we demonstrate the theoretical breakdowns of twin networks in certain cross-world counterfactual scenarios. To this end, we propose a novel teleporter theory to establish a general and simple graphical representation of counterfactuals, which provides criteria for determining teleporter variables to connect multiple worlds. In theoretical application, we determine that introducing the proposed teleporter theory can directly obtain the conditional independence between counterfactual variables and real-world variables from the cross-world SCM without requiring complex algebraic derivations. Accordingly, we can further identify counterfactual causal effects through cross-world symbolic derivation. We demonstrate the generality of the teleporter theory to the practical application. Adhering to the proposed theory, we build a plug-and-play module, and the effectiveness of which are substantiated by experiments on benchmarks.
Authors: Noah Golowich, Ankur Moitra
Abstract: One of the most natural approaches to reinforcement learning (RL) with function approximation is value iteration, which inductively generates approximations to the optimal value function by solving a sequence of regression problems. To ensure the success of value iteration, it is typically assumed that Bellman completeness holds, which ensures that these regression problems are well-specified. We study the problem of learning an optimal policy under Bellman completeness in the online model of RL with linear function approximation. In the linear setting, while statistically efficient algorithms are known under Bellman completeness (e.g., Jiang et al. (2017); Zanette et al. (2020)), these algorithms all rely on the principle of global optimism which requires solving a nonconvex optimization problem. In particular, it has remained open as to whether computationally efficient algorithms exist. In this paper we give the first polynomial-time algorithm for RL under linear Bellman completeness when the number of actions is any constant.
Authors: Constantinos Daskalakis, Noah Golowich
Abstract: The empirical risk minimization (ERM) principle has been highly impactful in machine learning, leading both to near-optimal theoretical guarantees for ERM-based learning algorithms as well as driving many of the recent empirical successes in deep learning. In this paper, we investigate the question of whether the ability to perform ERM, which computes a hypothesis minimizing empirical risk on a given dataset, is necessary for efficient learning: in particular, is there a weaker oracle than ERM which can nevertheless enable learnability? We answer this question affirmatively, showing that in the realizable setting of PAC learning for binary classification, a concept class can be learned using an oracle which only returns a single bit indicating whether a given dataset is realizable by some concept in the class. The sample complexity and oracle complexity of our algorithm depend polynomially on the VC dimension of the hypothesis class, thus showing that there is only a polynomial price to pay for use of our weaker oracle. Our results extend to the agnostic learning setting with a slight strengthening of the oracle, as well as to the partial concept, multiclass and real-valued learning settings. In the setting of partial concept classes, prior to our work no oracle-efficient algorithms were known, even with a standard ERM oracle. Thus, our results address a question of Alon et al. (2021) who asked whether there are algorithmic principles which enable efficient learnability in this setting.
Authors: Yibin Wang, Haizhou Shi, Ligong Han, Dimitris Metaxas, Hao Wang
Abstract: Large Language Models (LLMs) often suffer from overconfidence during inference, particularly when adapted to downstream domain-specific tasks with limited data. Previous work addresses this issue by employing approximate Bayesian estimation after the LLMs are trained, enabling them to quantify uncertainty. However, such post-training approaches' performance is severely limited by the parameters learned during training. In this paper, we go beyond post-training Bayesianization and propose Bayesian Low-Rank Adaptation by Backpropagation (BLoB), an algorithm that continuously and jointly adjusts both the mean and covariance of LLM parameters throughout the whole fine-tuning process. Our empirical results verify the effectiveness of BLoB in terms of generalization and uncertainty estimation, when evaluated on both in-distribution and out-of-distribution data.
Authors: Xueqi Cheng, Yu Wang, Yunchao Liu, Yuying Zhao, Charu C. Aggarwal, Tyler Derr
Abstract: Recent years have witnessed the remarkable success of applying Graph machine learning (GML) to node/graph classification and link prediction. However, edge classification task that enjoys numerous real-world applications such as social network analysis and cybersecurity, has not seen significant advancement. To address this gap, our study pioneers a comprehensive approach to edge classification. We identify a novel `Topological Imbalance Issue', which arises from the skewed distribution of edges across different classes, affecting the local subgraph of each edge and harming the performance of edge classifications. Inspired by the recent studies in node classification that the performance discrepancy exists with varying local structural patterns, we aim to investigate if the performance discrepancy in topological imbalanced edge classification can also be mitigated by characterizing the local class distribution variance. To overcome this challenge, we introduce Topological Entropy (TE), a novel topological-based metric that measures the topological imbalance for each edge. Our empirical studies confirm that TE effectively measures local class distribution variance, and indicate that prioritizing edges with high TE values can help address the issue of topological imbalance. Based on this, we develop two strategies - Topological Reweighting and TE Wedge-based Mixup - to focus training on (synthetic) edges based on their TEs. While topological reweighting directly manipulates training edge weights according to TE, our wedge-based mixup interpolates synthetic edges between high TE wedges. Ultimately, we integrate these strategies into a novel topological imbalance strategy for edge classification: TopoEdge. Through extensive experiments, we demonstrate the efficacy of our proposed strategies on newly curated datasets and thus establish a new benchmark for (imbalanced) edge classification.
Authors: Noah Golowich, Ankur Moitra
Abstract: In this paper, we study the offline RL problem with linear function approximation. Our main structural assumption is that the MDP has low inherent Bellman error, which stipulates that linear value functions have linear Bellman backups with respect to the greedy policy. This assumption is natural in that it is essentially the minimal assumption required for value iteration to succeed. We give a computationally efficient algorithm which succeeds under a single-policy coverage condition on the dataset, namely which outputs a policy whose value is at least that of any policy which is well-covered by the dataset. Even in the setting when the inherent Bellman error is 0 (termed linear Bellman completeness), our algorithm yields the first known guarantee under single-policy coverage. In the setting of positive inherent Bellman error ${\varepsilon_{\mathrm{BE}}} > 0$, we show that the suboptimality error of our algorithm scales with $\sqrt{\varepsilon_{\mathrm{BE}}}$. Furthermore, we prove that the scaling of the suboptimality with $\sqrt{\varepsilon_{\mathrm{BE}}}$ cannot be improved for any algorithm. Our lower bound stands in contrast to many other settings in reinforcement learning with misspecification, where one can typically obtain performance that degrades linearly with the misspecification error.
Authors: Jason Gross, Rajashree Agrawal, Thomas Kwa, Euan Ong, Chun Hei Yip, Alex Gibson, Soufiane Noubir, Lawrence Chan
Abstract: In this work, we propose using mechanistic interpretability -- techniques for reverse engineering model weights into human-interpretable algorithms -- to derive and compactly prove formal guarantees on model performance. We prototype this approach by formally proving lower bounds on the accuracy of 151 small transformers trained on a Max-of-$K$ task. We create 102 different computer-assisted proof strategies and assess their length and tightness of bound on each of our models. Using quantitative metrics, we find that shorter proofs seem to require and provide more mechanistic understanding. Moreover, we find that more faithful mechanistic understanding leads to tighter performance bounds. We confirm these connections by qualitatively examining a subset of our proofs. Finally, we identify compounding structureless noise as a key challenge for using mechanistic interpretability to generate compact proofs on model performance.
Authors: Jeffrey Li, Alex Fang, Georgios Smyrnis, Maor Ivgi, Matt Jordan, Samir Gadre, Hritik Bansal, Etash Guha, Sedrick Keh, Kushal Arora, Saurabh Garg, Rui Xin, Niklas Muennighoff, Reinhard Heckel, Jean Mercat, Mayee Chen, Suchin Gururangan, Mitchell Wortsman, Alon Albalak, Yonatan Bitton, Marianna Nezhurina, Amro Abbas, Cheng-Yu Hsieh, Dhruba Ghosh, Josh Gardner, Maciej Kilian, Hanlin Zhang, Rulin Shao, Sarah Pratt, Sunny Sanyal, Gabriel Ilharco, Giannis Daras, Kalyani Marathe, Aaron Gokaslan, Jieyu Zhang, Khyathi Chandu, Thao Nguyen, Igor Vasiljevic, Sham Kakade, Shuran Song, Sujay Sanghavi, Fartash Faghri, Sewoong Oh, Luke Zettlemoyer, Kyle Lo, Alaaeldin El-Nouby, Hadi Pouransari, Alexander Toshev, Stephanie Wang, Dirk Groeneveld, Luca Soldaini, Pang Wei Koh, Jenia Jitsev, Thomas Kollar, Alexandros G. Dimakis, Yair Carmon, Achal Dave, Ludwig Schmidt, Vaishaal Shankar
Abstract: We introduce DataComp for Language Models (DCLM), a testbed for controlled dataset experiments with the goal of improving language models. As part of DCLM, we provide a standardized corpus of 240T tokens extracted from Common Crawl, effective pretraining recipes based on the OpenLM framework, and a broad suite of 53 downstream evaluations. Participants in the DCLM benchmark can experiment with data curation strategies such as deduplication, filtering, and data mixing at model scales ranging from 412M to 7B parameters. As a baseline for DCLM, we conduct extensive experiments and find that model-based filtering is key to assembling a high-quality training set. The resulting dataset, DCLM-Baseline enables training a 7B parameter language model from scratch to 64% 5-shot accuracy on MMLU with 2.6T training tokens. Compared to MAP-Neo, the previous state-of-the-art in open-data language models, DCLM-Baseline represents a 6.6 percentage point improvement on MMLU while being trained with 40% less compute. Our baseline model is also comparable to Mistral-7B-v0.3 and Llama 3 8B on MMLU (63% & 66%), and performs similarly on an average of 53 natural language understanding tasks while being trained with 6.6x less compute than Llama 3 8B. Our results highlight the importance of dataset design for training language models and offer a starting point for further research on data curation.
