id
stringlengths
9
16
title
stringlengths
4
278
categories
stringlengths
5
104
abstract
stringlengths
6
4.09k
2502.06485
WyckoffDiff -- A Generative Diffusion Model for Crystal Symmetry
cond-mat.mtrl-sci cs.AI cs.LG
Crystalline materials often exhibit a high level of symmetry. However, most generative models do not account for symmetry, but rather model each atom without any constraints on its position or element. We propose a generative model, Wyckoff Diffusion (WyckoffDiff), which generates symmetry-based descriptions of crystals. This is enabled by considering a crystal structure representation that encodes all symmetry, and we design a novel neural network architecture which enables using this representation inside a discrete generative model framework. In addition to respecting symmetry by construction, the discrete nature of our model enables fast generation. We additionally present a new metric, Fr\'echet Wrenformer Distance, which captures the symmetry aspects of the materials generated, and we benchmark WyckoffDiff against recently proposed generative models for crystal generation.
2502.06486
Biomechanical Reconstruction with Confidence Intervals from Multiview Markerless Motion Capture
cs.CV
Advances in multiview markerless motion capture (MMMC) promise high-quality movement analysis for clinical practice and research. While prior validation studies show MMMC performs well on average, they do not provide what is needed in clinical practice or for large-scale utilization of MMMC -- confidence intervals over specific kinematic estimates from a specific individual analyzed using a possibly unique camera configuration. We extend our previous work using an implicit representation of trajectories optimized end-to-end through a differentiable biomechanical model to learn the posterior probability distribution over pose given all the detected keypoints. This posterior probability is learned through a variational approximation and estimates confidence intervals for individual joints at each moment in a trial, showing confidence intervals generally within 10-15 mm of spatial error for virtual marker locations, consistent with our prior validation studies. Confidence intervals over joint angles are typically only a few degrees and widen for more distal joints. The posterior also models the correlation structure over joint angles, such as correlations between hip and pelvis angles. The confidence intervals estimated through this method allow us to identify times and trials where kinematic uncertainty is high.
2502.06487
Adaptive Prompting: Ad-hoc Prompt Composition for Social Bias Detection
cs.CL
Recent advances on instruction fine-tuning have led to the development of various prompting techniques for large language models, such as explicit reasoning steps. However, the success of techniques depends on various parameters, such as the task, language model, and context provided. Finding an effective prompt is, therefore, often a trial-and-error process. Most existing approaches to automatic prompting aim to optimize individual techniques instead of compositions of techniques and their dependence on the input. To fill this gap, we propose an adaptive prompting approach that predicts the optimal prompt composition ad-hoc for a given input. We apply our approach to social bias detection, a highly context-dependent task that requires semantic understanding. We evaluate it with three large language models on three datasets, comparing compositions to individual techniques and other baselines. The results underline the importance of finding an effective prompt composition. Our approach robustly ensures high detection performance, and is best in several settings. Moreover, first experiments on other tasks support its generalizability.
2502.06490
Recent Advances in Discrete Speech Tokens: A Review
eess.AS cs.AI cs.MM cs.SD eess.SP
The rapid advancement of speech generation technologies in the era of large language models (LLMs) has established discrete speech tokens as a foundational paradigm for speech representation. These tokens, characterized by their discrete, compact, and concise nature, are not only advantageous for efficient transmission and storage, but also inherently compatible with the language modeling framework, enabling seamless integration of speech into text-dominated LLM architectures. Current research categorizes discrete speech tokens into two principal classes: acoustic tokens and semantic tokens, each of which has evolved into a rich research domain characterized by unique design philosophies and methodological approaches. This survey systematically synthesizes the existing taxonomy and recent innovations in discrete speech tokenization, conducts a critical examination of the strengths and limitations of each paradigm, and presents systematic experimental comparisons across token types. Furthermore, we identify persistent challenges in the field and propose potential research directions, aiming to offer actionable insights to inspire future advancements in the development and application of discrete speech tokens.
2502.06491
Model-Based Offline Reinforcement Learning with Reliability-Guaranteed Sequence Modeling
cs.LG cs.AI
Model-based offline reinforcement learning (MORL) aims to learn a policy by exploiting a dynamics model derived from an existing dataset. Applying conservative quantification to the dynamics model, most existing works on MORL generate trajectories that approximate the real data distribution to facilitate policy learning by using current information (e.g., the state and action at time step $t$). However, these works neglect the impact of historical information on environmental dynamics, leading to the generation of unreliable trajectories that may not align with the real data distribution. In this paper, we propose a new MORL algorithm \textbf{R}eliability-guaranteed \textbf{T}ransformer (RT), which can eliminate unreliable trajectories by calculating the cumulative reliability of the generated trajectory (i.e., using a weighted variational distance away from the real data). Moreover, by sampling candidate actions with high rewards, RT can efficiently generate high-return trajectories from the existing offline data. We theoretically prove the performance guarantees of RT in policy learning, and empirically demonstrate its effectiveness against state-of-the-art model-based methods on several benchmark tasks.
2502.06494
GuideLLM: Exploring LLM-Guided Conversation with Applications in Autobiography Interviewing
cs.CL cs.AI
Although Large Language Models (LLMs) succeed in human-guided conversations such as instruction following and question answering, the potential of LLM-guided conversations-where LLMs direct the discourse and steer the conversation's objectives-remains under-explored. In this study, we first characterize LLM-guided conversation into three fundamental components: (i) Goal Navigation; (ii) Context Management; (iii) Empathetic Engagement, and propose GuideLLM as an installation. We then implement an interviewing environment for the evaluation of LLM-guided conversation. Specifically, various topics are involved in this environment for comprehensive interviewing evaluation, resulting in around 1.4k turns of utterances, 184k tokens, and over 200 events mentioned during the interviewing for each chatbot evaluation. We compare GuideLLM with 6 state-of-the-art LLMs such as GPT-4o and Llama-3-70b-Instruct, from the perspective of interviewing quality, and autobiography generation quality. For automatic evaluation, we derive user proxies from multiple autobiographies and employ LLM-as-a-judge to score LLM behaviors. We further conduct a human-involved experiment by employing 45 human participants to chat with GuideLLM and baselines. We then collect human feedback, preferences, and ratings regarding the qualities of conversation and autobiography. Experimental results indicate that GuideLLM significantly outperforms baseline LLMs in automatic evaluation and achieves consistent leading performances in human ratings.
2502.06498
Decision Boundary Optimization-Informed Domain Adaptation
cs.CV
Maximum Mean Discrepancy (MMD) is widely used in a number of domain adaptation (DA) methods and shows its effectiveness in aligning data distributions across domains. However, in previous DA research, MMD-based DA methods focus mostly on distribution alignment, and ignore to optimize the decision boundary for classification-aware DA, thereby falling short in reducing the DA upper error bound. In this paper, we propose a strengthened MMD measurement, namely, Decision Boundary optimization-informed MMD (DB-MMD), which enables MMD to carefully take into account the decision boundaries, thereby simultaneously optimizing the distribution alignment and cross-domain classifier within a hybrid framework, and leading to a theoretical bound guided DA. We further seamlessly embed the proposed DB-MMD measurement into several popular DA methods, e.g., MEDA, DGA-DA, to demonstrate its effectiveness w.r.t different experimental settings. We carry out comprehensive experiments using 8 standard DA datasets. The experimental results show that the DB-MMD enforced DA methods improve their baseline models using plain vanilla MMD, with a margin that can be as high as 9.5.
2502.06501
Learning Clustering-based Prototypes for Compositional Zero-shot Learning
cs.CV
Learning primitive (i.e., attribute and object) concepts from seen compositions is the primary challenge of Compositional Zero-Shot Learning (CZSL). Existing CZSL solutions typically rely on oversimplified data assumptions, e.g., modeling each primitive with a single centroid primitive representation, ignoring the natural diversities of the attribute (resp. object) when coupled with different objects (resp. attribute). In this work, we develop ClusPro, a robust clustering-based prototype mining framework for CZSL that defines the conceptual boundaries of primitives through a set of diversified prototypes. Specifically, ClusPro conducts within-primitive clustering on the embedding space for automatically discovering and dynamically updating prototypes. These representative prototypes are subsequently used to repaint a well-structured and independent primitive embedding space, ensuring intra-primitive separation and inter-primitive decorrelation through prototype-based contrastive learning and decorrelation learning. Moreover, ClusPro efficiently performs prototype clustering in a non-parametric fashion without the introduction of additional learnable parameters or computational budget during testing. Experiments on three benchmarks demonstrate ClusPro outperforms various top-leading CZSL solutions under both closed-world and open-world settings.
2502.06516
Boost-and-Skip: A Simple Guidance-Free Diffusion for Minority Generation
cs.LG cs.AI cs.CV stat.ML
Minority samples are underrepresented instances located in low-density regions of a data manifold, and are valuable in many generative AI applications, such as data augmentation, creative content generation, etc. Unfortunately, existing diffusion-based minority generators often rely on computationally expensive guidance dedicated for minority generation. To address this, here we present a simple yet powerful guidance-free approach called Boost-and-Skip for generating minority samples using diffusion models. The key advantage of our framework requires only two minimal changes to standard generative processes: (i) variance-boosted initialization and (ii) timestep skipping. We highlight that these seemingly-trivial modifications are supported by solid theoretical and empirical evidence, thereby effectively promoting emergence of underrepresented minority features. Our comprehensive experiments demonstrate that Boost-and-Skip greatly enhances the capability of generating minority samples, even rivaling guidance-based state-of-the-art approaches while requiring significantly fewer computations.
2502.06519
SIREN: Semantic, Initialization-Free Registration of Multi-Robot Gaussian Splatting Maps
cs.RO cs.CV
We present SIREN for registration of multi-robot Gaussian Splatting (GSplat) maps, with zero access to camera poses, images, and inter-map transforms for initialization or fusion of local submaps. To realize these capabilities, SIREN harnesses the versatility and robustness of semantics in three critical ways to derive a rigorous registration pipeline for multi-robot GSplat maps. First, SIREN utilizes semantics to identify feature-rich regions of the local maps where the registration problem is better posed, eliminating the need for any initialization which is generally required in prior work. Second, SIREN identifies candidate correspondences between Gaussians in the local maps using robust semantic features, constituting the foundation for robust geometric optimization, coarsely aligning 3D Gaussian primitives extracted from the local maps. Third, this key step enables subsequent photometric refinement of the transformation between the submaps, where SIREN leverages novel-view synthesis in GSplat maps along with a semantics-based image filter to compute a high-accuracy non-rigid transformation for the generation of a high-fidelity fused map. We demonstrate the superior performance of SIREN compared to competing baselines across a range of real-world datasets, and in particular, across the most widely-used robot hardware platforms, including a manipulator, drone, and quadruped. In our experiments, SIREN achieves about 90x smaller rotation errors, 300x smaller translation errors, and 44x smaller scale errors in the most challenging scenes, where competing methods struggle. We will release the code and provide a link to the project page after the review process.
2502.06523
Tighter Value-Function Approximations for POMDPs
cs.AI
Solving partially observable Markov decision processes (POMDPs) typically requires reasoning about the values of exponentially many state beliefs. Towards practical performance, state-of-the-art solvers use value bounds to guide this reasoning. However, sound upper value bounds are often computationally expensive to compute, and there is a tradeoff between the tightness of such bounds and their computational cost. This paper introduces new and provably tighter upper value bounds than the commonly used fast informed bound. Our empirical evaluation shows that, despite their additional computational overhead, the new upper bounds accelerate state-of-the-art POMDP solvers on a wide range of benchmarks.
2502.06525
Properties of Wasserstein Gradient Flows for the Sliced-Wasserstein Distance
stat.ML cs.LG
In this paper, we investigate the properties of the Sliced Wasserstein Distance (SW) when employed as an objective functional. The SW metric has gained significant interest in the optimal transport and machine learning literature, due to its ability to capture intricate geometric properties of probability distributions while remaining computationally tractable, making it a valuable tool for various applications, including generative modeling and domain adaptation. Our study aims to provide a rigorous analysis of the critical points arising from the optimization of the SW objective. By computing explicit perturbations, we establish that stable critical points of SW cannot concentrate on segments. This stability analysis is crucial for understanding the behaviour of optimization algorithms for models trained using the SW objective. Furthermore, we investigate the properties of the SW objective, shedding light on the existence and convergence behavior of critical points. We illustrate our theoretical results through numerical experiments.
2502.06526
Convex Split Lemma without Inequalities
quant-ph cs.IT math-ph math.IT math.MP
We introduce a refinement to the convex split lemma by replacing the max mutual information with the collision mutual information, transforming the inequality into an equality. This refinement yields tighter achievability bounds for quantum source coding tasks, including state merging and state splitting. Furthermore, we derive a universal upper bound on the smoothed max mutual information, where "universal" signifies that the bound depends exclusively on R\'enyi entropies and is independent of the system's dimensions. This result has significant implications for quantum information processing, particularly in applications such as the reverse quantum Shannon theorem.