Authors: Chris Junchi Li
Abstract: We study the problem of solving strongly convex and smooth unconstrained optimization problems using stochastic first-order algorithms. We devise a novel algorithm, referred to as \emph{Recursive One-Over-T SGD} (\textsf{ROOT-SGD}), based on an easily implementable, recursive averaging of past stochastic gradients. We prove that it simultaneously achieves state-of-the-art performance in both a finite-sample, nonasymptotic sense and an asymptotic sense. On the nonasymptotic side, we prove risk bounds on the last iterate of \textsf{ROOT-SGD} with leading-order terms that match the optimal statistical risk with a unity pre-factor, along with a higher-order term that scales at the sharp rate of $O(n^{-3/2})$ under the Lipschitz condition on the Hessian matrix. On the asymptotic side, we show that when a mild, one-point Hessian continuity condition is imposed, the rescaled last iterate of (multi-epoch) \textsf{ROOT-SGD} converges asymptotically to a Gaussian limit with the Cram\'{e}r-Rao optimal asymptotic covariance, for a broad range of step-size choices.
Authors: Luke A. Bauer, James K. Howes IV, Sam A. Markelon, Vincent Bindschaedler, Thomas Shrimpton
Abstract: State of the art generative models of human-produced content are the focus of many recent papers that explore their use for steganographic communication. In particular, generative models of natural language text. Loosely, these works (invertibly) encode message-carrying bits into a sequence of samples from the model, ultimately yielding a plausible natural language covertext. By focusing on this narrow steganographic piece, prior work has largely ignored the significant algorithmic challenges, and performance-security tradeoffs, that arise when one actually tries to build a messaging pipeline around it. We make these challenges concrete, by considering the natural application of such a pipeline: namely, "dead-drop" covert messaging over large, public internet platforms (e.g. social media sites). We explicate the challenges and describe approaches to overcome them, surfacing in the process important performance and security tradeoffs that must be carefully tuned. We implement a system around this model-based format-transforming encryption pipeline, and give an empirical analysis of its performance and (heuristic) security.
Authors: Xiaowu Dai, Wenlu Xu, Yuan Qi, Michael I. Jordan
Abstract: Online platforms in the Internet Economy commonly incorporate recommender systems that recommend products (or "arms") to users (or "agents"). A key challenge in this domain arises from myopic agents who are naturally incentivized to exploit by choosing the optimal arm based on current information, rather than exploring various alternatives to gather information that benefits the collective. We propose a novel recommender system that aligns with agents' incentives while achieving asymptotically optimal performance, as measured by regret in repeated interactions. Our framework models this incentive-aware system as a multi-agent bandit problem in two-sided markets, where the interactions of agents and arms are facilitated by recommender systems on online platforms. This model incorporates incentive constraints induced by agents' opportunity costs. In scenarios where opportunity costs are known to the platform, we show the existence of an incentive-compatible recommendation algorithm. This algorithm pools recommendations between a genuinely good arm and an unknown arm using a randomized and adaptive strategy. Moreover, when these opportunity costs are unknown, we introduce an algorithm that randomly pools recommendations across all arms, utilizing the cumulative loss from each arm as feedback for strategic exploration. We demonstrate that both algorithms satisfy an ex-post fairness criterion, which protects agents from over-exploitation. All code for using the proposed algorithms and reproducing results is made available on GitHub.
Authors: Nayana Dasgupta, Mirco Musolesi
Abstract: Solving the problem of cooperation is fundamentally important for the creation and maintenance of functional societies. Problems of cooperation are omnipresent within human society, with examples ranging from navigating busy road junctions to negotiating treaties. As the use of AI becomes more pervasive throughout society, the need for socially intelligent agents capable of navigating these complex cooperative dilemmas is becoming increasingly evident. Direct punishment is a ubiquitous social mechanism that has been shown to foster the emergence of cooperation in both humans and non-humans. In the natural world, direct punishment is often strongly coupled with partner selection and reputation and used in conjunction with third-party punishment. The interactions between these mechanisms could potentially enhance the emergence of cooperation within populations. However, no previous work has evaluated the learning dynamics and outcomes emerging from Multi-Agent Reinforcement Learning (MARL) populations that combine these mechanisms. This paper addresses this gap. It presents a comprehensive analysis and evaluation of the behaviors and learning dynamics associated with direct punishment, third-party punishment, partner selection, and reputation. Finally, we discuss the implications of using these mechanisms on the design of cooperative AI systems.
Authors: Shiran Yuan, Kaizhu Huang
Abstract: Low-Rank Tensor Completion, a method which exploits the inherent structure of tensors, has been studied extensively as an effective approach to tensor completion. Whilst such methods attained great success, none have systematically considered exploiting the numerical priors of tensor elements. Ignoring numerical priors causes loss of important information regarding the data, and therefore prevents the algorithms from reaching optimal accuracy. Despite the existence of some individual works which consider ad hoc numerical priors for specific tasks, no generalizable frameworks for incorporating numerical priors have appeared. We present the Generalized CP Decomposition Tensor Completion (GCDTC) framework, the first generalizable framework for low-rank tensor completion that takes numerical priors of the data into account. We test GCDTC by further proposing the Smooth Poisson Tensor Completion (SPTC) algorithm, an instantiation of the GCDTC framework, whose performance exceeds current state-of-the-arts by considerable margins in the task of non-negative tensor completion, exemplifying GCDTC's effectiveness. Our code is open-source.
Authors: Chengkun Li, Gr\'egoire Clart\'e, Martin J{\o}rgensen, Luigi Acerbi
Abstract: In applied Bayesian inference scenarios, users may have access to a large number of pre-existing model evaluations, for example from maximum-a-posteriori (MAP) optimization runs. However, traditional approximate inference techniques make little to no use of this available information. We propose the framework of post-process Bayesian inference as a means to obtain a quick posterior approximation from existing target density evaluations, with no further model calls. Within this framework, we introduce Variational Sparse Bayesian Quadrature (VSBQ), a method for post-process approximate inference for models with black-box and potentially noisy likelihoods. VSBQ reuses existing target density evaluations to build a sparse Gaussian process (GP) surrogate model of the log posterior density function. Subsequently, we leverage sparse-GP Bayesian quadrature combined with variational inference to achieve fast approximate posterior inference over the surrogate. We validate our method on challenging synthetic scenarios and real-world applications from computational neuroscience. The experiments show that VSBQ builds high-quality posterior approximations by post-processing existing optimization traces, with no further model evaluations.
Authors: Harry Li, Gabriel Appleby, Camelia Daniela Brumar, Remco Chang, Ashley Suh
Abstract: This study presents insights from interviews with nineteen Knowledge Graph (KG) practitioners who work in both enterprise and academic settings on a wide variety of use cases. Through this study, we identify critical challenges experienced by KG practitioners when creating, exploring, and analyzing KGs that could be alleviated through visualization design. Our findings reveal three major personas among KG practitioners - KG Builders, Analysts, and Consumers - each of whom have their own distinct expertise and needs. We discover that KG Builders would benefit from schema enforcers, while KG Analysts need customizable query builders that provide interim query results. For KG Consumers, we identify a lack of efficacy for node-link diagrams, and the need for tailored domain-specific visualizations to promote KG adoption and comprehension. Lastly, we find that implementing KGs effectively in practice requires both technical and social solutions that are not addressed with current tools, technologies, and collaborative workflows. From the analysis of our interviews, we distill several visualization research directions to improve KG usability, including knowledge cards that balance digestibility and discoverability, timeline views to track temporal changes, interfaces that support organic discovery, and semantic explanations for AI and machine learning predictions.
Authors: Francisco Vargas, Teodora Reu, Anna Kerekes
Abstract: Denoising diffusion models are a class of generative models which have recently achieved state-of-the-art results across many domains. Gradual noise is added to the data using a diffusion process, which transforms the data distribution into a Gaussian. Samples from the generative model are then obtained by simulating an approximation of the time reversal of this diffusion initialized by Gaussian samples. Recent research has explored adapting diffusion models for sampling and inference tasks. In this paper, we leverage known connections to stochastic control akin to the F\"ollmer drift to extend established neural network approximation results for the F\"ollmer drift to denoising diffusion models and samplers.
Authors: Tjeerd Jan Heeringa, Len Spek, Felix Schwenninger, Christoph Brune
Abstract: The approximation properties of infinitely wide shallow neural networks heavily depend on the choice of the activation function. To understand this influence, we study embeddings between Barron spaces with different activation functions. These embeddings are proven by providing push-forward maps on the measures $\mu$ used to represent functions $f$. An activation function of particular interest is the rectified power unit ($\operatorname{RePU}$) given by $\operatorname{RePU}_s(x)=\max(0,x)^s$. For many commonly used activation functions, the well-known Taylor remainder theorem can be used to construct a push-forward map, which allows us to prove the embedding of the associated Barron space into a Barron space with a $\operatorname{RePU}$ as activation function. Moreover, the Barron spaces associated with the $\operatorname{RePU}_s$ have a hierarchical structure similar to the Sobolev spaces $H^m$.
Authors: Young Wu, Jeremy McMahan, Xiaojin Zhu, Qiaomin Xie
Abstract: We characterize offline data poisoning attacks on Multi-Agent Reinforcement Learning (MARL), where an attacker may change a data set in an attempt to install a (potentially fictitious) unique Markov-perfect Nash equilibrium for a two-player zero-sum Markov game. We propose the unique Nash set, namely the set of games, specified by their Q functions, with a specific joint policy being the unique Nash equilibrium. The unique Nash set is central to poisoning attacks because the attack is successful if and only if data poisoning pushes all plausible games inside the set. The unique Nash set generalizes the reward polytope commonly used in inverse reinforcement learning to MARL. For zero-sum Markov games, both the inverse Nash set and the set of plausible games induced by data are polytopes in the Q function space. We exhibit a linear program to efficiently compute the optimal poisoning attack. Our work sheds light on the structure of data poisoning attacks on offline MARL, a necessary step before one can design more robust MARL algorithms.
Authors: Jia-Qi Wang, Rong-Qiang He, Zhong-Yi Lu
Abstract: Neural-network quantum states (NQSs), variationally optimized by combining traditional methods and deep learning techniques, is a new way to find quantum many-body ground states and gradually becomes a competitor of traditional variational methods. However, there are still some difficulties in the optimization of NQSs, such as local minima, slow convergence, and sign structure optimization. Here, we split a quantum many-body variational wave function into a multiplication of a real-valued amplitude neural network and a sign structure, and focus on the optimization of the amplitude network while keeping the sign structure fixed. The amplitude network is a convolutional neural network (CNN) with residual blocks, namely a ResNet. Our method is tested on three typical quantum many-body systems. The obtained ground state energies are lower than or comparable to those from traditional variational Monte Carlo (VMC) methods and density matrix renormalization group (DMRG). Surprisingly, for the frustrated Heisenberg $J_1$-$J_2$ model, our results are better than those of the complex-valued CNN in the literature, implying that the sign structure of the complex-valued NQS is difficult to be optimized. We will study the optimization of the sign structure of NQSs in the future.
Authors: Jingqian Wu, Rongtao Xu, Zach Wood-Doughty, Changwei Wang, Shibiao Xu, Edmund Y. Lam
Abstract: Local feature detection and description play an important role in many computer vision tasks, which are designed to detect and describe keypoints in "any scene" and "any downstream task". Data-driven local feature learning methods need to rely on pixel-level correspondence for training, which is challenging to acquire at scale, thus hindering further improvements in performance. In this paper, we propose SAMFeat to introduce SAM (segment anything model), a fundamental model trained on 11 million images, as a teacher to guide local feature learning and thus inspire higher performance on limited datasets. To do so, first, we construct an auxiliary task of Attention-weighted Semantic Relation Distillation (ASRD), which distillates feature relations with category-agnostic semantic information learned by the SAM encoder into a local feature learning network, to improve local feature description using semantic discrimination. Second, we develop a technique called Weakly Supervised Contrastive Learning Based on Semantic Grouping (WSC), which utilizes semantic groupings derived from SAM as weakly supervised signals, to optimize the metric space of local descriptors. Third, we design an Edge Attention Guidance (EAG) to further improve the accuracy of local feature detection and description by prompting the network to pay more attention to the edge region guided by SAM. SAMFeat's performance on various tasks such as image matching on HPatches, and long-term visual localization on Aachen Day-Night showcases its superiority over previous local features. The release code is available at https://github.com/vignywang/SAMFeat.
Authors: Hua Ma, Shang Wang, Yansong Gao, Zhi Zhang, Huming Qiu, Minhui Xue, Alsharif Abuadbba, Anmin Fu, Surya Nepal, Derek Abbott
Abstract: All current backdoor attacks on deep learning (DL) models fall under the category of a vertical class backdoor (VCB) -- class-dependent. In VCB attacks, any sample from a class activates the implanted backdoor when the secret trigger is present. Existing defense strategies overwhelmingly focus on countering VCB attacks, especially those that are source-class-agnostic. This narrow focus neglects the potential threat of other simpler yet general backdoor types, leading to false security implications. This study introduces a new, simple, and general type of backdoor attack coined as the horizontal class backdoor (HCB) that trivially breaches the class dependence characteristic of the VCB, bringing a fresh perspective to the community. HCB is now activated when the trigger is presented together with an innocuous feature, regardless of class. For example, the facial recognition model misclassifies a person who wears sunglasses with a smiling innocuous feature into the targeted person, such as an administrator, regardless of which person. The key is that these innocuous features are horizontally shared among classes but are only exhibited by partial samples per class. Extensive experiments on attacking performance across various tasks, including MNIST, facial recognition, traffic sign recognition, object detection, and medical diagnosis, confirm the high efficiency and effectiveness of the HCB. We rigorously evaluated the evasiveness of the HCB against a series of eleven representative countermeasures, including Fine-Pruning (RAID 18'), STRIP (ACSAC 19'), Neural Cleanse (Oakland 19'), ABS (CCS 19'), Februus (ACSAC 20'), NAD (ICLR 21'), MNTD (Oakland 21'), SCAn (USENIX SEC 21'), MOTH (Oakland 22'), Beatrix (NDSS 23'), and MM-BD (Oakland 24'). None of these countermeasures prove robustness, even when employing a simplistic trigger, such as a small and static white-square patch.
Authors: Shi Chen, Qin Li, Oliver Tse, Stephen J. Wright
Abstract: The acceleration of gradient-based optimization methods is a subject of significant practical and theoretical importance, particularly within machine learning applications. While much attention has been directed towards optimizing within Euclidean space, the need to optimize over spaces of probability measures in machine learning motivates exploration of accelerated gradient methods in this context too. To this end, we introduce a Hamiltonian-flow approach analogous to momentum-based approaches in Euclidean space. We demonstrate that, in the continuous-time setting, algorithms based on this approach can achieve convergence rates of arbitrarily high order. We complement our findings with numerical examples.
Authors: Teyun Kwon, Norman Di Palo, Edward Johns
Abstract: Large Language Models (LLMs) have recently shown promise as high-level planners for robots when given access to a selection of low-level skills. However, it is often assumed that LLMs do not possess sufficient knowledge to be used for the low-level trajectories themselves. In this work, we address this assumption thoroughly, and investigate if an LLM (GPT-4) can directly predict a dense sequence of end-effector poses for manipulation tasks, when given access to only object detection and segmentation vision models. We designed a single, task-agnostic prompt, without any in-context examples, motion primitives, or external trajectory optimisers. Then we studied how well it can perform across 30 real-world language-based tasks, such as "open the bottle cap" and "wipe the plate with the sponge", and we investigated which design choices in this prompt are the most important. Our conclusions raise the assumed limit of LLMs for robotics, and we reveal for the first time that LLMs do indeed possess an understanding of low-level robot control sufficient for a range of common tasks, and that they can additionally detect failures and then re-plan trajectories accordingly. Videos, prompts, and code are available at: https://www.robot-learning.uk/language-models-trajectory-generators.
URLs: https://www.robot-learning.uk/language-models-trajectory-generators.
Authors: Md Zobaer Islam, Brenden Martin, Carly Gotcher, Tyler Martinez, John F. O'Hara, Sabit Ekin
Abstract: In this study, we present a non-contact respiratory anomaly detection method using incoherent light-wave signals reflected from the chest of a mechanical robot that can breathe like human beings. In comparison to existing radar and camera-based sensing systems for vitals monitoring, this technology uses only a low-cost ubiquitous infrared light source and sensor. This light-wave sensing system recognizes different breathing anomalies from the variations of light intensity reflected from the chest of the robot within a 0.5m-1.5m range with an average classification accuracy of up to 96.6% using machine learning.
Authors: Ravil Mussabayev, Rustam Mussabayev
Abstract: This paper introduces a novel formulation of the clustering problem, namely the Minimum Sum-of-Squares Clustering of Infinitely Tall Data (MSSC-ITD), and presents HPClust, an innovative set of hybrid parallel approaches for its effective solution. By utilizing modern high-performance computing techniques, HPClust enhances key clustering metrics: effectiveness, computational efficiency, and scalability. In contrast to vanilla data parallelism, which only accelerates processing time through the MapReduce framework, our approach unlocks superior performance by leveraging the multi-strategy competitive-cooperative parallelism and intricate properties of the objective function landscape. Unlike other available algorithms that struggle to scale, our algorithm is inherently parallel in nature, improving solution quality through increased scalability and parallelism, and outperforming even advanced algorithms designed for small and medium-sized datasets. Our evaluation of HPClust, featuring four parallel strategies, demonstrates its superiority over traditional and cutting-edge methods by offering better performance in the key metrics. These results also show that parallel processing not only enhances the clustering efficiency, but the accuracy as well. Additionally, we explore the balance between computational efficiency and clustering quality, providing insights into optimal parallel strategies based on dataset specifics and resource availability. This research advances our understanding of parallelism in clustering algorithms, demonstrating that a judicious hybridization of advanced parallel approaches yields optimal results for MSSC-ITD. Experiments on synthetic data further confirm HPClust's exceptional scalability and robustness to noise.