2502.06527
CustomVideoX: 3D Reference Attention Driven Dynamic Adaptation for Zero-Shot Customized Video Diffusion Transformers
cs.CV
Customized generation has achieved significant progress in image synthesis, yet personalized video generation remains challenging due to temporal inconsistencies and quality degradation. In this paper, we introduce CustomVideoX, an innovative framework leveraging the video diffusion transformer for personalized video generation from a reference image. CustomVideoX capitalizes on pre-trained video networks by exclusively training the LoRA parameters to extract reference features, ensuring both efficiency and adaptability. To facilitate seamless interaction between the reference image and video content, we propose 3D Reference Attention, which enables direct and simultaneous engagement of reference image features with all video frames across spatial and temporal dimensions. To mitigate the excessive influence of reference image features and textual guidance on generated video content during inference, we implement the Time-Aware Reference Attention Bias (TAB) strategy, dynamically modulating reference bias over different time steps. Additionally, we introduce the Entity Region-Aware Enhancement (ERAE) module, aligning highly activated regions of key entity tokens with reference feature injection by adjusting attention bias. To thoroughly evaluate personalized video generation, we establish a new benchmark, VideoBench, comprising over 50 objects and 100 prompts for extensive assessment. Experimental results show that CustomVideoX significantly outperforms existing methods in terms of video consistency and quality.
2502.06533
Ignore the KL Penalty! Boosting Exploration on Critical Tokens to Enhance RL Fine-Tuning
cs.CL cs.LG
The ability to achieve long-term goals is a key challenge in the current development of large language models (LLMs). To address this, pre-trained LLMs can be fine-tuned with reinforcement learning (RL) to explore solutions that optimize a given goal. However, exploration with LLMs is difficult, as a balance has to be struck between discovering new solutions and staying close enough to the pre-trained model, so as not to degrade basic capabilities. This is typically controlled with a Kullback-Leibler (KL) penalty. In this paper, we investigate the exploration dynamics of a small language model on a simple arithmetic task. We show how varying degrees of pre-training influence exploration and demonstrate the importance of "critical tokens" which have a dramatic impact on the final outcome. Consequently, we introduce a simple modification to the KL penalty that favors exploration on critical tokens, increasing the efficiency of the RL fine-tuning stage.
2502.06536
Sample-efficient Learning of Concepts with Theoretical Guarantees: from Data to Concepts without Interventions
stat.ML cs.LG
Machine learning is a vital part of many real-world systems, but several concerns remain about the lack of interpretability, explainability and robustness of black-box AI systems. Concept-based models (CBM) address some of these challenges by learning interpretable concepts from high-dimensional data, e.g. images, which are used to predict labels. An important issue in CBMs is concept leakage, i.e., spurious information in the learned concepts, which effectively leads to learning "wrong" concepts. Current mitigating strategies are heuristic, have strong assumptions, e.g., they assume that the concepts are statistically independent of each other, or require substantial human interaction in terms of both interventions and labels provided by annotators. In this paper, we describe a framework that provides theoretical guarantees on the correctness of the learned concepts and on the number of required labels, without requiring any interventions. Our framework leverages causal representation learning (CRL) to learn high-level causal variables from low-level data, and learns to align these variables with interpretable concepts. We propose a linear and a non-parametric estimator for this mapping, providing a finite-sample high probability result in the linear case and an asymptotic consistency result for the non-parametric estimator. We implement our framework with state-of-the-art CRL methods, and show its efficacy in learning the correct concepts in synthetic and image benchmarks.
2502.06543
Unsupervised Learning for Feature Extraction and Temporal Alignment of 3D+t Point Clouds of Zebrafish Embryos
cs.CV
Zebrafish are widely used in biomedical research and developmental stages of their embryos often need to be synchronized for further analysis. We present an unsupervised approach to extract descriptive features from 3D+t point clouds of zebrafish embryos and subsequently use those features to temporally align corresponding developmental stages. An autoencoder architecture is proposed to learn a descriptive representation of the point clouds and we designed a deep regression network for their temporal alignment. We achieve a high alignment accuracy with an average mismatch of only 3.83 minutes over an experimental duration of 5.3 hours. As a fully-unsupervised approach, there is no manual labeling effort required and unlike manual analyses the method easily scales. Besides, the alignment without human annotation of the data also avoids any influence caused by subjective bias.
2502.06544
Sequence Transferability and Task Order Selection in Continual Learning
cs.LG cs.CV
In continual learning, understanding the properties of task sequences and their relationships to model performance is important for developing advanced algorithms with better accuracy. However, efforts in this direction remain underdeveloped despite encouraging progress in methodology development. In this work, we investigate the impacts of sequence transferability on continual learning and propose two novel measures that capture the total transferability of a task sequence, either in the forward or backward direction. Based on the empirical properties of these measures, we then develop a new method for the task order selection problem in continual learning. Our method can be shown to offer a better performance than the conventional strategy of random task selection.
2502.06545
Dimension-free Regret for Learning Asymmetric Linear Dynamical Systems
cs.LG stat.ML
Previously, methods for learning marginally stable linear dynamical systems either required the transition matrix to be symmetric or incurred regret bounds that scale polynomially with the system's hidden dimension. In this work, we introduce a novel method that overcomes this trade-off, achieving dimension-free regret despite the presence of asymmetric matrices and marginal stability. Our method combines spectral filtering with linear predictors and employs Chebyshev polynomials in the complex plane to construct a novel spectral filtering basis. This construction guarantees sublinear regret in an online learning framework, without relying on any statistical or generative assumptions. Specifically, we prove that as long as the transition matrix has eigenvalues with complex component bounded by $1/\mathrm{poly} \log T$, then our method achieves regret $\tilde{O}(T^{9/10})$ when compared to the best linear dynamical predictor in hindsight.
2502.06547
Data Augmentation and Regularization for Learning Group Equivariance
stat.ML cs.LG math.OC
In many machine learning tasks, known symmetries can be used as an inductive bias to improve model performance. In this paper, we consider learning group equivariance through training with data augmentation. We summarize results from a previous paper of our own, and extend the results to show that equivariance of the trained model can be achieved through training on augmented data in tandem with regularization.
2502.06551
Efficient Scientific Full Text Classification: The Case of EICAT Impact Assessments
cs.CL
This study explores strategies for efficiently classifying scientific full texts using both small, BERT-based models and local large language models like Llama-3.1 8B. We focus on developing methods for selecting subsets of input sentences to reduce input size while simultaneously enhancing classification performance. To this end, we compile a novel dataset consisting of full-text scientific papers from the field of invasion biology, specifically addressing the impacts of invasive species. These papers are aligned with publicly available impact assessments created by researchers for the International Union for Conservation of Nature (IUCN). Through extensive experimentation, we demonstrate that various sources like human evidence annotations, LLM-generated annotations or explainability scores can be used to train sentence selection models that improve the performance of both encoder- and decoder-based language models while optimizing efficiency through the reduction in input length, leading to improved results even if compared to models like ModernBERT that are able to handle the complete text as input. Additionally, we find that repeated sampling of shorter inputs proves to be a very effective strategy that, at a slightly increased cost, can further improve classification performance.
2502.06552
Diffusion Models for Computational Neuroimaging: A Survey
cs.CV
Computational neuroimaging involves analyzing brain images or signals to provide mechanistic insights and predictive tools for human cognition and behavior. While diffusion models have shown stability and high-quality generation in natural images, there is increasing interest in adapting them to analyze brain data for various neurological tasks such as data enhancement, disease diagnosis and brain decoding. This survey provides an overview of recent efforts to integrate diffusion models into computational neuroimaging. We begin by introducing the common neuroimaging data modalities, follow with the diffusion formulations and conditioning mechanisms. Then we discuss how the variations of the denoising starting point, condition input and generation target of diffusion models are developed and enhance specific neuroimaging tasks. For a comprehensive overview of the ongoing research, we provide a publicly available repository at https://github.com/JoeZhao527/dm4neuro.
2502.06555
Is API Access to LLMs Useful for Generating Private Synthetic Tabular Data?
cs.LG cs.CR
Differentially private (DP) synthetic data is a versatile tool for enabling the analysis of private data. Recent advancements in large language models (LLMs) have inspired a number of algorithm techniques for improving DP synthetic data generation. One family of approaches uses DP finetuning on the foundation model weights; however, the model weights for state-of-the-art models may not be public. In this work we propose two DP synthetic tabular data algorithms that only require API access to the foundation model. We adapt the Private Evolution algorithm (Lin et al., 2023; Xie et al., 2024) -- which was designed for image and text data -- to the tabular data domain. In our extension of Private Evolution, we define a query workload-based distance measure, which may be of independent interest. We propose a family of algorithms that use one-shot API access to LLMs, rather than adaptive queries to the LLM. Our findings reveal that API-access to powerful LLMs does not always improve the quality of DP synthetic data compared to established baselines that operate without such access. We provide insights into the underlying reasons and propose improvements to LLMs that could make them more effective for this application.
2502.06556
ProjectTest: A Project-level LLM Unit Test Generation Benchmark and Impact of Error Fixing Mechanisms
cs.SE cs.CL
Unit test generation has become a promising and important use case of LLMs. However, existing evaluation benchmarks for assessing LLM unit test generation capabilities focus on function- or class-level code rather than more practical and challenging project-level codebases. To address such limitation, we propose ProjectTest, a project-level benchmark for unit test generation covering Python, Java, and JavaScript. ProjectTest features 20 moderate-sized and high-quality projects per language. We evaluate nine frontier LLMs on ProjectTest and the results show that all frontier LLMs tested exhibit moderate performance on ProjectTest on Python and Java, highlighting the difficulty of ProjectTest. We also conduct a thorough error analysis, which shows that even frontier LLMs, such as Claude-3.5-Sonnet, have significant basic yet critical errors, including compilation and cascade errors. Motivated by this observation, we further evaluate all frontier LLMs under manual error-fixing and self-error-fixing scenarios to assess their potential when equipped with error-fixing mechanisms. Our code and dataset is available at \href{https://github.com/YiboWANG214/ProjectTest}{ProjectTest}.
2502.06557
LiveForesighter: Generating Future Information for Live-Streaming Recommendations at Kuaishou
cs.IR
Live-streaming, as a new-generation media to connect users and authors, has attracted a lot of attention and experienced rapid growth in recent years. Compared with the content-static short-video recommendation, the live-streaming recommendation faces more challenges in giving our users a satisfactory experience: (1) Live-streaming content is dynamically ever-changing along time. (2) valuable behaviors (e.g., send digital-gift, buy products) always require users to watch for a long-time (>10 min). Combining the two attributes, here raising a challenging question for live-streaming recommendation: How to discover the live-streamings that the content user is interested in at the current moment, and further a period in the future?
2502.06559
Can We Trust AI Benchmarks? An Interdisciplinary Review of Current Issues in AI Evaluation
cs.AI
Quantitative Artificial Intelligence (AI) Benchmarks have emerged as fundamental tools for evaluating the performance, capability, and safety of AI models and systems. Currently, they shape the direction of AI development and are playing an increasingly prominent role in regulatory frameworks. As their influence grows, however, so too does concerns about how and with what effects they evaluate highly sensitive topics such as capabilities, including high-impact capabilities, safety and systemic risks. This paper presents an interdisciplinary meta-review of about 100 studies that discuss shortcomings in quantitative benchmarking practices, published in the last 10 years. It brings together many fine-grained issues in the design and application of benchmarks (such as biases in dataset creation, inadequate documentation, data contamination, and failures to distinguish signal from noise) with broader sociotechnical issues (such as an over-focus on evaluating text-based AI models according to one-time testing logic that fails to account for how AI models are increasingly multimodal and interact with humans and other technical systems). Our review also highlights a series of systemic flaws in current benchmarking practices, such as misaligned incentives, construct validity issues, unknown unknowns, and problems with the gaming of benchmark results. Furthermore, it underscores how benchmark practices are fundamentally shaped by cultural, commercial and competitive dynamics that often prioritise state-of-the-art performance at the expense of broader societal concerns. By providing an overview of risks associated with existing benchmarking procedures, we problematise disproportionate trust placed in benchmarks and contribute to ongoing efforts to improve the accountability and relevance of quantitative AI benchmarks within the complexities of real-world scenarios.
2502.06560
Position: It's Time to Act on the Risk of Efficient Personalized Text Generation
cs.CL cs.CY
The recent surge in high-quality open-sourced Generative AI text models (colloquially: LLMs), as well as efficient finetuning techniques, has opened the possibility of creating high-quality personalized models, i.e., models generating text attuned to a specific individual's needs and capable of credibly imitating their writing style by leveraging that person's own data to refine an open-source model. The technology to create such models is accessible to private individuals, and training and running such models can be done cheaply on consumer-grade hardware. These advancements are a huge gain for usability and privacy. This position paper argues, however, that these advancements also introduce new safety risks by making it practically feasible for malicious actors to impersonate specific individuals at scale, for instance for the purpose of phishing emails, based on small amounts of publicly available text. We further argue that these risks are complementary to - and distinct from - the much-discussed risks of other impersonation attacks such as image, voice, or video deepfakes, and are not adequately addressed by the larger research community, or the current generation of open - and closed-source models.