Authors: Vignesh Gopakumar, Stanislas Pamela, Lorenzo Zanisi, Zongyi Li, Ander Gray, Daniel Brennand, Nitesh Bhatia, Gregory Stathopoulos, Matt Kusner, Marc Peter Deisenroth, Anima Anandkumar, JOREK Team, MAST Team
Abstract: Predicting plasma evolution within a Tokamak reactor is crucial to realizing the goal of sustainable fusion. Capabilities in forecasting the spatio-temporal evolution of plasma rapidly and accurately allow us to quickly iterate over design and control strategies on current Tokamak devices and future reactors. Modelling plasma evolution using numerical solvers is often expensive, consuming many hours on supercomputers, and hence, we need alternative inexpensive surrogate models. We demonstrate accurate predictions of plasma evolution both in simulation and experimental domains using deep learning-based surrogate modelling tools, viz., Fourier Neural Operators (FNO). We show that FNO has a speedup of six orders of magnitude over traditional solvers in predicting the plasma dynamics simulated from magnetohydrodynamic models, while maintaining a high accuracy (MSE in the normalised domain $\approx$ $10^{-5}$). Our modified version of the FNO is capable of solving multi-variable Partial Differential Equations (PDE), and can capture the dependence among the different variables in a single model. FNOs can also predict plasma evolution on real-world experimental data observed by the cameras positioned within the MAST Tokamak, i.e., cameras looking across the central solenoid and the divertor in the Tokamak. We show that FNOs are able to accurately forecast the evolution of plasma and have the potential to be deployed for real-time monitoring. We also illustrate their capability in forecasting the plasma shape, the locations of interactions of the plasma with the central solenoid and the divertor for the full (available) duration of the plasma shot within MAST. The FNO offers a viable alternative for surrogate modelling as it is quick to train and infer, and requires fewer data points, while being able to do zero-shot super-resolution and getting high-fidelity solutions.
Authors: Qi Hu, Haoran Li, Jiaxin Bai, Zihao Wang, Yangqiu Song
Abstract: In the era of large language models (LLMs), efficient and accurate data retrieval has become increasingly crucial for the use of domain-specific or private data in the retrieval augmented generation (RAG). Neural graph databases (NGDBs) have emerged as a powerful paradigm that combines the strengths of graph databases (GDBs) and neural networks to enable efficient storage, retrieval, and analysis of graph-structured data which can be adaptively trained with LLMs. The usage of neural embedding storage and Complex neural logical Query Answering (CQA) provides NGDBs with generalization ability. When the graph is incomplete, by extracting latent patterns and representations, neural graph databases can fill gaps in the graph structure, revealing hidden relationships and enabling accurate query answering. Nevertheless, this capability comes with inherent trade-offs, as it introduces additional privacy risks to the domain-specific or private databases. Malicious attackers can infer more sensitive information in the database using well-designed queries such as from the answer sets of where Turing Award winners born before 1950 and after 1940 lived, the living places of Turing Award winner Hinton are probably exposed, although the living places may have been deleted in the training stage due to the privacy concerns. In this work, we propose a privacy-preserved neural graph database (P-NGDB) framework to alleviate the risks of privacy leakage in NGDBs. We introduce adversarial training techniques in the training stage to enforce the NGDBs to generate indistinguishable answers when queried with private information, enhancing the difficulty of inferring sensitive information through combinations of multiple innocuous queries.
Authors: Diego Carraro, Derek Bridge
Abstract: It has long been recognized that it is not enough for a Recommender System (RS) to provide recommendations based only on their relevance to users. Among many other criteria, the set of recommendations may need to be diverse. Diversity is one way of handling recommendation uncertainty and ensuring that recommendations offer users a meaningful choice. The literature reports many ways of measuring diversity and improving the diversity of a set of recommendations, most notably by re-ranking and selecting from a larger set of candidate recommendations. Driven by promising insights from the literature on how to incorporate versatile Large Language Models (LLMs) into the RS pipeline, in this paper we show how LLMs can be used for diversity re-ranking. We begin with an informal study that verifies that LLMs can be used for re-ranking tasks and do have some understanding of the concept of item diversity. Then, we design a more rigorous methodology where LLMs are prompted to generate a diverse ranking from a candidate ranking using various prompt templates with different re-ranking instructions in a zero-shot fashion. We conduct comprehensive experiments testing state-of-the-art LLMs from the GPT and Llama families. We compare their re-ranking capabilities with random re-ranking and various traditional re-ranking methods from the literature. We open-source the code of our experiments for reproducibility. Our findings suggest that the trade-offs (in terms of performance and costs, among others) of LLM-based re-rankers are superior to those of random re-rankers but, as yet, inferior to the ones of traditional re-rankers. However, the LLM approach is promising. LLMs exhibit improved performance on many natural language processing and recommendation tasks and lower inference costs. Given these trends, we can expect LLM-based re-ranking to become more competitive soon.
Authors: Saqib Qamar, Abu Imran Baba, St\'ephane Verger, Magnus Andersson
Abstract: Wood comprises different cell types, such as fibers, tracheids and vessels, defining its properties. Studying cells' shape, size, and arrangement in microscopy images is crucial for understanding wood characteristics. Typically, this involves macerating (soaking) samples in a solution to separate cells, then spreading them on slides for imaging with a microscope that covers a wide area, capturing thousands of cells. However, these cells often cluster and overlap in images, making the segmentation difficult and time-consuming using standard image-processing methods. In this work, we developed an automatic deep learning segmentation approach that utilizes the one-stage YOLOv8 model for fast and accurate segmentation and characterization of macerated fiber and vessel form aspen trees in microscopy images. The model can analyze 32,640 x 25,920 pixels images and demonstrate effective cell detection and segmentation, achieving a mAP_{0.5-0.95} of 78 %. To assess the model's robustness, we examined fibers from a genetically modified tree line known for longer fibers. The outcomes were comparable to previous manual measurements. Additionally, we created a user-friendly web application for image analysis and provided the code for use on Google Colab. By leveraging YOLOv8's advances, this work provides a deep learning solution to enable efficient quantification and analysis of wood cells suitable for practical applications.
Authors: Jacob Fein-Ashley, Tian Ye, Sachini Wickramasinghe, Bingyi Zhang, Rajgopal Kannan, Viktor Prasanna
Abstract: Image classifiers often rely on convolutional neural networks (CNN) for their tasks, which are inherently more heavyweight than multilayer perceptrons (MLPs), which can be problematic in real-time applications. Additionally, many image classification models work on both RGB and grayscale datasets. Classifiers that operate solely on grayscale images are much less common. Grayscale image classification has diverse applications, including but not limited to medical image classification and synthetic aperture radar (SAR) automatic target recognition (ATR). Thus, we present a novel grayscale (single channel) image classification approach using a vectorized view of images. We exploit the lightweightness of MLPs by viewing images as a vector and reducing our problem setting to the grayscale image classification setting. We find that using a single graph convolutional layer batch-wise increases accuracy and reduces variance in the performance of our model. Moreover, we develop a customized accelerator on FPGA for the proposed model with several optimizations to improve its performance. Our experimental results on benchmark grayscale image datasets demonstrate the effectiveness of the proposed model, achieving vastly lower latency (up to 16$\times$ less) and competitive or leading performance compared to other state-of-the-art image classification models on various domain-specific grayscale image classification datasets.
Authors: Miranda Christ, Sam Gunn
Abstract: We construct pseudorandom error-correcting codes (or simply pseudorandom codes), which are error-correcting codes with the property that any polynomial number of codewords are pseudorandom to any computationally-bounded adversary. Efficient decoding of corrupted codewords is possible with the help of a decoding key. We build pseudorandom codes that are robust to substitution and deletion errors, where pseudorandomness rests on standard cryptographic assumptions. Specifically, pseudorandomness is based on either $2^{O(\sqrt{n})}$-hardness of LPN, or polynomial hardness of LPN and the planted XOR problem at low density. As our primary application of pseudorandom codes, we present an undetectable watermarking scheme for outputs of language models that is robust to cropping and a constant rate of random substitutions and deletions. The watermark is undetectable in the sense that any number of samples of watermarked text are computationally indistinguishable from text output by the original model. This is the first undetectable watermarking scheme that can tolerate a constant rate of errors. Our second application is to steganography, where a secret message is hidden in innocent-looking content. We present a constant-rate stateless steganography scheme with robustness to a constant rate of substitutions. Ours is the first stateless steganography scheme with provable steganographic security and any robustness to errors.