2502.06563
Large Language Models Meet Symbolic Provers for Logical Reasoning Evaluation
cs.CL
First-order logic (FOL) reasoning, which involves sequential deduction, is pivotal for intelligent systems and serves as a valuable task for evaluating reasoning capabilities, particularly in chain-of-thought (CoT) contexts. Existing benchmarks often rely on extensive human annotation or handcrafted templates, making it difficult to achieve the necessary complexity, scalability, and diversity for robust evaluation. To address these limitations, we propose a novel framework called ProverGen that synergizes the generative strengths of Large Language Models (LLMs) with the rigor and precision of symbolic provers, enabling the creation of a scalable, diverse, and high-quality FOL reasoning dataset, ProverQA. ProverQA is also distinguished by its inclusion of accessible and logically coherent intermediate reasoning steps for each problem. Our evaluation shows that state-of-the-art LLMs struggle to solve ProverQA problems, even with CoT prompting, highlighting the dataset's challenging nature. We also finetune Llama3.1-8B-Instruct on a separate training set generated by our framework. The finetuned model demonstrates consistent improvements on both in-distribution and out-of-distribution test sets, suggesting the value of our proposed data generation framework. Code available at: https://github.com/opendatalab/ProverGen
2502.06564
Robust Scatter Matrix Estimation for Elliptical Distributions in Polynomial Time
cs.DS cs.LG math.ST stat.ML stat.TH
We study the problem of computationally efficient robust estimation of scatter matrices of elliptical distributions under the strong contamination model. We design polynomial time algorithms that achieve dimension-independent error in Frobenius norm. Our first result is a sequence of efficient algorithms that approaches nearly optimal error. Specifically, under a mild assumption on the eigenvalues of the scatter matrix $\Sigma$, for every $t \in \mathbb{N}$, we design an estimator that, given $n = d^{O(t)}$ samples, in time $n^{O(t)}$ finds $\hat{\Sigma}$ such that $ \Vert{\Sigma^{-1/2}\, ({\hat{\Sigma} - \Sigma})\, \Sigma^{-1/2}}\Vert_{\text{F}} \le O(t \cdot \varepsilon^{1-\frac{1}{t}})$, where $\varepsilon$ is the fraction of corruption. We do not require any assumptions on the moments of the distribution, while all previously known computationally efficient algorithms for robust covariance/scatter estimation with dimension-independent error rely on strong assumptions on the moments, such as sub-Gaussianity or (certifiable) hypercontractivity. Furthermore, under a stronger assumption on the eigenvalues of $\Sigma$ (that, in particular, is satisfied by all matrices with constant condition number), we provide a fast (sub-quadratic in the input size) algorithm that, given nearly optimal number of samples $n = \tilde{O}(d^2/\varepsilon)$, in time $\tilde{O}({nd^2 poly(1/\varepsilon)})$ finds $\hat{\Sigma}$ such that $\Vert\hat{\Sigma} - \Sigma\Vert_{\text{F}} \le O(\Vert{\Sigma}\Vert \cdot \sqrt{\varepsilon})$. Our approach is based on robust covariance estimation of the spatial sign (the projection onto the sphere of radius $\sqrt{d}$) of elliptical distributions.
2502.06567
Membership Inference Risks in Quantized Models: A Theoretical and Empirical Study
stat.ML cs.LG
Quantizing machine learning models has demonstrated its effectiveness in lowering memory and inference costs while maintaining performance levels comparable to the original models. In this work, we investigate the impact of quantization procedures on the privacy of data-driven models, specifically focusing on their vulnerability to membership inference attacks. We derive an asymptotic theoretical analysis of Membership Inference Security (MIS), characterizing the privacy implications of quantized algorithm weights against the most powerful (and possibly unknown) attacks. Building on these theoretical insights, we propose a novel methodology to empirically assess and rank the privacy levels of various quantization procedures. Using synthetic datasets, we demonstrate the effectiveness of our approach in assessing the MIS of different quantizers. Furthermore, we explore the trade-off between privacy and performance using real-world data and models in the context of molecular modeling.
2502.06572
LawGPT: Knowledge-Guided Data Generation and Its Application to Legal LLM
cs.CL cs.AI
Large language models (LLMs), both proprietary and open-source, have demonstrated remarkable capabilities across various natural language processing tasks. However, they face significant limitations in legal reasoning tasks. Proprietary models introduce data privacy risks and high inference costs, while open-source models underperform due to insufficient legal domain training data. To address these limitations, we study data generation for legal reasoning to improve the legal reasoning performance of open-source LLMs with the help of proprietary LLMs. This is challenging due to the lack of legal knowledge in proprietary LLMs and the difficulty in verifying the generated data. We propose KgDG, a knowledge-guided data generation framework for legal reasoning. Our framework enables leveraging legal knowledge to enhance generation diversity and introduces a refinement and verification process to ensure the quality of generated data. Moreover, we expand the generated dataset to further enhance the LLM reasoning capabilities. Using KgDG, we create a synthetic legal reasoning dataset containing 50K high-quality examples. Our trained model LawGPT outperforms existing legal-specific LLMs and achieves performance comparable to proprietary LLMs, demonstrating the effectiveness of KgDG and LawGPT. Our code and resources is publicly available at https://github.com/LAMDASZ-ML/Knowledge-Guide-Data-Generation .
2502.06574
On the Impact of the Utility in Semivalue-based Data Valuation
cs.AI cs.GT cs.LG
Semivalue-based data valuation in machine learning (ML) quantifies the contribution of individual data points to a downstream ML task by leveraging principles from cooperative game theory and the notion of utility. While this framework has been used in practice for assessing data quality, our experiments reveal inconsistent valuation outcomes across different utilities, albeit all related to ML performance. Beyond raising concerns about the reliability of data valuation, this inconsistency is challenging to interpret, as it stems from the complex interaction of the utility with data points and semivalue weights, which has barely been studied in prior work. In this paper, we take a first step toward clarifying the utility impact on semivalue-based data valuation. Specifically, we provide geometric interpretations of this impact for a broad family of classification utilities, which includes the accuracy and the arithmetic mean. We introduce the notion of spatial signatures: given a semivalue, data points can be embedded into a two-dimensional space, and utility functions map to the dual of this space. This geometric perspective separates the influence of the dataset and semivalue from that of the utility, providing a theoretical explanation for the experimentally observed sensitivity of valuation outcomes to the utility choice.
2502.06575
Predictive Red Teaming: Breaking Policies Without Breaking Robots
cs.RO cs.AI cs.LG cs.SY eess.SY
Visuomotor policies trained via imitation learning are capable of performing challenging manipulation tasks, but are often extremely brittle to lighting, visual distractors, and object locations. These vulnerabilities can depend unpredictably on the specifics of training, and are challenging to expose without time-consuming and expensive hardware evaluations. We propose the problem of predictive red teaming: discovering vulnerabilities of a policy with respect to environmental factors, and predicting the corresponding performance degradation without hardware evaluations in off-nominal scenarios. In order to achieve this, we develop RoboART: an automated red teaming (ART) pipeline that (1) modifies nominal observations using generative image editing to vary different environmental factors, and (2) predicts performance under each variation using a policy-specific anomaly detector executed on edited observations. Experiments across 500+ hardware trials in twelve off-nominal conditions for visuomotor diffusion policies demonstrate that RoboART predicts performance degradation with high accuracy (less than 0.19 average difference between predicted and real success rates). We also demonstrate how predictive red teaming enables targeted data collection: fine-tuning with data collected under conditions predicted to be adverse boosts baseline performance by 2-7x.
2502.06577
The Minimal Search Space for Conditional Causal Bandits
cs.LG cs.AI stat.ML
Causal knowledge can be used to support decision-making problems. This has been recognized in the causal bandits literature, where a causal (multi-armed) bandit is characterized by a causal graphical model and a target variable. The arms are then interventions on the causal model, and rewards are samples of the target variable. Causal bandits were originally studied with a focus on hard interventions. We focus instead on cases where the arms are conditional interventions, which more accurately model many real-world decision-making problems by allowing the value of the intervened variable to be chosen based on the observed values of other variables. This paper presents a graphical characterization of the minimal set of nodes guaranteed to contain the optimal conditional intervention, which maximizes the expected reward. We then propose an efficient algorithm with a time complexity of $O(|V| + |E|)$ to identify this minimal set of nodes. We prove that the graphical characterization and the proposed algorithm are correct. Finally, we empirically demonstrate that our algorithm significantly prunes the search space and substantially accelerates convergence rates when integrated into standard multi-armed bandit algorithms.
2502.06580
Inventory Consensus Control in Supply Chain Networks using Dissipativity-Based Control and Topology Co-Design
eess.SY cs.SY
Recent global and local phenomena have exposed vulnerabilities in critical supply chain networks (SCNs), drawing significant attention from researchers across various fields. Typically, SCNs are viewed as static entities regularly optimized to maintain their optimal operation. However, the dynamic nature of SCNs and their associated uncertainties have motivated researchers to treat SCNs as dynamic networked systems requiring robust control techniques. In this paper, we address the SCN inventory consensus problem, which aims to synchronize multiple parallel supply chains, enhancing coordination and robustness of the overall SCN. To achieve this, we take a novel approach exploiting dissipativity theory. In particular, we propose a dissipativity-based co-design strategy for distributed consensus controllers and communication topology in SCNs. It requires only the dissipativity information of the individual supply chains and involves solving a set of convex optimization problems, thus contributing to scalability, compositionality, and computational efficiency. Moreover, it optimizes the robustness of the SCN to various associated uncertainties, mitigating both bullwhip and ripple effects. We demonstrate our contributions using numerical examples, mainly by comparing the consensus performance with respect to standard steady-state control, feedback control, and consensus control strategies.
2502.06581
A Survey on Video Analytics in Cloud-Edge-Terminal Collaborative Systems
cs.NI cs.CV cs.LG
The explosive growth of video data has driven the development of distributed video analytics in cloud-edge-terminal collaborative (CETC) systems, enabling efficient video processing, real-time inference, and privacy-preserving analysis. Among multiple advantages, CETC systems can distribute video processing tasks and enable adaptive analytics across cloud, edge, and terminal devices, leading to breakthroughs in video surveillance, autonomous driving, and smart cities. In this survey, we first analyze fundamental architectural components, including hierarchical, distributed, and hybrid frameworks, alongside edge computing platforms and resource management mechanisms. Building upon these foundations, edge-centric approaches emphasize on-device processing, edge-assisted offloading, and edge intelligence, while cloud-centric methods leverage powerful computational capabilities for complex video understanding and model training. Our investigation also covers hybrid video analytics incorporating adaptive task offloading and resource-aware scheduling techniques that optimize performance across the entire system. Beyond conventional approaches, recent advances in large language models and multimodal integration reveal both opportunities and challenges in platform scalability, data protection, and system reliability. Future directions also encompass explainable systems, efficient processing mechanisms, and advanced video analytics, offering valuable insights for researchers and practitioners in this dynamic field.
2502.06583
Adaptive Perception for Unified Visual Multi-modal Object Tracking
cs.CV
Recently, many multi-modal trackers prioritize RGB as the dominant modality, treating other modalities as auxiliary, and fine-tuning separately various multi-modal tasks. This imbalance in modality dependence limits the ability of methods to dynamically utilize complementary information from each modality in complex scenarios, making it challenging to fully perceive the advantages of multi-modal. As a result, a unified parameter model often underperforms in various multi-modal tracking tasks. To address this issue, we propose APTrack, a novel unified tracker designed for multi-modal adaptive perception. Unlike previous methods, APTrack explores a unified representation through an equal modeling strategy. This strategy allows the model to dynamically adapt to various modalities and tasks without requiring additional fine-tuning between different tasks. Moreover, our tracker integrates an adaptive modality interaction (AMI) module that efficiently bridges cross-modality interactions by generating learnable tokens. Experiments conducted on five diverse multi-modal datasets (RGBT234, LasHeR, VisEvent, DepthTrack, and VOT-RGBD2022) demonstrate that APTrack not only surpasses existing state-of-the-art unified multi-modal trackers but also outperforms trackers designed for specific multi-modal tasks.
2502.06584
Deep Reinforcement Learning based Triggering Function for Early Classifiers of Time Series
cs.LG
Early Classification of Time Series (ECTS) has been recognized as an important problem in many areas where decisions have to be taken as soon as possible, before the full data availability, while time pressure increases. Numerous ECTS approaches have been proposed, based on different triggering functions, each taking into account various pieces of information related to the incoming time series and/or the output of a classifier. Although their performances have been empirically compared in the literature, no studies have been carried out on the optimality of these triggering functions that involve ``man-tailored'' decision rules. Based on the same information, could there be better triggering functions? This paper presents one way to investigate this question by showing first how to translate ECTS problems into Reinforcement Learning (RL) ones, where the very same information is used in the state space. A thorough comparison of the performance obtained by ``handmade'' approaches and their ``RL-based'' counterparts has been carried out. A second question investigated in this paper is whether a different combination of information, defining the state space in RL systems, can achieve even better performance. Experiments show that the system we describe, called \textsc{Alert}, significantly outperforms its state-of-the-art competitors on a large number of datasets.