Authors: Yuzhe Zhang, Yipeng Zhang, Yidong Gan, Lina Yao, Chen Wang
Abstract: Causal graph recovery is traditionally done using statistical estimation-based methods or based on individual's knowledge about variables of interests. They often suffer from data collection biases and limitations of individuals' knowledge. The advance of large language models (LLMs) provides opportunities to address these problems. We propose a novel method that leverages LLMs to deduce causal relationships in general causal graph recovery tasks. This method leverages knowledge compressed in LLMs and knowledge LLMs extracted from scientific publication database as well as experiment data about factors of interest to achieve this goal. Our method gives a prompting strategy to extract associational relationships among those factors and a mechanism to perform causality verification for these associations. Comparing to other LLM-based methods that directly instruct LLMs to do the highly complex causal reasoning, our method shows clear advantage on causal graph quality on benchmark datasets. More importantly, as causality among some factors may change as new research results emerge, our method show sensitivity to new evidence in the literature and can provide useful information for updating causal graphs accordingly.
Authors: Anej Svete, Robin Shing Moon Chan, Ryan Cotterell
Abstract: Recent work by Hewitt et al. (2020) provides an interpretation of the empirical success of recurrent neural networks (RNNs) as language models (LMs). It shows that RNNs can efficiently represent bounded hierarchical structures that are prevalent in human language. This suggests that RNNs' success might be linked to their ability to model hierarchy. However, a closer inspection of Hewitt et al.'s (2020) construction shows that it is not inherently limited to hierarchical structures. This poses a natural question: What other classes of LMs can RNNs efficiently represent? To this end, we generalize Hewitt et al.'s (2020) construction and show that RNNs can efficiently represent a larger class of LMs than previously claimed -- specifically, those that can be represented by a pushdown automaton with a bounded stack and a specific stack update function. Altogether, the efficiency of representing this diverse class of LMs with RNN LMs suggests novel interpretations of their inductive bias.
Authors: Navid Mohammadi Foumani, Geoffrey Mackellar, Soheila Ghane, Saad Irtza, Nam Nguyen, Mahsa Salehi
Abstract: Self-supervised approaches for electroencephalography (EEG) representation learning face three specific challenges inherent to EEG data: (1) The low signal-to-noise ratio which challenges the quality of the representation learned, (2) The wide range of amplitudes from very small to relatively large due to factors such as the inter-subject variability, risks the models to be dominated by higher amplitude ranges, and (3) The absence of explicit segmentation in the continuous-valued sequences which can result in less informative representations. To address these challenges, we introduce \textit{EEG2Rep}, a self-prediction approach for self-supervised representation learning from EEG. Two core novel components of EEG2Rep are as follows: 1) Instead of learning to predict the masked input from raw EEG, EEG2Rep learns to predict masked input in latent representation space, and 2) Instead of conventional masking methods, EEG2Rep uses a new semantic subsequence preserving (SSP) method which provides informative masked inputs to guide EEG2Rep to generate rich semantic representations. In experiments on 6 diverse EEG tasks with subject variability, EEG2Rep significantly outperforms state-of-the-art methods. We show that our semantic subsequence preserving improves the existing masking methods in self-prediction literature and find that preserving 50\% of EEG recordings will result in the most accurate results on all 6 tasks on average. Finally, we show that EEG2Rep is robust to noise addressing a significant challenge that exists in EEG data. Models and code are available at:\url{https://github.com/Navidfoumani/EEG2Rep}
Authors: Zhongwei Wan, Che Liu, Xin Wang, Chaofan Tao, Hui Shen, Zhenwu Peng, Jie Fu, Rossella Arcucci, Huaxiu Yao, Mi Zhang
Abstract: Electrocardiogram (ECG) is the primary non-invasive diagnostic tool for monitoring cardiac conditions and is crucial in assisting clinicians. Recent studies have concentrated on classifying cardiac conditions using ECG data but have overlooked ECG report generation, which is time-consuming and requires clinical expertise. To automate ECG report generation and ensure its versatility, we propose the Multimodal ECG Instruction Tuning (MEIT) framework, the first attempt to tackle ECG report generation with LLMs and multimodal instructions. To facilitate future research, we establish a benchmark to evaluate MEIT with various LLMs backbones across two large-scale ECG datasets. Our approach uniquely aligns the representations of the ECG signal and the report, and we conduct extensive experiments to benchmark MEIT with nine open-source LLMs using more than 800,000 ECG reports. MEIT's results underscore the superior performance of instruction-tuned LLMs, showcasing their proficiency in quality report generation, zero-shot capabilities, and resilience to signal perturbation. These findings emphasize the efficacy of our MEIT framework and its potential for real-world clinical application.
Authors: Zhengrui Guo, Jiabo Ma, Yingxue Xu, Yihui Wang, Liansheng Wang, Hao Chen
Abstract: Histopathology serves as the gold standard in cancer diagnosis, with clinical reports being vital in interpreting and understanding this process, guiding cancer treatment and patient care. The automation of histopathology report generation with deep learning stands to significantly enhance clinical efficiency and lessen the labor-intensive, time-consuming burden on pathologists in report writing. In pursuit of this advancement, we introduce HistGen, a multiple instance learning-empowered framework for histopathology report generation together with the first benchmark dataset for evaluation. Inspired by diagnostic and report-writing workflows, HistGen features two delicately designed modules, aiming to boost report generation by aligning whole slide images (WSIs) and diagnostic reports from local and global granularity. To achieve this, a local-global hierarchical encoder is developed for efficient visual feature aggregation from a region-to-slide perspective. Meanwhile, a cross-modal context module is proposed to explicitly facilitate alignment and interaction between distinct modalities, effectively bridging the gap between the extensive visual sequences of WSIs and corresponding highly summarized reports. Experimental results on WSI report generation show the proposed model outperforms state-of-the-art (SOTA) models by a large margin. Moreover, the results of fine-tuning our model on cancer subtyping and survival analysis tasks further demonstrate superior performance compared to SOTA methods, showcasing strong transfer learning capability. Dataset, model weights, and source code are available in https://github.com/dddavid4real/HistGen.
Authors: Yuxuan Li, Xiang Li, Yimain Dai, Qibin Hou, Li Liu, Yongxiang Liu, Ming-Ming Cheng, Jian Yang
Abstract: Remote sensing images pose distinct challenges for downstream tasks due to their inherent complexity. While a considerable amount of research has been dedicated to remote sensing classification, object detection and semantic segmentation, most of these studies have overlooked the valuable prior knowledge embedded within remote sensing scenarios. Such prior knowledge can be useful because remote sensing objects may be mistakenly recognized without referencing a sufficiently long-range context, which can vary for different objects. This paper considers these priors and proposes a lightweight Large Selective Kernel Network (LSKNet) backbone. LSKNet can dynamically adjust its large spatial receptive field to better model the ranging context of various objects in remote sensing scenarios. To our knowledge, large and selective kernel mechanisms have not been previously explored in remote sensing images. Without bells and whistles, our lightweight LSKNet sets new state-of-the-art scores on standard remote sensing classification, object detection and semantic segmentation benchmarks. Our comprehensive analysis further validated the significance of the identified priors and the effectiveness of LSKNet. The code is available at https://github.com/zcablii/LSKNet.
Authors: Lingkai Kong, Molei Tao
Abstract: Explicit, momentum-based dynamics for optimizing functions defined on Lie groups was recently constructed, based on techniques such as variational optimization and left trivialization. We appropriately add tractable noise to the optimization dynamics to turn it into a sampling dynamics, leveraging the advantageous feature that the trivialized momentum variable is Euclidean despite that the potential function lives on a manifold. We then propose a Lie-group MCMC sampler, by delicately discretizing the resulting kinetic-Langevin-type sampling dynamics. The Lie group structure is exactly preserved by this discretization. Exponential convergence with explicit convergence rate for both the continuous dynamics and the discrete sampler are then proved under $W_2$ distance. Only compactness of the Lie group and geodesically $L$-smoothness of the potential function are needed. To the best of our knowledge, this is the first convergence result for kinetic Langevin on curved spaces, and also the first quantitative result that requires no convexity or, at least not explicitly, any common relaxation such as isoperimetry.
Authors: Zhiqiu Lin, Deepak Pathak, Baiqi Li, Jiayao Li, Xide Xia, Graham Neubig, Pengchuan Zhang, Deva Ramanan
Abstract: Despite significant progress in generative AI, comprehensive evaluation remains challenging because of the lack of effective metrics and standardized benchmarks. For instance, the widely-used CLIPScore measures the alignment between a (generated) image and text prompt, but it fails to produce reliable scores for complex prompts involving compositions of objects, attributes, and relations. One reason is that text encoders of CLIP can notoriously act as a "bag of words", conflating prompts such as "the horse is eating the grass" with "the grass is eating the horse". To address this, we introduce the VQAScore, which uses a visual-question-answering (VQA) model to produce an alignment score by computing the probability of a "Yes" answer to a simple "Does this figure show '{text}'?" question. Though simpler than prior art, VQAScore computed with off-the-shelf models produces state-of-the-art results across many (8) image-text alignment benchmarks. We also compute VQAScore with an in-house model that follows best practices in the literature. For example, we use a bidirectional image-question encoder that allows image embeddings to depend on the question being asked (and vice versa). Our in-house model, CLIP-FlanT5, outperforms even the strongest baselines that make use of the proprietary GPT-4V. Interestingly, although we train with only images, VQAScore can also align text with video and 3D models. VQAScore allows researchers to benchmark text-to-visual generation using complex texts that capture the compositional structure of real-world prompts. We introduce GenAI-Bench, a more challenging benchmark with 1,600 compositional text prompts that require parsing scenes, objects, attributes, relationships, and high-order reasoning like comparison and logic. GenAI-Bench also offers over 15,000 human ratings for leading image and video generation models such as Stable Diffusion, DALL-E 3, and Gen2.
Authors: Ying Li, Zhidi Lin, Feng Yin, Michael Minyi Zhang
Abstract: Gaussian process latent variable models (GPLVMs) are a versatile family of unsupervised learning models commonly used for dimensionality reduction. However, common challenges in modeling data with GPLVMs include inadequate kernel flexibility and improper selection of the projection noise, leading to a type of model collapse characterized by vague latent representations that do not reflect the underlying data structure. This paper addresses these issues by, first, theoretically examining the impact of projection variance on model collapse through the lens of a linear GPLVM. Second, we tackle model collapse due to inadequate kernel flexibility by integrating the spectral mixture (SM) kernel and a differentiable random Fourier feature (RFF) kernel approximation, which ensures computational scalability and efficiency through off-the-shelf automatic differentiation tools for learning the kernel hyperparameters, projection variance, and latent representations within the variational inference framework. The proposed GPLVM, named advisedRFLVM, is evaluated across diverse datasets and consistently outperforms various salient competing models, including state-of-the-art variational autoencoders (VAEs) and other GPLVM variants, in terms of informative latent representations and missing data imputation.
Authors: Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
Abstract: We show that even the most recent safety-aligned LLMs are not robust to simple adaptive jailbreaking attacks. First, we demonstrate how to successfully leverage access to logprobs for jailbreaking: we initially design an adversarial prompt template (sometimes adapted to the target LLM), and then we apply random search on a suffix to maximize a target logprob (e.g., of the token ``Sure''), potentially with multiple restarts. In this way, we achieve nearly 100% attack success rate -- according to GPT-4 as a judge -- on Vicuna-13B, Mistral-7B, Phi-3-Mini, Nemotron-4-340B, Llama-2-Chat-7B/13B/70B, Llama-3-Instruct-8B, Gemma-7B, GPT-3.5, GPT-4, and R2D2 from HarmBench that was adversarially trained against the GCG attack. We also show how to jailbreak all Claude models -- that do not expose logprobs -- via either a transfer or prefilling attack with a 100% success rate. In addition, we show how to use random search on a restricted set of tokens for finding trojan strings in poisoned models -- a task that shares many similarities with jailbreaking -- which is the algorithm that brought us the first place in the SaTML'24 Trojan Detection Competition. The common theme behind these attacks is that adaptivity is crucial: different models are vulnerable to different prompting templates (e.g., R2D2 is very sensitive to in-context learning prompts), some models have unique vulnerabilities based on their APIs (e.g., prefilling for Claude), and in some settings, it is crucial to restrict the token search space based on prior knowledge (e.g., for trojan detection). For reproducibility purposes, we provide the code, logs, and jailbreak artifacts in the JailbreakBench format at https://github.com/tml-epfl/llm-adaptive-attacks.
Authors: Rachmad Vidya Wicaksana Putra, Alberto Marchisio, Muhammad Shafique
Abstract: Recent trends have shown that autonomous agents, such as Autonomous Ground Vehicles (AGVs), Unmanned Aerial Vehicles (UAVs), and mobile robots, effectively improve human productivity in solving diverse tasks. However, since these agents are typically powered by portable batteries, they require extremely low power/energy consumption to operate in a long lifespan. To solve this challenge, neuromorphic computing has emerged as a promising solution, where bio-inspired Spiking Neural Networks (SNNs) use spikes from event-based cameras or data conversion pre-processing to perform sparse computations efficiently. However, the studies of SNN deployments for autonomous agents are still at an early stage. Hence, the optimization stages for enabling efficient embodied SNN deployments for autonomous agents have not been defined systematically. Toward this, we propose a novel framework called SNN4Agents that consists of a set of optimization techniques for designing energy-efficient embodied SNNs targeting autonomous agent applications. Our SNN4Agents employs weight quantization, timestep reduction, and attention window reduction to jointly improve the energy efficiency, reduce the memory footprint, optimize the processing latency, while maintaining high accuracy. In the evaluation, we investigate use cases of event-based car recognition, and explore the trade-offs among accuracy, latency, memory, and energy consumption. The experimental results show that our proposed framework can maintain high accuracy (i.e., 84.12% accuracy) with 68.75% memory saving, 3.58x speed-up, and 4.03x energy efficiency improvement as compared to the state-of-the-art work for NCARS dataset. In this manner, our SNN4Agents framework paves the way toward enabling energy-efficient embodied SNN deployments for autonomous agents.
Authors: Liwei Kang, Zirui Zhao, David Hsu, Wee Sun Lee
Abstract: Chain-of-thought (CoT), tree-of-thought (ToT), and related techniques work surprisingly well in practice for some complex reasoning tasks with Large Language Models (LLMs), but why? This work seeks the underlying reasons by conducting experimental case studies and linking the performance benefits to well-established sample and computational complexity principles in machine learning. We experimented with 6 reasoning tasks, ranging from grade school math, air travel planning, ..., to Blocksworld. The results suggest that (i) both CoT and ToT benefit significantly from task decomposition, which breaks a complex reasoning task into a sequence of steps with low sample complexity and explicitly outlines the reasoning structure, and (ii) for computationally hard reasoning tasks, the more sophisticated tree structure of ToT outperforms the linear structure of CoT. These findings provide useful guidelines for the use of LLM in solving reasoning tasks in practice.
Authors: Ross Greer, Bj{\o}rk Antoniussen, Andreas M{\o}gelmose, Mohan Trivedi
Abstract: Object detection is crucial for ensuring safe autonomous driving. However, data-driven approaches face challenges when encountering minority or novel objects in the 3D driving scene. In this paper, we propose VisLED, a language-driven active learning framework for diverse open-set 3D Object Detection. Our method leverages active learning techniques to query diverse and informative data samples from an unlabeled pool, enhancing the model's ability to detect underrepresented or novel objects. Specifically, we introduce the Vision-Language Embedding Diversity Querying (VisLED-Querying) algorithm, which operates in both open-world exploring and closed-world mining settings. In open-world exploring, VisLED-Querying selects data points most novel relative to existing data, while in closed-world mining, it mines novel instances of known classes. We evaluate our approach on the nuScenes dataset and demonstrate its efficiency compared to random sampling and entropy-querying methods. Our results show that VisLED-Querying consistently outperforms random sampling and offers competitive performance compared to entropy-querying despite the latter's model-optimality, highlighting the potential of VisLED for improving object detection in autonomous driving scenarios. We make our code publicly available at https://github.com/Bjork-crypto/VisLED-Querying
Authors: Haniyeh Ehsani Oskouie, Mohammad-Shahram Moin, Shohreh Kasaei
Abstract: Deep neural networks have made significant advancements in accurately estimating scene flow using point clouds, which is vital for many applications like video analysis, action recognition, and navigation. The robustness of these techniques, however, remains a concern, particularly in the face of adversarial attacks that have been proven to deceive state-of-the-art deep neural networks in many domains. Surprisingly, the robustness of scene flow networks against such attacks has not been thoroughly investigated. To address this problem, the proposed approach aims to bridge this gap by introducing adversarial white-box attacks specifically tailored for scene flow networks. Experimental results show that the generated adversarial examples obtain up to 33.7 relative degradation in average end-point error on the KITTI and FlyingThings3D datasets. The study also reveals the significant impact that attacks targeting point clouds in only one dimension or color channel have on average end-point error. Analyzing the success and failure of these attacks on the scene flow networks and their 2D optical flow network variants shows a higher vulnerability for the optical flow networks.
Authors: Anej Svete, Ryan Cotterell
Abstract: Existing work has analyzed the representational capacity of the transformer architecture by means of formal models of computation. However, the focus so far has been on analyzing the architecture in terms of language \emph{acceptance}. We contend that this is an ill-suited problem in the study of \emph{language models} (LMs), which are definitionally \emph{probability distributions} over strings. In this paper, we focus on the relationship between transformer LMs and $n$-gram LMs, a simple and historically relevant class of language models. We show that transformer LMs using the hard or sparse attention mechanisms can exactly represent any $n$-gram LM, giving us a concrete lower bound on their probabilistic representational capacity. This provides a first step towards understanding the mechanisms that transformer LMs can use to represent probability distributions over strings.