2502.06585
Extract-QD Framework: A Generic Approach for Quality-Diversity in Noisy, Stochastic or Uncertain Domains
cs.NE
Quality-Diversity (QD) has demonstrated potential in discovering collections of diverse solutions to optimisation problems. Originally designed for deterministic environments, QD has been extended to noisy, stochastic, or uncertain domains through various Uncertain-QD (UQD) methods. However, the large number of UQD methods, each with unique constraints, makes selecting the most suitable one challenging. To remedy this situation, we present two contributions: first, the Extract-QD Framework (EQD Framework), and second, Extract-ME (EME), a new method derived from it. The EQD Framework unifies existing approaches within a modular view, and facilitates developing novel methods by interchanging modules. We use it to derive EME, a novel method that consistently outperforms or matches the best existing methods on standard benchmarks, while previous methods show varying performance. In a second experiment, we show how our EQD Framework can be used to augment existing QD algorithms and in particular the well-established Policy-Gradient-Assisted-MAP-Elites method, and demonstrate improved performance in uncertain domains at no additional evaluation cost. For any new uncertain task, our contributions now provide EME as a reliable "first guess" method, and the EQD Framework as a tool for developing task-specific approaches. Together, these contributions aim to lower the cost of adopting UQD insights in QD applications.
2502.06587
evclust: Python library for evidential clustering
cs.SE cs.CV cs.LG
A recent developing trend in clustering is the advancement of algorithms that not only identify clusters within data, but also express and capture the uncertainty of cluster membership. Evidential clustering addresses this by using the Dempster-Shafer theory of belief functions, a framework designed to manage and represent uncertainty. This approach results in a credal partition, a structured set of mass functions that quantify the uncertain assignment of each object to potential groups. The Python framework evclust, presented in this paper, offers a suite of efficient evidence clustering algorithms as well as tools for visualizing, evaluating and analyzing credal partitions.
2502.06589
Hephaestus: Improving Fundamental Agent Capabilities of Large Language Models through Continual Pre-Training
cs.CL cs.AI cs.LG
Due to the scarcity of agent-oriented pre-training data, LLM-based autonomous agents typically rely on complex prompting or extensive fine-tuning, which often fails to introduce new capabilities while preserving strong generalizability. We introduce Hephaestus-Forge, the first large-scale pre-training corpus designed to enhance the fundamental capabilities of LLM agents in API function calling, intrinsic reasoning and planning, and adapting to environmental feedback. Hephaestus-Forge comprises 103B agent-specific data encompassing 76,537 APIs, including both tool documentation to introduce knowledge of API functions and function calling trajectories to strengthen intrinsic reasoning. To explore effective training protocols, we investigate scaling laws to identify the optimal recipe in data mixing ratios. By continual pre-training on Hephaestus-Forge, Hephaestus outperforms small- to medium-scale open-source LLMs and rivals commercial LLMs on three agent benchmarks, demonstrating the effectiveness of our pre-training corpus in enhancing fundamental agentic capabilities and generalization of LLMs to new tasks or environments.
2502.06591
Diffeomorphic Temporal Alignment Nets for Time-series Joint Alignment and Averaging
cs.LG
In time-series analysis, nonlinear temporal misalignment remains a pivotal challenge that forestalls even simple averaging. Since its introduction, the Diffeomorphic Temporal Alignment Net (DTAN), which we first introduced (Weber et al., 2019) and further developed in (Weber & Freifeld, 2023), has proven itself as an effective solution for this problem (these conference papers are earlier partial versions of the current manuscript). DTAN predicts and applies diffeomorphic transformations in an input-dependent manner, thus facilitating the joint alignment (JA) and averaging of time-series ensembles in an unsupervised or a weakly-supervised manner. The inherent challenges of the weakly/unsupervised setting, particularly the risk of trivial solutions through excessive signal distortion, are mitigated using either one of two distinct strategies: 1) a regularization term for warps; 2) using the Inverse Consistency Averaging Error (ICAE). The latter is a novel, regularization-free approach which also facilitates the JA of variable-length signals. We also further extend our framework to incorporate multi-task learning (MT-DTAN), enabling simultaneous time-series alignment and classification. Additionally, we conduct a comprehensive evaluation of different backbone architectures, demonstrating their efficacy in time-series alignment tasks. Finally, we showcase the utility of our approach in enabling Principal Component Analysis (PCA) for misaligned time-series data. Extensive experiments across 128 UCR datasets validate the superiority of our approach over contemporary averaging methods, including both traditional and learning-based approaches, marking a significant advancement in the field of time-series analysis.
2502.06593
A Large-scale AI-generated Image Inpainting Benchmark
cs.CV
Recent advances in generative models enable highly realistic image manipulations, creating an urgent need for robust forgery detection methods. Current datasets for training and evaluating these methods are limited in scale and diversity. To address this, we propose a methodology for creating high-quality inpainting datasets and apply it to create DiQuID, comprising over 95,000 inpainted images generated from 78,000 original images sourced from MS-COCO, RAISE, and OpenImages. Our methodology consists of three components: (1) Semantically Aligned Object Replacement (SAOR) that identifies suitable objects through instance segmentation and generates contextually appropriate prompts, (2) Multiple Model Image Inpainting (MMII) that employs various state-of-the-art inpainting pipelines primarily based on diffusion models to create diverse manipulations, and (3) Uncertainty-Guided Deceptiveness Assessment (UGDA) that evaluates image realism through comparative analysis with originals. The resulting dataset surpasses existing ones in diversity, aesthetic quality, and technical quality. We provide comprehensive benchmarking results using state-of-the-art forgery detection methods, demonstrating the dataset's effectiveness in evaluating and improving detection algorithms. Through a human study with 42 participants on 1,000 images, we show that while humans struggle with images classified as deceiving by our methodology, models trained on our dataset maintain high performance on these challenging cases. Code and dataset are available at https://github.com/mever-team/DiQuID.
2502.06594
A Review of Conceptualizations of Safety and Risk in Current Automated Driving Regulation
eess.SY cs.SY
"Safety" and "Risk" are key concepts for the design and development of automated vehicles. For the market introduction or large-scale field tests, both concepts are not only relevant for engineers developing the vehicles, but for all stakeholders (e.g., regulators, lawyers, or the general public) who have stakes in the technology. In the communication between stakeholder groups, common notions of these abstract concepts are key for efficient communication and setting mutual expectations. In the European market, automated vehicles require Europe-wide type approval or at least operating permits in the individual states. For this, a central means of communication between regulators and engineers are regulatory documents. Flawed terminology regarding the safety expectations for automated vehicles can unnecessarily complicate relations between regulators and manufacturers, and thus hinder the introduction of the technology. In this paper, we review relevant documents at the UN- and EU-level, for the UK, and Germany regarding their (implied) notions of safety and risk. We contrast the regulatory notions with established and more recently developing notions of safety and risk in the field of automated driving. Based on the analysis, we provide recommendations on how explicit definitions of safety and risk in regulatory documents can support rather than hinder the market introduction of automated vehicles.
2502.06595
Surrogate models for diffusion on graphs via sparse polynomials
math.NA cs.LG cs.NA
Diffusion kernels over graphs have been widely utilized as effective tools in various applications due to their ability to accurately model the flow of information through nodes and edges. However, there is a notable gap in the literature regarding the development of surrogate models for diffusion processes on graphs. In this work, we fill this gap by proposing sparse polynomial-based surrogate models for parametric diffusion equations on graphs with community structure. In tandem, we provide convergence guarantees for both least squares and compressed sensing-based approximations by showing the holomorphic regularity of parametric solutions to these diffusion equations. Our theoretical findings are accompanied by a series of numerical experiments conducted on both synthetic and real-world graphs that demonstrate the applicability of our methodology.
2502.06597
Continual Release Moment Estimation with Differential Privacy
cs.LG stat.ML
We propose Joint Moment Estimation (JME), a method for continually and privately estimating both the first and second moments of data with reduced noise compared to naive approaches. JME uses the matrix mechanism and a joint sensitivity analysis to allow the second moment estimation with no additional privacy cost, thereby improving accuracy while maintaining privacy. We demonstrate JME's effectiveness in two applications: estimating the running mean and covariance matrix for Gaussian density estimation, and model training with DP-Adam on CIFAR-10.
2502.06599
Joint parameter and state estimation for regularized time-discrete multibody dynamics
math.OC cs.SY eess.SY
We develop a method for offline parameter estimation of discrete multibody dynamics with regularized and frictional kinematic constraints. This setting leads to unobserved degrees of freedom, which we handle using joint state and parameter estimation. Our method finds the states and parameters as the solution to a nonlinear least squares optimization problem based on the inverse dynamics and the observation error. The solution is found using a Levenberg-Marquardt algorithm with derivatives from automatic differentiation and custom differentiation rules for the complementary conditions that appear due to dry frictional constraints. We reduce the number of method parameters to the choice of the time-step, regularization coefficients, and a parameter that controls the relative weighting of inverse dynamics and observation errors. We evaluate the method using synthetic and real measured data, focusing on performance and sensitivity to method parameters. In particular, we optimize over a 13-dimensional parameter space, including inertial, frictional, tilt, and motor parameters, using data from a real Furuta pendulum. Results show fast convergence, in the order of seconds, and good agreement for different time-series of recorded data over multiple method parameter choices. However, very stiff constraints may cause difficulties in solving the optimization problem. We conclude that our method can be very fast and has method parameters that are robust and easy to set in the tested scenarios.
2502.06600
Evaluation of Multilingual Image Captioning: How far can we get with CLIP models?
cs.CL cs.AI
The evaluation of image captions, looking at both linguistic fluency and semantic correspondence to visual contents, has witnessed a significant effort. Still, despite advancements such as the CLIPScore metric, multilingual captioning evaluation has remained relatively unexplored. This work presents several strategies, and extensive experiments, related to evaluating CLIPScore variants in multilingual settings. To address the lack of multilingual test data, we consider two different strategies: (1) using quality aware machine-translated datasets with human judgements, and (2) re-purposing multilingual datasets that target semantic inference and reasoning. Our results highlight the potential of finetuned multilingual models to generalize across languages and to handle complex linguistic challenges. Tests with machine-translated data show that multilingual CLIPScore models can maintain a high correlation with human judgements across different languages, and additional tests with natively multilingual and multicultural data further attest to the high-quality assessments.
2502.06601
Amortized In-Context Bayesian Posterior Estimation
cs.LG cs.AI stat.ML
Bayesian inference provides a natural way of incorporating prior beliefs and assigning a probability measure to the space of hypotheses. Current solutions rely on iterative routines like Markov Chain Monte Carlo (MCMC) sampling and Variational Inference (VI), which need to be re-run whenever new observations are available. Amortization, through conditional estimation, is a viable strategy to alleviate such difficulties and has been the guiding principle behind simulation-based inference, neural processes and in-context methods using pre-trained models. In this work, we conduct a thorough comparative analysis of amortized in-context Bayesian posterior estimation methods from the lens of different optimization objectives and architectural choices. Such methods train an amortized estimator to perform posterior parameter inference by conditioning on a set of data examples passed as context to a sequence model such as a transformer. In contrast to language models, we leverage permutation invariant architectures as the true posterior is invariant to the ordering of context examples. Our empirical study includes generalization to out-of-distribution tasks, cases where the assumed underlying model is misspecified, and transfer from simulated to real problems. Subsequently, it highlights the superiority of the reverse KL estimator for predictive problems, especially when combined with the transformer architecture and normalizing flows.
2502.06604
Do we really have to filter out random noise in pre-training data for language models?
cs.CL
Web-scale pre-training datasets are the cornerstone of LLMs' success. However, text data curated from the internet inevitably contains random noise caused by decoding errors or unregulated web content. In contrast to previous works that focus on low quality or synthetic data, our study \textbf{provides the first systematic investigation into such random noise through a cohesive ``What-Why-How'' framework.} Surprisingly, we observed that the resulting increase in next-token prediction (NTP) loss was significantly lower than the proportion of random noise. We provide a theoretical justification for this phenomenon, which also elucidates the success of multilingual models. On the other hand, experiments show that the model's performance in downstream tasks is not based solely on the NTP loss, which means that random noise may result in degraded downstream performance. To address the potential adverse effects, we introduce a novel plug-and-play Local Gradient Matching loss, which explicitly enhances the denoising capability of the downstream task head by aligning the gradient of normal and perturbed features without requiring knowledge of the model's parameters. Additional experiments on 8 language and 14 vision benchmarks further validate its effectiveness.
2502.06606
MaterialFusion: High-Quality, Zero-Shot, and Controllable Material Transfer with Diffusion Models
cs.CV
Manipulating the material appearance of objects in images is critical for applications like augmented reality, virtual prototyping, and digital content creation. We present MaterialFusion, a novel framework for high-quality material transfer that allows users to adjust the degree of material application, achieving an optimal balance between new material properties and the object's original features. MaterialFusion seamlessly integrates the modified object into the scene by maintaining background consistency and mitigating boundary artifacts. To thoroughly evaluate our approach, we have compiled a dataset of real-world material transfer examples and conducted complex comparative analyses. Through comprehensive quantitative evaluations and user studies, we demonstrate that MaterialFusion significantly outperforms existing methods in terms of quality, user control, and background preservation. Code is available at https://github.com/ControlGenAI/MaterialFusion.
2502.06607
Illegal Waste Detection in Remote Sensing Images: A Case Study
cs.CV cs.AI
Environmental crime currently represents the third largest criminal activity worldwide while threatening ecosystems as well as human health. Among the crimes related to this activity, improper waste management can nowadays be countered more easily thanks to the increasing availability and decreasing cost of Very-High-Resolution Remote Sensing images, which enable semi-automatic territory scanning in search of illegal landfills. This paper proposes a pipeline, developed in collaboration with professionals from a local environmental agency, for detecting candidate illegal dumping sites leveraging a classifier of Remote Sensing images. To identify the best configuration for such classifier, an extensive set of experiments was conducted and the impact of diverse image characteristics and training settings was thoroughly analyzed. The local environmental agency was then involved in an experimental exercise where outputs from the developed classifier were integrated in the experts' everyday work, resulting in time savings with respect to manual photo-interpretation. The classifier was eventually run with valuable results on a location outside of the training area, highlighting potential for cross-border applicability of the proposed pipeline.
2502.06608
TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Models
cs.CV cs.AI
Recent advancements in diffusion techniques have propelled image and video generation to unprecedented levels of quality, significantly accelerating the deployment and application of generative AI. However, 3D shape generation technology has so far lagged behind, constrained by limitations in 3D data scale, complexity of 3D data processing, and insufficient exploration of advanced techniques in the 3D domain. Current approaches to 3D shape generation face substantial challenges in terms of output quality, generalization capability, and alignment with input conditions. We present TripoSG, a new streamlined shape diffusion paradigm capable of generating high-fidelity 3D meshes with precise correspondence to input images. Specifically, we propose: 1) A large-scale rectified flow transformer for 3D shape generation, achieving state-of-the-art fidelity through training on extensive, high-quality data. 2) A hybrid supervised training strategy combining SDF, normal, and eikonal losses for 3D VAE, achieving high-quality 3D reconstruction performance. 3) A data processing pipeline to generate 2 million high-quality 3D samples, highlighting the crucial rules for data quality and quantity in training 3D generative models. Through comprehensive experiments, we have validated the effectiveness of each component in our new framework. The seamless integration of these parts has enabled TripoSG to achieve state-of-the-art performance in 3D shape generation. The resulting 3D shapes exhibit enhanced detail due to high-resolution capabilities and demonstrate exceptional fidelity to input images. Moreover, TripoSG demonstrates improved versatility in generating 3D models from diverse image styles and contents, showcasing strong generalization capabilities. To foster progress and innovation in the field of 3D generation, we will make our model publicly available.
2502.06615
Multi-Scale Feature Fusion with Image-Driven Spatial Integration for Left Atrium Segmentation from Cardiac MRI Images
cs.CV eess.IV
Accurate segmentation of the left atrium (LA) from late gadolinium-enhanced magnetic resonance imaging plays a vital role in visualizing diseased atrial structures, enabling the diagnosis and management of cardiovascular diseases. It is particularly essential for planning treatment with ablation therapy, a key intervention for atrial fibrillation (AF). However, manual segmentation is time-intensive and prone to inter-observer variability, underscoring the need for automated solutions. Class-agnostic foundation models like DINOv2 have demonstrated remarkable feature extraction capabilities in vision tasks. However, their lack of domain specificity and task-specific adaptation can reduce spatial resolution during feature extraction, impacting the capture of fine anatomical detail in medical imaging. To address this limitation, we propose a segmentation framework that integrates DINOv2 as an encoder with a UNet-style decoder, incorporating multi-scale feature fusion and input image integration to enhance segmentation accuracy. The learnable weighting mechanism dynamically prioritizes hierarchical features from different encoder blocks of the foundation model, optimizing feature selection for task relevance. Additionally, the input image is reintroduced during the decoding stage to preserve high-resolution spatial details, addressing limitations of downsampling in the encoder. We validate our approach on the LAScarQS 2022 dataset and demonstrate improved performance with a 92.3% Dice and 84.1% IoU score for giant architecture compared to the nnUNet baseline model. These findings emphasize the efficacy of our approach in advancing the field of automated left atrium segmentation from cardiac MRI.
2502.06617
Scaling Multi-Document Event Summarization: Evaluating Compression vs. Full-Text Approaches
cs.CL
Automatically summarizing large text collections is a valuable tool for document research, with applications in journalism, academic research, legal work, and many other fields. In this work, we contrast two classes of systems for large-scale multi-document summarization (MDS): compression and full-text. Compression-based methods use a multi-stage pipeline and often lead to lossy summaries. Full-text methods promise a lossless summary by relying on recent advances in long-context reasoning. To understand their utility on large-scale MDS, we evaluated them on three datasets, each containing approximately one hundred documents per summary. Our experiments cover a diverse set of long-context transformers (Llama-3.1, Command-R, Jamba-1.5-Mini) and compression methods (retrieval-augmented, hierarchical, incremental). Overall, we find that full-text and retrieval methods perform the best in most settings. With further analysis into the salient information retention patterns, we show that compression-based methods show strong promise at intermediate stages, even outperforming full-context. However, they suffer information loss due to their multi-stage pipeline and lack of global context. Our results highlight the need to develop hybrid approaches that combine compression and full-text approaches for optimal performance on large-scale multi-document summarization.
2502.06618
On the Reliability of Information Retrieval From MDS Coded Data in DNA Storage
cs.IT cs.ET math.IT
This work presents a theoretical analysis of the probability of successfully retrieving data encoded with MDS codes (e.g., Reed-Solomon codes) in DNA storage systems. We study this probability under independent and identically distributed (i.i.d.) substitution errors, focusing on a common code design strategy that combines inner and outer MDS codes. Our analysis demonstrates how this probability depends on factors such as the total number of sequencing reads, their distribution across strands, the rates of the inner and outer codes, and the substitution error probabilities. These results provide actionable insights into optimizing DNA storage systems under reliability constraints, including determining the minimum number of sequencing reads needed for reliable data retrieval and identifying the optimal balance between the rates of inner and outer MDS codes.
2502.06619
Unleashing the Potential of Pre-Trained Diffusion Models for Generalizable Person Re-Identification
cs.CV
Domain-generalizable re-identification (DG Re-ID) aims to train a model on one or more source domains and evaluate its performance on unseen target domains, a task that has attracted growing attention due to its practical relevance. While numerous methods have been proposed, most rely on discriminative or contrastive learning frameworks to learn generalizable feature representations. However, these approaches often fail to mitigate shortcut learning, leading to suboptimal performance. In this work, we propose a novel method called diffusion model-assisted representation learning with a correlation-aware conditioning scheme (DCAC) to enhance DG Re-ID. Our method integrates a discriminative and contrastive Re-ID model with a pre-trained diffusion model through a correlation-aware conditioning scheme. By incorporating ID classification probabilities generated from the Re-ID model with a set of learnable ID-wise prompts, the conditioning scheme injects dark knowledge that captures ID correlations to guide the diffusion process. Simultaneously, feedback from the diffusion model is back-propagated through the conditioning scheme to the Re-ID model, effectively improving the generalization capability of Re-ID features. Extensive experiments on both single-source and multi-source DG Re-ID tasks demonstrate that our method achieves state-of-the-art performance. Comprehensive ablation studies further validate the effectiveness of the proposed approach, providing insights into its robustness. Codes will be available at https://github.com/RikoLi/DCAC.
2502.06627
Towards Closing the Gap between Model-Based Systems Engineering and Automated Vehicle Assurance: Tailoring Generic Methods by Integrating Domain Knowledge
eess.SY cs.SY
Designing, assuring and releasing safe automated vehicles is a highly interdisciplinary process. As complex systems, automated driving systems will inevitably be subject to emergent properties, i. e., the properties of the overall system will be more than just a sum of the properties of its integrated elements. Safety is one example of such emergent properties. In this regard, it must be ensured that effects of emergence do not render an overall system that is composed of safety-approved sub systems unsafe. The key challenges in this regard are twofold: Regarding the interdisciplinary character of the development and assurance processes, all relevant stakeholders must speak a common language and have a common understanding of the key concepts that influence system safety. Additionally, the individual properties of system elements should remain traceable to the system level. Model-Based Systems Engineering (MBSE) provides an interdisciplinary mindset, as well as methods and processes to manage emergent system properties over the entire system lifecycle. By this, MBSE provides tools that can assist the assurance process for automated vehicles. However, concepts from the domain of MBSE have a reputation for not being directly accessible for domain experts who are no experts in the field of Systems Engineering. This paper highlights challenges when applying MBSE methods to the design and development of automated driving systems. It will present an approach to create and apply domain-specific SysML profiles, which can be a first step for enhancing communication between different stake-holders in the development and safety assurance processes.
2502.06631
Conformal Predictions for Human Action Recognition with Vision-Language Models
cs.CV cs.AI
Human-In-The-Loop (HITL) frameworks are integral to many real-world computer vision systems, enabling human operators to make informed decisions with AI assistance. Conformal Predictions (CP), which provide label sets with rigorous guarantees on ground truth inclusion probabilities, have recently gained traction as a valuable tool in HITL settings. One key application area is video surveillance, closely associated with Human Action Recognition (HAR). This study explores the application of CP on top of state-of-the-art HAR methods that utilize extensively pre-trained Vision-Language Models (VLMs). Our findings reveal that CP can significantly reduce the average number of candidate classes without modifying the underlying VLM. However, these reductions often result in distributions with long tails. To address this, we introduce a method based on tuning the temperature parameter of the VLMs to minimize these tails without requiring additional calibration data. Our code is made available on GitHub at the address https://github.com/tbary/CP4VLM.
2502.06632
Few-Shot Classification and Anatomical Localization of Tissues in SPECT Imaging
cs.CV cs.AI cs.LG
Accurate classification and anatomical localization are essential for effective medical diagnostics and research, which may be efficiently performed using deep learning techniques. However, availability of limited labeled data poses a significant challenge. To address this, we adapted Prototypical Networks and the Propagation-Reconstruction Network (PRNet) for few-shot classification and localization, respectively, in Single Photon Emission Computed Tomography (SPECT) images. For the proof of concept we used a 2D-sliced image cropped around heart. The Prototypical Network, with a pre-trained ResNet-18 backbone, classified ventricles, myocardium, and liver tissues with 96.67% training and 93.33% validation accuracy. PRNet, adapted for 2D imaging with an encoder-decoder architecture and skip connections, achieved a training loss of 1.395, accurately reconstructing patches and capturing spatial relationships. These results highlight the potential of Prototypical Networks for tissue classification with limited labeled data and PRNet for anatomical landmark localization, paving the way for improved performance in deep learning frameworks.
2502.06633
Combining Large Language Models with Static Analyzers for Code Review Generation
cs.SE cs.AI
Code review is a crucial but often complex, subjective, and time-consuming activity in software development. Over the past decades, significant efforts have been made to automate this process. Early approaches focused on knowledge-based systems (KBS) that apply rule-based mechanisms to detect code issues, providing precise feedback but struggling with complex, context-dependent cases. More recent work has shifted toward fine-tuning pre-trained language models for code review, enabling broader issue coverage but often at the expense of precision. In this paper, we propose a hybrid approach that combines the strengths of KBS and learning-based systems (LBS) to generate high-quality, comprehensive code reviews. Our method integrates knowledge at three distinct stages of the language model pipeline: during data preparation (Data-Augmented Training, DAT), at inference (Retrieval-Augmented Generation, RAG), and after inference (Naive Concatenation of Outputs, NCO). We empirically evaluate our combination strategies against standalone KBS and LBS fine-tuned on a real-world dataset. Our results show that these hybrid strategies enhance the relevance, completeness, and overall quality of review comments, effectively bridging the gap between rule-based tools and deep learning models.
2502.06634
Automatic Annotation Augmentation Boosts Translation between Molecules and Natural Language
cs.LG cs.AI q-bio.BM
Recent advancements in AI for biological research focus on integrating molecular data with natural language to accelerate drug discovery. However, the scarcity of high-quality annotations limits progress in this area. This paper introduces LA$^3$, a Language-based Automatic Annotation Augmentation framework that leverages large language models to augment existing datasets, thereby improving AI training. We demonstrate the effectiveness of LA$^3$ by creating an enhanced dataset, LaChEBI-20, where we systematically rewrite the annotations of molecules from an established dataset. These rewritten annotations preserve essential molecular information while providing more varied sentence structures and vocabulary. Using LaChEBI-20, we train LaMolT5 based on a benchmark architecture to learn the mapping between molecular representations and augmented annotations. Experimental results on text-based *de novo* molecule generation and molecule captioning demonstrate that LaMolT5 outperforms state-of-the-art models. Notably, incorporating LA$^3$ leads to improvements of up to 301% over the benchmark architecture. Furthermore, we validate the effectiveness of LA$^3$ notable applications in *image*, *text* and *graph* tasks, affirming its versatility and utility.