Authors: Yuxi Xie, Anirudh Goyal, Wenyue Zheng, Min-Yen Kan, Timothy P. Lillicrap, Kenji Kawaguchi, Michael Shieh
Abstract: We introduce an approach aimed at enhancing the reasoning capabilities of Large Language Models (LLMs) through an iterative preference learning process inspired by the successful strategy employed by AlphaZero. Our work leverages Monte Carlo Tree Search (MCTS) to iteratively collect preference data, utilizing its look-ahead ability to break down instance-level rewards into more granular step-level signals. To enhance consistency in intermediate steps, we combine outcome validation and stepwise self-evaluation, continually updating the quality assessment of newly generated data. The proposed algorithm employs Direct Preference Optimization (DPO) to update the LLM policy using this newly generated step-level preference data. Theoretical analysis reveals the importance of using on-policy sampled data for successful self-improving. Extensive evaluations on various arithmetic and commonsense reasoning tasks demonstrate remarkable performance improvements over existing models. For instance, our approach outperforms the Mistral-7B Supervised Fine-Tuning (SFT) baseline on GSM8K, MATH, and ARC-C, with substantial increases in accuracy to $81.8\%$ (+$5.9\%$), $34.7\%$ (+$5.8\%$), and $76.4\%$ (+$15.8\%$), respectively. Additionally, our research delves into the training and inference compute tradeoff, providing insights into how our method effectively maximizes performance gains. Our code is publicly available at https://github.com/YuxiXie/MCTS-DPO.
Authors: Zhenqiao Song, Tinglin Huang, Lei Li, Wengong Jin
Abstract: How can we design proteins with desired functions? We are motivated by a chemical intuition that both geometric structure and biochemical properties are critical to a protein's function. In this paper, we propose SurfPro, a new method to generate functional proteins given a desired surface and its associated biochemical properties. SurfPro comprises a hierarchical encoder that progressively models the geometric shape and biochemical features of a protein surface, and an autoregressive decoder to produce an amino acid sequence. We evaluate SurfPro on a standard inverse folding benchmark CATH 4.2 and two functional protein design tasks: protein binder design and enzyme design. Our SurfPro consistently surpasses previous state-of-the-art inverse folding methods, achieving a recovery rate of 57.78% on CATH 4.2 and higher success rates in terms of protein-protein binding and enzyme-substrate interaction scores.
Authors: Johannes R\"uckert, Louise Bloch, Raphael Br\"ungel, Ahmad Idrissi-Yaghir, Henning Sch\"afer, Cynthia S. Schmidt, Sven Koitka, Obioma Pelka, Asma Ben Abacha, Alba G. Seco de Herrera, Henning M\"uller, Peter A. Horn, Felix Nensa, Christoph M. Friedrich
Abstract: Automated medical image analysis systems often require large amounts of training data with high quality labels, which are difficult and time consuming to generate. This paper introduces Radiology Object in COntext version 2 (ROCOv2), a multimodal dataset consisting of radiological images and associated medical concepts and captions extracted from the PMC Open Access subset. It is an updated version of the ROCO dataset published in 2018, and adds 35,705 new images added to PMC since 2018. It further provides manually curated concepts for imaging modalities with additional anatomical and directional concepts for X-rays. The dataset consists of 79,789 images and has been used, with minor modifications, in the concept detection and caption prediction tasks of ImageCLEFmedical Caption 2023. The dataset is suitable for training image annotation models based on image-caption pairs, or for multi-label image classification using Unified Medical Language System (UMLS) concepts provided with each image. In addition, it can serve for pre-training of medical domain models, and evaluation of deep learning models for multi-task learning.
Authors: Yuyang Xue, Jingshuai Liu, Steven McDonagh, Sotirios A. Tsaftaris
Abstract: Machine unlearning is a promising paradigm for removing unwanted data samples from a trained model, towards ensuring compliance with privacy regulations and limiting harmful biases. Although unlearning has been shown in, e.g., classification and recommendation systems, its potential in medical image-to-image translation, specifically in image recon-struction, has not been thoroughly investigated. This paper shows that machine unlearning is possible in MRI tasks and has the potential to benefit for bias removal. We set up a protocol to study how much shared knowledge exists between datasets of different organs, allowing us to effectively quantify the effect of unlearning. Our study reveals that combining training data can lead to hallucinations and reduced image quality in the reconstructed data. We use unlearning to remove hallucinations as a proxy exemplar of undesired data removal. Indeed, we show that machine unlearning is possible without full retraining. Furthermore, our observations indicate that maintaining high performance is feasible even when using only a subset of retain data. We have made our code publicly accessible.
Authors: Xiaodong Liu
Abstract: This paper presents a significant improvement on the previous conference paper known as DefSent. The prior study seeks to improve sentence embeddings of language models by projecting definition sentences into the vector space of dictionary entries. We discover that this approach is not fully explored due to the methodological limitation of using word embeddings of language models to represent dictionary entries. This leads to two hindrances. First, dictionary entries are constrained by the single-word vocabulary, and thus cannot be fully exploited. Second, semantic representations of language models are known to be anisotropic, but pre-processing word embeddings for DefSent is not allowed because its weight is frozen during training and tied to the prediction layer. In this paper, we propose a novel method to progressively build entry embeddings not subject to the limitations. As a result, definition sentences can be projected into a quasi-isotropic or isotropic vector space of unlimited dictionary entries, so that sentence embeddings of noticeably better quality are attainable. We abbreviate our approach as DefSent+ (a plus version of DefSent), involving the following strengths: 1) the task performance on measuring sentence similarities is significantly improved compared to DefSent; 2) when DefSent+ is used to further train data-augmented models like SIMCSE, SNCSE, and SynCSE, state-of-the-art performance on measuring sentence similarities can be achieved among the approaches without using manually labeled datasets; 3) DefSent+ is also competitive in feature-based transfer for NLP downstream tasks.
Authors: F\'elix Marcoccia, C\'edric Adjih, Paul M\"uhlethaler
Abstract: Most of multiple link prediction or graph generation techniques rely on the attention mechanism or on Graph Neural Networks (GNNs), which consist in leveraging node-level information exchanges in order to form proper link predictions. Such node-level interactions do not process nodes as an ordered sequence, which would imply some kind of natural ordering of the nodes: they are said to be permutation invariant mechanisms. They are well suited for graph problems, but struggle at providing a global orchestration of the predicted links, which can result in a loss of performance. Some typical issues can be the difficulty to ensure high-level properties such as global connectedness, fixed diameter or to avoid information bottleneck effects such as oversmoothing and oversquashing, which respectively consist in abundant smoothing in dense areas leading to a loss of information and a tendency to exclude isolated nodes from the message passing scheme, and often result in irrelevant, unbalanced link predictions. To tackle this problem, we hereby present Cross-Attentive Modulation (CAM) tokens, which introduce cross-attentive units used to condition node and edge-level modulations in order to enable context-aware computations that improve the global consistency of the prediction links. We will implement it on a few permutation invariant architectures, and showcase benchmarks that prove the merits of our work.
Authors: Owen Dugan, Donato Manuel Jimenez Beneto, Charlotte Loh, Zhuo Chen, Rumen Dangovski, Marin Solja\v{c}i\'c
Abstract: Despite significant advancements in text generation and reasoning, Large Language Models (LLMs) still face challenges in accurately performing complex arithmetic operations. To achieve accurate calculations, language model systems often enable LLMs to generate code for arithmetic operations. However, this approach compromises speed and security and, if finetuning is involved, risks the language model losing prior capabilities. We propose a framework that enables exact arithmetic in \textit{a single autoregressive step}, providing faster, more secure, and more interpretable LLM systems with arithmetic capabilities. We use the hidden states of an LLM to control a symbolic architecture which performs arithmetic. Our implementation using Llama 3 8B Instruct with OccamNet as a symbolic model (OccamLlama) achieves 100\% accuracy on single arithmetic operations ($+,-,\times,\div,\sin{},\cos{},\log{},\exp{},\sqrt{}$), outperforming GPT 4o and on par with GPT 4o using a code interpreter. OccamLlama also outperforms GPT 4o both with and without a code interpreter on mathematical problem solving benchmarks involving challenging arithmetic, thus enabling small LLMs to match the arithmetic performance of even much larger models. We will make our code public shortly.
Authors: Alan Nawzad Amin, Andrew Gordon Wilson
Abstract: To make accurate predictions, understand mechanisms, and design interventions in systems of many variables, we wish to learn causal graphs from large scale data. Unfortunately the space of all possible causal graphs is enormous so scalably and accurately searching for the best fit to the data is a challenge. In principle we could substantially decrease the search space, or learn the graph entirely, by testing the conditional independence of variables. However, deciding if two variables are adjacent in a causal graph may require an exponential number of tests. Here we build a scalable and flexible method to evaluate if two variables are adjacent in a causal graph, the Differentiable Adjacency Test (DAT). DAT replaces an exponential number of tests with a provably equivalent relaxed problem. It then solves this problem by training two neural networks. We build a graph learning method based on DAT, DAT-Graph, that can also learn from data with interventions. DAT-Graph can learn graphs of 1000 variables with state of the art accuracy. Using the graph learned by DAT-Graph, we also build models that make much more accurate predictions of the effects of interventions on large scale RNA sequencing data.