2502.06635
Steel-LLM:From Scratch to Open Source -- A Personal Journey in Building a Chinese-Centric LLM
cs.CL cs.AI
Steel-LLM is a Chinese-centric language model developed from scratch with the goal of creating a high-quality, open-source model despite limited computational resources. Launched in March 2024, the project aimed to train a 1-billion-parameter model on a large-scale dataset, prioritizing transparency and the sharing of practical insights to assist others in the community. The training process primarily focused on Chinese data, with a small proportion of English data included, addressing gaps in existing open-source LLMs by providing a more detailed and practical account of the model-building journey. Steel-LLM has demonstrated competitive performance on benchmarks such as CEVAL and CMMLU, outperforming early models from larger institutions. This paper provides a comprehensive summary of the project's key contributions, including data collection, model design, training methodologies, and the challenges encountered along the way, offering a valuable resource for researchers and practitioners looking to develop their own LLMs. The model checkpoints and training script are available at https://github.com/zhanshijinwat/Steel-LLM.
2502.06636
Enhancing healthcare infrastructure resilience through agent-based simulation methods
cs.MA cs.SY eess.SY
Critical infrastructures face demanding challenges due to natural and human-generated threats, such as pandemics, workforce shortages or cyber-attacks, which might severely compromise service quality. To improve system resilience, decision-makers would need intelligent tools for quick and efficient resource allocation. This article explores an agent-based simulation model that intends to capture a part of the complexity of critical infrastructure systems, particularly considering the interdependencies of healthcare systems with information and telecommunication systems. Such a model enables to implement a simulation-based optimization approach in which the exposure of critical systems to risks is evaluated, while comparing the mitigation effects of multiple tactical and strategical decision alternatives to enhance their resilience. The proposed model is designed to be parameterizable, to enable adapting it to risk scenarios with different severity, and it facilitates the compilation of relevant performance indicators enabling monitoring at both agent level and system level. To validate the agent-based model, a literature-supported methodology has been used to perform cross-validation, sensitivity analysis and test the usefulness of the proposed model through a use case. The use case analyzes the impact of a concurrent pandemic and a cyber-attack on a hospital and compares different resiliency-enhancing countermeasures using contingency tables. Overall, the use case illustrates the feasibility and versatility of the proposed approach to enhance resiliency.
2502.06643
MoETuner: Optimized Mixture of Expert Serving with Balanced Expert Placement and Token Routing
cs.LG cs.DC
Mixture-of-Experts (MoE) model architecture has emerged as a promising solution for scaling transformer models efficiently, offering sparse activation that reduces computational costs while increasing model capacity. However, as MoE models scale, they need to be distributed across GPU devices, thus face critical performance bottlenecks due to their large memory footprint. Expert parallelism distributes experts across GPUs, however, faces key challenges including an unbalanced token routing and expert activation, resulting in communication tail latency and processing inefficiencies. While existing solutions address some of these issues, they fail to resolve the dual challenges of load imbalance and communication skew. The imbalance in token processing load across experts causes uneven processing times on different GPUs, while communication skew between GPUs leads to unbalanced inter-GPU data transfers. These factors degrade the performance of MoE models by increasing tail latency and reducing overall throughput. To address these limitations, we propose an Integer Linear Programming (ILP) formulation to optimize expert placement by jointly considering token load, communication, and computation costs. We exploit the property that there is a token routing dependency across layers, where tokens routed to a specific expert in one layer are likely to be routed to a limited set of experts in the subsequent layer. Our solution, MoETuner, offers an optimal expert-to-GPU assignment that minimizes inter-GPU token routing costs and balances token processing across devices, thereby reducing tail latency and end-to-end execution time. Experimental results demonstrate 9.3% and 17.5% of end-to-end speedups for single-node and multi-node inference respectively, showcasing the potential of our ILP-based optimization for offering expert parallel solutions for next-generation MoEs.
2502.06645
Koopman-Equivariant Gaussian Processes
cs.LG cs.SY eess.SY stat.ML
Credible forecasting and representation learning of dynamical systems are of ever-increasing importance for reliable decision-making. To that end, we propose a family of Gaussian processes (GP) for dynamical systems with linear time-invariant responses, which are nonlinear only in initial conditions. This linearity allows us to tractably quantify forecasting and representational uncertainty, simultaneously alleviating the challenge of computing the distribution of trajectories from a GP-based dynamical system and enabling a new probabilistic treatment of learning Koopman operator representations. Using a trajectory-based equivariance -- which we refer to as \textit{Koopman equivariance} -- we obtain a GP model with enhanced generalization capabilities. To allow for large-scale regression, we equip our framework with variational inference based on suitable inducing points. Experiments demonstrate on-par and often better forecasting performance compared to kernel-based methods for learning dynamical systems.
2502.06648
The 2021 Tokyo Olympics Multilingual News Article Dataset
cs.IR cs.AI cs.CL
In this paper, we introduce a dataset of multilingual news articles covering the 2021 Tokyo Olympics. A total of 10,940 news articles were gathered from 1,918 different publishers, covering 1,350 sub-events of the 2021 Olympics, and published between July 1, 2021, and August 14, 2021. These articles are written in nine languages from different language families and in different scripts. To create the dataset, the raw news articles were first retrieved via a service that collects and analyzes news articles. Then, the articles were grouped using an online clustering algorithm, with each group containing articles reporting on the same sub-event. Finally, the groups were manually annotated and evaluated. The development of this dataset aims to provide a resource for evaluating the performance of multilingual news clustering algorithms, for which limited datasets are available. It can also be used to analyze the dynamics and events of the 2021 Tokyo Olympics from different perspectives. The dataset is available in CSV format and can be accessed from the CLARIN.SI repository.
2502.06649
Estimation of Food Intake Quantity Using Inertial Signals from Smartwatches
eess.SP cs.LG
Accurate monitoring of eating behavior is crucial for managing obesity and eating disorders such as bulimia nervosa. At the same time, existing methods rely on multiple and/or specialized sensors, greatly harming adherence and ultimately, the quality and continuity of data. This paper introduces a novel approach for estimating the weight of a bite, from a commercial smartwatch. Our publicly-available dataset contains smartwatch inertial data from ten participants, with manually annotated start and end times of each bite along with their corresponding weights from a smart scale, under semi-controlled conditions. The proposed method combines extracted behavioral features such as the time required to load the utensil with food, with statistical features of inertial signals, that serve as input to a Support Vector Regression model to estimate bite weights. Under a leave-one-subject-out cross-validation scheme, our approach achieves a mean absolute error (MAE) of 3.99 grams per bite. To contextualize this performance, we introduce the improvement metric, that measures the relative MAE difference compared to a baseline model. Our method demonstrates a 17.41% improvement, while the adapted state-of-the art method shows a -28.89% performance against that same baseline. The results presented in this work establish the feasibility of extracting meaningful bite weight estimates from commercial smartwatch inertial sensors alone, laying the groundwork for future accessible, non-invasive dietary monitoring systems.
2502.06650
Prototype Contrastive Consistency Learning for Semi-Supervised Medical Image Segmentation
cs.CV
Medical image segmentation is a crucial task in medical image analysis, but it can be very challenging especially when there are less labeled data but with large unlabeled data. Contrastive learning has proven to be effective for medical image segmentation in semi-supervised learning by constructing contrastive samples from partial pixels. However, although previous contrastive learning methods can mine semantic information from partial pixels within images, they ignore the whole context information of unlabeled images, which is very important to precise segmentation. In order to solve this problem, we propose a novel prototype contrastive learning method called Prototype Contrastive Consistency Segmentation (PCCS) for semi-supervised medical image segmentation. The core idea is to enforce the prototypes of the same semantic class to be closer and push the prototypes in different semantic classes far away from each other. Specifically, we construct a signed distance map and an uncertainty map from unlabeled images. The signed distance map is used to construct prototypes for contrastive learning, and then we estimate the prototype uncertainty from the uncertainty map as trade-off among prototypes. In order to obtain better prototypes, based on the student-teacher architecture, a new mechanism named prototype updating prototype is designed to assist in updating the prototypes for contrastive learning. In addition, we propose an uncertainty-consistency loss to mine more reliable information from unlabeled data. Extensive experiments on medical image segmentation demonstrate that PCCS achieves better segmentation performance than the state-of-the-art methods. The code is available at https://github.com/comphsh/PCCS.
2502.06652
Transparent NLP: Using RAG and LLM Alignment for Privacy Q&A
cs.CL
The transparency principle of the General Data Protection Regulation (GDPR) requires data processing information to be clear, precise, and accessible. While language models show promise in this context, their probabilistic nature complicates truthfulness and comprehensibility. This paper examines state-of-the-art Retrieval Augmented Generation (RAG) systems enhanced with alignment techniques to fulfill GDPR obligations. We evaluate RAG systems incorporating an alignment module like Rewindable Auto-regressive Inference (RAIN) and our proposed multidimensional extension, MultiRAIN, using a Privacy Q&A dataset. Responses are optimized for preciseness and comprehensibility and are assessed through 21 metrics, including deterministic and large language model-based evaluations. Our results show that RAG systems with an alignment module outperform baseline RAG systems on most metrics, though none fully match human answers. Principal component analysis of the results reveals complex interactions between metrics, highlighting the need to refine metrics. This study provides a foundation for integrating advanced natural language processing systems into legal compliance frameworks.
2502.06653
In-Context Learning (and Unlearning) of Length Biases
cs.CL
Large language models have demonstrated strong capabilities to learn in-context, where exemplar input-output pairings are appended to the prompt for demonstration. However, existing work has demonstrated the ability of models to learn lexical and label biases in-context, which negatively impacts both performance and robustness of models. The impact of other statistical data biases remains under-explored, which this work aims to address. We specifically investigate the impact of length biases on in-context learning. We demonstrate that models do learn length biases in the context window for their predictions, and further empirically analyze the factors that modulate the level of bias exhibited by the model. In addition, we show that learning length information in-context can be used to counter the length bias that has been encoded in models (e.g., via fine-tuning). This reveals the power of in-context learning in debiasing model prediction behaviors without the need for costly parameter updates.
2502.06655
Unbiased Evaluation of Large Language Models from a Causal Perspective
cs.AI
Benchmark contamination has become a significant concern in the LLM evaluation community. Previous Agents-as-an-Evaluator address this issue by involving agents in the generation of questions. Despite their success, the biases in Agents-as-an-Evaluator methods remain largely unexplored. In this paper, we present a theoretical formulation of evaluation bias, providing valuable insights into designing unbiased evaluation protocols. Furthermore, we identify two type of bias in Agents-as-an-Evaluator through carefully designed probing tasks on a minimal Agents-as-an-Evaluator setup. To address these issues, we propose the Unbiased Evaluator, an evaluation protocol that delivers a more comprehensive, unbiased, and interpretable assessment of LLMs.Extensive experiments reveal significant room for improvement in current LLMs. Additionally, we demonstrate that the Unbiased Evaluator not only offers strong evidence of benchmark contamination but also provides interpretable evaluation results.
2502.06656
A Frontier AI Risk Management Framework: Bridging the Gap Between Current AI Practices and Established Risk Management
cs.AI
The recent development of powerful AI systems has highlighted the need for robust risk management frameworks in the AI industry. Although companies have begun to implement safety frameworks, current approaches often lack the systematic rigor found in other high-risk industries. This paper presents a comprehensive risk management framework for the development of frontier AI that bridges this gap by integrating established risk management principles with emerging AI-specific practices. The framework consists of four key components: (1) risk identification (through literature review, open-ended red-teaming, and risk modeling), (2) risk analysis and evaluation using quantitative metrics and clearly defined thresholds, (3) risk treatment through mitigation measures such as containment, deployment controls, and assurance processes, and (4) risk governance establishing clear organizational structures and accountability. Drawing from best practices in mature industries such as aviation or nuclear power, while accounting for AI's unique challenges, this framework provides AI developers with actionable guidelines for implementing robust risk management. The paper details how each component should be implemented throughout the life-cycle of the AI system - from planning through deployment - and emphasizes the importance and feasibility of conducting risk management work prior to the final training run to minimize the burden associated with it.
2502.06658
Generating Samples to Question Trained Models
cs.LG
There is a growing need for investigating how machine learning models operate. With this work, we aim to understand trained machine learning models by questioning their data preferences. We propose a mathematical framework that allows us to probe trained models and identify their preferred samples in various scenarios including prediction-risky, parameter-sensitive, or model-contrastive samples. To showcase our framework, we pose these queries to a range of models trained on a range of classification and regression tasks, and receive answers in the form of generated data.
2502.06659
Who Taught You That? Tracing Teachers in Model Distillation
cs.CL
Model distillation -- using outputs from a large teacher model to teach a small student model -- is a practical means of creating efficient models for a particular task. We ask: Can we identify a students' teacher based on its outputs? Such "footprints" left by teacher LLMs would be interesting artifacts. Beyond this, reliable teacher inference may have practical implications as actors seek to distill specific capabilities of massive proprietary LLMs into deployed smaller LMs, potentially violating terms of service. We consider practical task distillation targets including summarization, question answering, and instruction-following. We assume a finite set of candidate teacher models, which we treat as blackboxes. We design discriminative models that operate over lexical features. We find that $n$-gram similarity alone is unreliable for identifying teachers, but part-of-speech (PoS) templates preferred by student models mimic those of their teachers.
2502.06661
iLOCO: Distribution-Free Inference for Feature Interactions
stat.ML cs.LG
Feature importance measures are widely studied and are essential for understanding model behavior, guiding feature selection, and enhancing interpretability. However, many machine learning fitted models involve complex, higher-order interactions between features. Existing feature importance metrics fail to capture these higher-order effects while existing interaction metrics often suffer from limited applicability or excessive computation; no methods exist to conduct statistical inference for feature interactions. To bridge this gap, we first propose a new model-agnostic metric, interaction Leave-One-Covariate-Out iLOCO, for measuring the importance of higher-order feature interactions. Next, we leverage recent advances in LOCO inference to develop distribution-free and assumption-light confidence intervals for our iLOCO metric. To address computational challenges, we also introduce an ensemble learning method for calculating the iLOCO metric and confidence intervals that we show is both computationally and statistically efficient. We validate our iLOCO metric and our confidence intervals on both synthetic and real data sets, showing that our approach outperforms existing methods and provides the first inferential approach to detecting feature interactions.
2502.06663
EfficientLLM: Scalable Pruning-Aware Pretraining for Architecture-Agnostic Edge Language Models
cs.LG
Modern large language models (LLMs) driven by scaling laws, achieve intelligence emergency in large model sizes. Recently, the increasing concerns about cloud costs, latency, and privacy make it an urgent requirement to develop compact edge language models. Distinguished from direct pretraining that bounded by the scaling law, this work proposes the pruning-aware pretraining, focusing on retaining performance of much larger optimized models. It features following characteristics: 1) Data-scalable: we introduce minimal parameter groups in LLM and continuously optimize structural pruning, extending post-training pruning methods like LLM-Pruner and SparseGPT into the pretraining phase. 2) Architecture-agnostic: the LLM architecture is auto-designed using saliency-driven pruning, which is the first time to exceed SoTA human-designed LLMs in modern pretraining. We reveal that it achieves top-quality edge language models, termed EfficientLLM, by scaling up LLM compression and extending its boundary. EfficientLLM significantly outperforms SoTA baselines with $100M \sim 1B$ parameters, such as MobileLLM, SmolLM, Qwen2.5-0.5B, OLMo-1B, Llama3.2-1B in common sense benchmarks. As the first attempt, EfficientLLM bridges the performance gap between traditional LLM compression and direct pretraining methods, and we will fully open source at https://github.com/Xingrun-Xing2/EfficientLLM.
2502.06664
Evaluation of Deep Audio Representations for Hearables
cs.SD cs.AI cs.LG
Effectively steering hearable devices requires understanding the acoustic environment around the user. In the computational analysis of sound scenes, foundation models have emerged as the state of the art to produce high-performance, robust, multi-purpose audio representations. We introduce and release Deep Evaluation of Audio Representations (DEAR), the first dataset and benchmark to evaluate the efficacy of foundation models in capturing essential acoustic properties for hearables. The dataset includes 1,158 audio tracks, each 30 seconds long, created by spatially mixing proprietary monologues with commercial, high-quality recordings of everyday acoustic scenes. Our benchmark encompasses eight tasks that assess the general context, speech sources, and technical acoustic properties of the audio scenes. Through our evaluation of four general-purpose audio representation models, we demonstrate that the BEATs model significantly surpasses its counterparts. This superiority underscores the advantage of models trained on diverse audio collections, confirming their applicability to a wide array of auditory tasks, including encoding the environment properties necessary for hearable steering. The DEAR dataset and associated code are available at https://dear-dataset.github.io.
2502.06666
Automatic Evaluation of Healthcare LLMs Beyond Question-Answering
cs.CL cs.AI
Current Large Language Models (LLMs) benchmarks are often based on open-ended or close-ended QA evaluations, avoiding the requirement of human labor. Close-ended measurements evaluate the factuality of responses but lack expressiveness. Open-ended capture the model's capacity to produce discourse responses but are harder to assess for correctness. These two approaches are commonly used, either independently or together, though their relationship remains poorly understood. This work is focused on the healthcare domain, where both factuality and discourse matter greatly. It introduces a comprehensive, multi-axis suite for healthcare LLM evaluation, exploring correlations between open and close benchmarks and metrics. Findings include blind spots and overlaps in current methodologies. As an updated sanity check, we release a new medical benchmark --CareQA-- with both open and closed variants. Finally, we propose a novel metric for open-ended evaluations -- Relaxed Perplexity -- to mitigate the identified limitations.
2502.06669
Boosting Self-Efficacy and Performance of Large Language Models via Verbal Efficacy Stimulations
cs.CL cs.AI
Significant improvements have been observed in the zero-shot capabilities of the Large Language Models (LLMs). Due to their high sensitivity to input, research has increasingly focused on enhancing LLMs' performance via direct and simple prompt engineering rather than intricate domain adaptation. Studies suggest that LLMs exhibit emotional intelligence, and both positive and negative emotions can potentially enhance task performances. However, prior interaction prompts have predominantly concentrated on a single stimulus type, neglecting to compare different stimulus effects, examine the influence of varying task difficulties, or explore underlying mechanisms. This paper, inspired by the positive correlation between self-efficacy and task performance within the social cognitive theory, introduces Verbal Efficacy Stimulations (VES). Our VES comprises three types of verbal prompts: encouraging, provocative, and critical, addressing six aspects such as helpfulness and competence. And we further categorize task difficulty, aiming to extensively investigate how distinct VES influence the self-efficacy and task achievements of language models at varied levels of difficulty. The experimental results show that the three types of VES improve the performance of LLMs on most tasks, and the most effective VES varies for different models. In extensive experiments, we have obtained some findings consistent with psychological theories, providing novel insights for future research.
2502.06673
Selecting Optimal Sampling Rate for Stable Super-Resolution
math.NA cs.IT cs.NA math.IT
We investigate the recovery of nodes and amplitudes from noisy frequency samples in spike train signals, also known as the super-resolution (SR) problem. When the node separation falls below the Rayleigh limit, the problem becomes ill-conditioned. Admissible sampling rates, or decimation parameters, improve the conditioning of the SR problem, enabling more accurate recovery. We propose an efficient preprocessing method to identify the optimal sampling rate, significantly enhancing the performance of SR techniques.
2502.06674
RAILS: Risk-Aware Iterated Local Search for Joint SLA Decomposition and Service Provider Management in Multi-Domain Networks
cs.NI cs.LG
The emergence of the fifth generation (5G) technology has transformed mobile networks into multi-service environments, necessitating efficient network slicing to meet diverse Service Level Agreements (SLAs). SLA decomposition across multiple network domains, each potentially managed by different service providers, poses a significant challenge due to limited visibility into real-time underlying domain conditions. This paper introduces Risk-Aware Iterated Local Search (RAILS), a novel risk model-driven meta-heuristic framework designed to jointly address SLA decomposition and service provider selection in multi-domain networks. By integrating online risk modeling with iterated local search principles, RAILS effectively navigates the complex optimization landscape, utilizing historical feedback from domain controllers. We formulate the joint problem as a Mixed-Integer Nonlinear Programming (MINLP) problem and prove its NP-hardness. Extensive simulations demonstrate that RAILS achieves near-optimal performance, offering an efficient, real-time solution for adaptive SLA management in modern multi-domain networks.
2502.06676
Discovery of skill switching criteria for learning agile quadruped locomotion
cs.RO
This paper develops a hierarchical learning and optimization framework that can learn and achieve well-coordinated multi-skill locomotion. The learned multi-skill policy can switch between skills automatically and naturally in tracking arbitrarily positioned goals and recover from failures promptly. The proposed framework is composed of a deep reinforcement learning process and an optimization process. First, the contact pattern is incorporated into the reward terms for learning different types of gaits as separate policies without the need for any other references. Then, a higher level policy is learned to generate weights for individual policies to compose multi-skill locomotion in a goal-tracking task setting. Skills are automatically and naturally switched according to the distance to the goal. The proper distances for skill switching are incorporated in reward calculation for learning the high level policy and updated by an outer optimization loop as learning progresses. We first demonstrated successful multi-skill locomotion in comprehensive tasks on a simulated Unitree A1 quadruped robot. We also deployed the learned policy in the real world showcasing trotting, bounding, galloping, and their natural transitions as the goal position changes. Moreover, the learned policy can react to unexpected failures at any time, perform prompt recovery, and resume locomotion successfully. Compared to discrete switch between single skills which failed to transition to galloping in the real world, our proposed approach achieves all the learned agile skills, with smoother and more continuous skill transitions.
2502.06678
Quantile Multi-Armed Bandits with 1-bit Feedback
stat.ML cs.IT cs.LG math.IT
In this paper, we study a variant of best-arm identification involving elements of risk sensitivity and communication constraints. Specifically, the goal of the learner is to identify the arm with the highest quantile reward, while the communication from an agent (who observes rewards) and the learner (who chooses actions) is restricted to only one bit of feedback per arm pull. We propose an algorithm that utilizes noisy binary search as a subroutine, allowing the learner to estimate quantile rewards through 1-bit feedback. We derive an instance-dependent upper bound on the sample complexity of our algorithm and provide an algorithm-independent lower bound for specific instances, with the two matching to within logarithmic factors under mild conditions, or even to within constant factors in certain low error probability scaling regimes. The lower bound is applicable even in the absence of communication constraints, and thus we conclude that restricting to 1-bit feedback has a minimal impact on the scaling of the sample complexity.
2502.06681
CHIRLA: Comprehensive High-resolution Identification and Re-identification for Large-scale Analysis
cs.CV cs.AI cs.LG
Person re-identification (Re-ID) is a key challenge in computer vision, requiring the matching of individuals across different cameras, locations, and time periods. While most research focuses on short-term scenarios with minimal appearance changes, real-world applications demand robust Re-ID systems capable of handling long-term scenarios, where persons' appearances can change significantly due to variations in clothing and physical characteristics. In this paper, we present CHIRLA, Comprehensive High-resolution Identification and Re-identification for Large-scale Analysis, a novel dataset specifically designed for long-term person Re-ID. CHIRLA consists of recordings from strategically placed cameras over a seven-month period, capturing significant variations in both temporal and appearance attributes, including controlled changes in participants' clothing and physical features. The dataset includes 22 individuals, four connected indoor environments, and seven cameras. We collected more than five hours of video that we semi-automatically labeled to generate around one million bounding boxes with identity annotations. By introducing this comprehensive benchmark, we aim to facilitate the development and evaluation of Re-ID algorithms that can reliably perform in challenging, long-term real-world scenarios.
2502.06682
Transfer Your Perspective: Controllable 3D Generation from Any Viewpoint in a Driving Scene
cs.CV
Self-driving cars relying solely on ego-centric perception face limitations in sensing, often failing to detect occluded, faraway objects. Collaborative autonomous driving (CAV) seems like a promising direction, but collecting data for development is non-trivial. It requires placing multiple sensor-equipped agents in a real-world driving scene, simultaneously! As such, existing datasets are limited in locations and agents. We introduce a novel surrogate to the rescue, which is to generate realistic perception from different viewpoints in a driving scene, conditioned on a real-world sample - the ego-car's sensory data. This surrogate has huge potential: it could potentially turn any ego-car dataset into a collaborative driving one to scale up the development of CAV. We present the very first solution, using a combination of simulated collaborative data and real ego-car data. Our method, Transfer Your Perspective (TYP), learns a conditioned diffusion model whose output samples are not only realistic but also consistent in both semantics and layouts with the given ego-car data. Empirical results demonstrate TYP's effectiveness in aiding in a CAV setting. In particular, TYP enables us to (pre-)train collaborative perception algorithms like early and late fusion with little or no real-world collaborative data, greatly facilitating downstream CAV applications.
2502.06683
Solving Optimal Power Flow on a Data-Budget: Feature Selection on Smart Meter Data
eess.SY cs.SY
How much data is needed to optimally schedule distributed energy resources (DERs)? Does the distribution system operator (DSO) have to precisely know load demands and solar injections at each bus of the feeder to solve an optimal power flow (OPF)? This work exploits redundancies in OPF's structure and data to avoid communicating such data deluge, and explores the trade-off between data compression and grid's performance. We propose an OPF data distillation framework involving two steps. The DSO first collects OPF data from only a subset of nodes. The DSO subsequently reconstructs the complete OPF data from the partial ones, and feeds them into the OPF solver. Selecting and reconstructing OPF data may be performed to either maximize the fidelity of reconstructed OPF data, or maximize the fidelity of OPF solutions corresponding to reconstructed data. Under the first objective, OPF data distillation is posed as a sparsity-regularized convex problem. Under the second objective, it is posed as a sparsity-regularized bilevel program. Both problems are solved using accelerated proximal gradient (PGD) algorithms. Numerical tests corroborate that the bilevel formulation enhances fidelity and feasibility of reconstructed OPF solutions, and that OPF solutions can be approximated reasonably well even from partial data.