Authors: Sumio Watanabe
Abstract: Mathematical equivalence between statistical mechanics and machine learning theory has been known since the 20th century, and researches based on such equivalence have provided novel methodology in both theoretical physics and statistical learning theory. For example, algebraic approach in statistical mechanics such as operator algebra enables us to analyze phase transition phenomena mathematically. In this paper, for theoretical physicists who are interested in artificial intelligence, we review and prospect algebraic researches in machine learning theory. If a learning machine has hierarchical structure or latent variables, then the random Hamiltonian cannot be expressed by any quadratic perturbation because it has singularities. To study an equilibrium state defined by such a singular random Hamiltonian, algebraic approach is necessary to derive asymptotic form of the free energy and the generalization error. We also introduce the most recent advance, in fact, theoretical foundation for alignment of artificial intelligence is now being constructed based on algebraic learning theory. This paper is devoted to the memory of Professor Huzihiro Araki who is a pioneer founder of algebraic research in both statistical mechanics and quantum field theory.
Authors: Saturnino Luz, Sofia De La Fuente Garcia, Fasih Haider, Davida Fromm, Brian MacWhinney, Alyssa Lanzi, Ya-Ning Chang, Chia-Ju Chou, Yi-Chien Liu
Abstract: We present a novel benchmark dataset and prediction tasks for investigating approaches to assess cognitive function through analysis of connected speech. The dataset consists of speech samples and clinical information for speakers of Mandarin Chinese and English with different levels of cognitive impairment as well as individuals with normal cognition. These data have been carefully matched by age and sex by propensity score analysis to ensure balance and representativity in model training. The prediction tasks encompass mild cognitive impairment diagnosis and cognitive test score prediction. This framework was designed to encourage the development of approaches to speech-based cognitive assessment which generalise across languages. We illustrate it by presenting baseline prediction models that employ language-agnostic and comparable features for diagnosis and cognitive test score prediction. The models achieved unweighted average recall was 59.2% in diagnosis, and root mean squared error of 2.89 in score prediction.
Authors: Minghong Fang (Kevin), Zifan Zhang (Kevin), Hairi (Kevin), Prashant Khanduri (Kevin), Jia (Kevin), Liu, Songtao Lu, Yuchen Liu, Neil Gong
Abstract: Federated learning (FL) enables multiple clients to collaboratively train machine learning models without revealing their private training data. In conventional FL, the system follows the server-assisted architecture (server-assisted FL), where the training process is coordinated by a central server. However, the server-assisted FL framework suffers from poor scalability due to a communication bottleneck at the server, and trust dependency issues. To address challenges, decentralized federated learning (DFL) architecture has been proposed to allow clients to train models collaboratively in a serverless and peer-to-peer manner. However, due to its fully decentralized nature, DFL is highly vulnerable to poisoning attacks, where malicious clients could manipulate the system by sending carefully-crafted local models to their neighboring clients. To date, only a limited number of Byzantine-robust DFL methods have been proposed, most of which are either communication-inefficient or remain vulnerable to advanced poisoning attacks. In this paper, we propose a new algorithm called BALANCE (Byzantine-robust averaging through local similarity in decentralization) to defend against poisoning attacks in DFL. In BALANCE, each client leverages its own local model as a similarity reference to determine if the received model is malicious or benign. We establish the theoretical convergence guarantee for BALANCE under poisoning attacks in both strongly convex and non-convex settings. Furthermore, the convergence rate of BALANCE under poisoning attacks matches those of the state-of-the-art counterparts in Byzantine-free settings. Extensive experiments also demonstrate that BALANCE outperforms existing DFL methods and effectively defends against poisoning attacks.
Authors: Oluwaseun T. Ajayi, Yu Cheng
Abstract: With the prevailing efforts to combat the coronavirus disease 2019 (COVID-19) pandemic, there are still uncertainties that are yet to be discovered about its spread, future impact, and resurgence. In this paper, we present a three-stage data-driven approach to distill the hidden information about COVID-19. The first stage employs a Bayesian network structure learning method to identify the causal relationships among COVID-19 symptoms and their intrinsic demographic variables. As a second stage, the output from the Bayesian network structure learning, serves as a useful guide to train an unsupervised machine learning (ML) algorithm that uncovers the similarities in patients' symptoms through clustering. The final stage then leverages the labels obtained from clustering to train a demographic symptom identification (DSID) model which predicts a patient's symptom class and the corresponding demographic probability distribution. We applied our method on the COVID-19 dataset obtained from the Centers for Disease Control and Prevention (CDC) in the United States. Results from the experiments show a testing accuracy of 99.99%, as against the 41.15% accuracy of a heuristic ML method. This strongly reveals the viability of our Bayesian network and ML approach in understanding the relationship between the virus symptoms, and providing insights on patients' stratification towards reducing the severity of the virus.
Authors: Eden Grad, Moshe Kimhi, Lion Halika, Chaim Baskin
Abstract: Obtaining accurate labels for instance segmentation is particularly challenging due to the complex nature of the task. Each image necessitates multiple annotations, encompassing not only the object's class but also its precise spatial boundaries. These requirements elevate the likelihood of errors and inconsistencies in both manual and automated annotation processes. By simulating different noise conditions, we provide a realistic scenario for assessing the robustness and generalization capabilities of instance segmentation models in different segmentation tasks, introducing COCO-N and Cityscapes-N. We also propose a benchmark for weakly annotation noise, dubbed COCO-WAN, which utilizes foundation models and weak annotations to simulate semi-automated annotation tools and their noisy labels. This study sheds light on the quality of segmentation masks produced by various models and challenges the efficacy of popular methods designed to address learning with label noise.
Authors: Amit Das, Zheng Zhang, Fatemeh Jamshidi, Vinija Jain, Aman Chadha, Nilanjana Raychawdhary, Mary Sandage, Lauramarie Pope, Gerry Dozier, Cheryl Seals
Abstract: Data annotation, the practice of assigning descriptive labels to raw data, is pivotal in optimizing the performance of machine learning models. However, it is a resource-intensive process susceptible to biases introduced by annotators. The emergence of sophisticated Large Language Models (LLMs), like ChatGPT presents a unique opportunity to modernize and streamline this complex procedure. While existing research extensively evaluates the efficacy of LLMs, as annotators, this paper delves into the biases present in LLMs, specifically GPT 3.5 and GPT 4o when annotating hate speech data. Our research contributes to understanding biases in four key categories: gender, race, religion, and disability. Specifically targeting highly vulnerable groups within these categories, we analyze annotator biases. Furthermore, we conduct a comprehensive examination of potential factors contributing to these biases by scrutinizing the annotated data. We introduce our custom hate speech detection dataset, HateSpeechCorpus, to conduct this research. Additionally, we perform the same experiments on the ETHOS (Mollas et al., 2022) dataset also for comparative analysis. This paper serves as a crucial resource, guiding researchers and practitioners in harnessing the potential of LLMs for dataannotation, thereby fostering advancements in this critical field. The HateSpeechCorpus dataset is available here: https://github.com/AmitDasRup123/HateSpeechCorpus
Authors: Ruoyu Wang, Chen Cai, Wenqian Wang, Jianjun Gao, Dan Lin, Wenyang Liu, Kim-Hui Yap
Abstract: Driver action recognition has significantly advanced in enhancing driver-vehicle interactions and ensuring driving safety by integrating multiple modalities, such as infrared and depth. Nevertheless, compared to RGB modality only, it is always laborious and costly to collect extensive data for all types of non-RGB modalities in car cabin environments. Therefore, previous works have suggested independently learning each non-RGB modality by fine-tuning a model pre-trained on RGB videos, but these methods are less effective in extracting informative features when faced with newly-incoming modalities due to large domain gaps. In contrast, we propose a Continual Cross-Modal Mapping Network (CM2-Net) to continually learn each newly-incoming modality with instructive prompts from the previously-learned modalities. Specifically, we have developed Accumulative Cross-modal Mapping Prompting (ACMP), to map the discriminative and informative features learned from previous modalities into the feature space of newly-incoming modalities. Then, when faced with newly-incoming modalities, these mapped features are able to provide effective prompts for which features should be extracted and prioritized. These prompts are accumulating throughout the continual learning process, thereby boosting further recognition performances. Extensive experiments conducted on the Drive&Act dataset demonstrate the performance superiority of CM2-Net on both uni- and multi-modal driver action recognition.
Authors: Huaiguang Cai
Abstract: Understanding the decision-making process of machine learning models is crucial for ensuring trustworthy machine learning. Data Shapley, a landmark study on data valuation, advances this understanding by assessing the contribution of each datum to model accuracy. However, the resource-intensive and time-consuming nature of multiple model retraining poses challenges for applying Data Shapley to large datasets. To address this, we propose the CHG (Conduct of Hardness and Gradient) score, which approximates the utility of each data subset on model accuracy during a single model training. By deriving the closed-form expression of the Shapley value for each data point under the CHG score utility function, we reduce the computational complexity to the equivalent of a single model retraining, an exponential improvement over existing methods. Additionally, we employ CHG Shapley for real-time data selection, demonstrating its effectiveness in identifying high-value and noisy data. CHG Shapley facilitates trustworthy model training through efficient data valuation, introducing a novel data-centric perspective on trustworthy machine learning.