2502.06684
EquiTabPFN: A Target-Permutation Equivariant Prior Fitted Networks
cs.LG cs.AI
Recent foundational models for tabular data, such as TabPFN, have demonstrated remarkable effectiveness in adapting to new tasks through in-context learning. However, these models overlook a crucial equivariance property: the arbitrary ordering of target dimensions should not influence model predictions. In this study, we identify this oversight as a source of incompressible error, termed the equivariance gap, which introduces instability in predictions. To mitigate these issues, we propose a novel model designed to preserve equivariance across output dimensions. Our experimental results indicate that our proposed model not only addresses these pitfalls effectively but also achieves competitive benchmark performance.
2502.06685
No Trick, No Treat: Pursuits and Challenges Towards Simulation-free Training of Neural Samplers
cs.LG stat.ML
We consider the sampling problem, where the aim is to draw samples from a distribution whose density is known only up to a normalization constant. Recent breakthroughs in generative modeling to approximate a high-dimensional data distribution have sparked significant interest in developing neural network-based methods for this challenging problem. However, neural samplers typically incur heavy computational overhead due to simulating trajectories during training. This motivates the pursuit of simulation-free training procedures of neural samplers. In this work, we propose an elegant modification to previous methods, which allows simulation-free training with the help of a time-dependent normalizing flow. However, it ultimately suffers from severe mode collapse. On closer inspection, we find that nearly all successful neural samplers rely on Langevin preconditioning to avoid mode collapsing. We systematically analyze several popular methods with various objective functions and demonstrate that, in the absence of Langevin preconditioning, most of them fail to adequately cover even a simple target. Finally, we draw attention to a strong baseline by combining the state-of-the-art MCMC method, Parallel Tempering (PT), with an additional generative model to shed light on future explorations of neural samplers.
2502.06689
Neumann eigenmaps for landmark embedding
math.ST cs.LG cs.NA math.NA stat.ML stat.TH
We present Neumann eigenmaps (NeuMaps), a novel approach for enhancing the standard diffusion map embedding using landmarks, i.e distinguished samples within the dataset. By interpreting these landmarks as a subgraph of the larger data graph, NeuMaps are obtained via the eigendecomposition of a renormalized Neumann Laplacian. We show that NeuMaps offer two key advantages: (1) they provide a computationally efficient embedding that accurately recovers the diffusion distance associated with the reflecting random walk on the subgraph, and (2) they naturally incorporate the Nystr\"om extension within the diffusion map framework through the discrete Neumann boundary condition. Through examples in digit classification and molecular dynamics, we demonstrate that NeuMaps not only improve upon existing landmark-based embedding methods but also enhance the stability of diffusion map embeddings to the removal of highly significant points.
2502.06692
Multi-label Scandinavian Language Identification (SLIDE)
cs.CL cs.AI
Identifying closely related languages at sentence level is difficult, in particular because it is often impossible to assign a sentence to a single language. In this paper, we focus on multi-label sentence-level Scandinavian language identification (LID) for Danish, Norwegian Bokm\r{a}l, Norwegian Nynorsk, and Swedish. We present the Scandinavian Language Identification and Evaluation, SLIDE, a manually curated multi-label evaluation dataset and a suite of LID models with varying speed-accuracy tradeoffs. We demonstrate that the ability to identify multiple languages simultaneously is necessary for any accurate LID method, and present a novel approach to training such multi-label LID models.
2502.06693
Recent Advances, Applications and Open Challenges in Machine Learning for Health: Reflections from Research Roundtables at ML4H 2024 Symposium
cs.LG cs.AI cs.CY
The fourth Machine Learning for Health (ML4H) symposium was held in person on December 15th and 16th, 2024, in the traditional, ancestral, and unceded territories of the Musqueam, Squamish, and Tsleil-Waututh Nations in Vancouver, British Columbia, Canada. The symposium included research roundtable sessions to foster discussions between participants and senior researchers on timely and relevant topics for the ML4H community. The organization of the research roundtables at the conference involved 13 senior and 27 junior chairs across 13 tables. Each roundtable session included an invited senior chair (with substantial experience in the field), junior chairs (responsible for facilitating the discussion), and attendees from diverse backgrounds with an interest in the session's topic.
2502.06695
FairDropout: Using Example-Tied Dropout to Enhance Generalization of Minority Groups
cs.LG
Deep learning models frequently exploit spurious features in training data to achieve low training error, often resulting in poor generalization when faced with shifted testing distributions. To address this issue, various methods from imbalanced learning, representation learning, and classifier recalibration have been proposed to enhance the robustness of deep neural networks against spurious correlations. In this paper, we observe that models trained with empirical risk minimization tend to generalize well for examples from the majority groups while memorizing instances from minority groups. Building on recent findings that show memorization can be localized to a limited number of neurons, we apply example-tied dropout as a method we term FairDropout, aimed at redirecting this memorization to specific neurons that we subsequently drop out during inference. We empirically evaluate FairDropout using the subpopulation benchmark suite encompassing vision, language, and healthcare tasks, demonstrating that it significantly reduces reliance on spurious correlations, and outperforms state-of-the-art methods.
2502.06698
Heisenberg-limited calibration of entangling gates with robust phase estimation
quant-ph cs.SY eess.SY
The calibration of high-quality two-qubit entangling gates is an essential component in engineering large-scale, fault-tolerant quantum computers. However, many standard calibration techniques are based on randomized circuits that are only quadratically sensitive to calibration errors. As a result, these approaches are inefficient, requiring many experimental shots to achieve acceptable performance. In this work, we demonstrate that robust phase estimation can enable high-precision, Heisenberg-limited estimates of coherent errors in multi-qubit gates. Equipped with an efficient estimator, the calibration problem may be reduced to a simple optimization loop that minimizes the estimated coherent error. We experimentally demonstrate our calibration protocols by improving the operation of a two-qubit controlled-Z gate on a superconducting processor, and we validate the improved performance with gate set tomography. Our methods are applicable to gates in other quantum hardware platforms such as ion traps and neutral atoms, and on other multi-qubit gates, such as CNOT or iSWAP.
2502.06703
Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling
cs.CL
Test-Time Scaling (TTS) is an important method for improving the performance of Large Language Models (LLMs) by using additional computation during the inference phase. However, current studies do not systematically analyze how policy models, Process Reward Models (PRMs), and problem difficulty influence TTS. This lack of analysis limits the understanding and practical use of TTS methods. In this paper, we focus on two core questions: (1) What is the optimal approach to scale test-time computation across different policy models, PRMs, and problem difficulty levels? (2) To what extent can extended computation improve the performance of LLMs on complex tasks, and can smaller language models outperform larger ones through this approach? Through comprehensive experiments on MATH-500 and challenging AIME24 tasks, we have the following observations: (1) The compute-optimal TTS strategy is highly dependent on the choice of policy model, PRM, and problem difficulty. (2) With our compute-optimal TTS strategy, extremely small policy models can outperform larger models. For example, a 1B LLM can exceed a 405B LLM on MATH-500. Moreover, on both MATH-500 and AIME24, a 0.5B LLM outperforms GPT-4o, a 3B LLM surpasses a 405B LLM, and a 7B LLM beats o1 and DeepSeek-R1, while with higher inference efficiency. These findings show the significance of adapting TTS strategies to the specific characteristics of each task and model and indicate that TTS is a promising approach for enhancing the reasoning abilities of LLMs.
2502.06705
RSAttAE: An Information-Aware Attention-based Autoencoder Recommender System
cs.LG cs.IR
Recommender systems play a crucial role in modern life, including information retrieval, the pharmaceutical industry, retail, and entertainment. The entertainment sector, in particular, attracts significant attention and generates substantial profits. This work proposes a new method for predicting unknown user-movie ratings to enhance customer satisfaction. To achieve this, we utilize the MovieLens 100K dataset. Our approach introduces an attention-based autoencoder to create meaningful representations and the XGBoost method for rating predictions. The results demonstrate that our proposal outperforms most of the existing state-of-the-art methods. Availability: github.com/ComputationIASBS/RecommSys
2502.06707
FinMamba: Market-Aware Graph Enhanced Multi-Level Mamba for Stock Movement Prediction
cs.CE
Recently, combining stock features with inter-stock correlations has become a common and effective approach for stock movement prediction. However, financial data presents significant challenges due to its low signal-to-noise ratio and the dynamic complexity of the market, which give rise to two key limitations in existing methods. First, the relationships between stocks are highly influenced by multifaceted factors including macroeconomic market dynamics, and current models fail to adaptively capture these evolving interactions under specific market conditions. Second, for the accuracy and timeliness required by real-world trading, existing financial data mining methods struggle to extract beneficial pattern-oriented dependencies from long historical data while maintaining high efficiency and low memory consumption. To address the limitations, we propose FinMamba, a Mamba-GNN-based framework for market-aware and multi-level hybrid stock movement prediction. Specifically, we devise a dynamic graph to learn the changing representations of inter-stock relationships by integrating a pruning module that adapts to market trends. Afterward, with a selective mechanism, the multi-level Mamba discards irrelevant information and resets states to skillfully recall historical patterns across multiple time scales with linear time costs, which are then jointly optimized for reliable prediction. Extensive experiments on U.S. and Chinese stock markets demonstrate the effectiveness of our proposed FinMamba, achieving state-of-the-art prediction accuracy and trading profitability, while maintaining low computational complexity. The code is available at https://github.com/TROUBADOUR000/FinMamba.
2502.06708
TEMSET-24K: Densely Annotated Dataset for Indexing Multipart Endoscopic Videos using Surgical Timeline Segmentation
cs.CV
Indexing endoscopic surgical videos is vital in surgical data science, forming the basis for systematic retrospective analysis and clinical performance evaluation. Despite its significance, current video analytics rely on manual indexing, a time-consuming process. Advances in computer vision, particularly deep learning, offer automation potential, yet progress is limited by the lack of publicly available, densely annotated surgical datasets. To address this, we present TEMSET-24K, an open-source dataset comprising 24,306 trans-anal endoscopic microsurgery (TEMS) video micro-clips. Each clip is meticulously annotated by clinical experts using a novel hierarchical labeling taxonomy encompassing phase, task, and action triplets, capturing intricate surgical workflows. To validate this dataset, we benchmarked deep learning models, including transformer-based architectures. Our in silico evaluation demonstrates high accuracy (up to 0.99) and F1 scores (up to 0.99) for key phases like Setup and Suturing. The STALNet model, tested with ConvNeXt, ViT, and SWIN V2 encoders, consistently segmented well-represented phases. TEMSET-24K provides a critical benchmark, propelling state-of-the-art solutions in surgical data science.
2502.06710
Learning Musical Representations for Music Performance Question Answering
cs.CV cs.MM cs.SD eess.AS
Music performances are representative scenarios for audio-visual modeling. Unlike common scenarios with sparse audio, music performances continuously involve dense audio signals throughout. While existing multimodal learning methods on the audio-video QA demonstrate impressive capabilities in general scenarios, they are incapable of dealing with fundamental problems within the music performances: they underexplore the interaction between the multimodal signals in performance and fail to consider the distinctive characteristics of instruments and music. Therefore, existing methods tend to answer questions regarding musical performances inaccurately. To bridge the above research gaps, (i) given the intricate multimodal interconnectivity inherent to music data, our primary backbone is designed to incorporate multimodal interactions within the context of music; (ii) to enable the model to learn music characteristics, we annotate and release rhythmic and music sources in the current music datasets; (iii) for time-aware audio-visual modeling, we align the model's music predictions with the temporal dimension. Our experiments show state-of-the-art effects on the Music AVQA datasets. Our code is available at https://github.com/xid32/Amuse.
2502.06715
HoneyComb: A Parallel Worst-Case Optimal Join on Multicores
cs.DB
To achieve true scalability on massive datasets, a modern query engine needs to be able to take advantage of large, shared-memory, multicore systems. Binary joins are conceptually easy to parallelize on a multicore system; however, several applications require a different approach to query evaluation, using a Worst-Case Optimal Join (WCOJ) algorithm. WCOJ is known to outperform traditional query plans for cyclic queries. However, there is no obvious adaptation of WCOJ to parallel architectures. The few existing systems that parallelize WCOJ do this by partitioning only the top variable of the WCOJ algorithm. This leads to work skew (since some relations end up being read entirely by every thread), possible contention between threads (when the hierarchical trie index is built lazily, which is the case on most recent WCOJ systems), and exacerbates the redundant computations already existing in WCOJ. We introduce HoneyComb, a parallel version of WCOJ, optimized for large multicore, shared-memory systems. HoneyComb partitions the domains of all query variables, not just that of the top loop. We adapt the partitioning idea from the HyperCube algorithm, developed by the theory community for computing multi-join queries on a massively parallel shared-nothing architecture, and introduce new methods for computing the shares, optimized for a shared-memory architecture. To avoid the contention created by the lazy construction of the trie-index, we introduce CoCo, a new and very simple index structure, which we build eagerly, by sorting the entire relation. Finally, in order to remove some of the redundant computations of WCOJ, we introduce a rewriting technique of the WCOJ plan that factors out some of these redundant computations. Our experimental evaluation compares HoneyComb with several recent implementations of WCOJ.