arXiv subjects

Lei Wang

Publications and source records attributed to Lei Wang.

At least 469 records · Page 26Linked to original sources

arXiv2025

Linear Convergence of Distributed Compressed Optimization with Equality Constraints

In this paper, the distributed strongly convex optimization problem is studied with spatio-temporal compressed communication and equality constraints. For the case where each agent holds an distributed local equality constraint, a distributed saddle-point algorithm is proposed by employing distributed filters to derive errors of the transmitted states for spatio-temporal compression purposes. It is shown that the resulting distributed compressed algorithm achieves linear convergence. Furthermore, the algorithm is generalized to the case where each agent holds a portion of the global equality constraint, i.e., the constraints across agents are coupled. By introducing an additional design freedom, the global equality constraint is shown to be equivalent to the one where each agent holds an equality constraint, for which the proposed distributed compressed saddle-point algorithm can be adapted to achieve linear convergence. Numerical simulations are adopted to validate the effectiveness of the proposed algorithms.

eess.SY↗

arXiv2025

Probing a light long-lived pseudo-scalar from Higgs decay via displaced taus at the LHC

A light (GeV mass) long-lived ($cτ$ around dozens of millimeters) CP-odd scalar can be readily predicted in new physics models. In this work we investigate the Higgs decay into such a light scalar plus a $Z$-boson and take the aligned two-Higgs-doublet model (2HDM) as an example. This light long-lived scalar, with the dominant decay to tau leptons, will fly over a distance from the production point and present a displaced vertex in an Inner Detector of a generally purposed experiment like ATLAS or CMS. In our study we focus on the LHC experiment and perform Monte Carlo simulations for the signal and backgrounds. We demonstrate some benchmark points for the aligned 2HDM and find the signal to be detectable when the luminosity is accumulated to 300 fb$^{-1}$. So our study suggests an experimental search for this process in the ongoing LHC.

hep-ph↗

arXiv2025

Improving Retrospective Language Agents via Joint Policy Gradient Optimization

In recent research advancements within the community, large language models (LLMs) have sparked great interest in creating autonomous agents. However, current prompt-based agents often heavily rely on large-scale LLMs. Meanwhile, although fine-tuning methods significantly enhance the capabilities of smaller LLMs, the fine-tuned agents often lack the potential for self-reflection and self-improvement. To address these challenges, we introduce a novel agent framework named RetroAct, which is a framework that jointly optimizes both task-planning and self-reflective evolution capabilities in language agents. Specifically, we develop a two-stage joint optimization process that integrates imitation learning and reinforcement learning, and design an off-policy joint policy gradient optimization algorithm with imitation learning regularization to enhance the data efficiency and training stability in agent tasks. RetroAct significantly improves the performance of open-source models, reduces dependency on closed-source LLMs, and enables fine-tuned agents to learn and evolve continuously. We conduct extensive experiments across various testing environments, demonstrating RetroAct has substantial improvements in task performance and decision-making processes.

cs.CL↗

arXiv2025

A Survey on Large Language Model based Autonomous Agents

Autonomous agents have long been a prominent research focus in both academic and industry communities. Previous research in this field often focuses on training agents with limited knowledge within isolated environments, which diverges significantly from human learning processes, and thus makes the agents hard to achieve human-like decisions. Recently, through the acquisition of vast amounts of web knowledge, large language models (LLMs) have demonstrated remarkable potential in achieving human-level intelligence. This has sparked an upsurge in studies investigating LLM-based autonomous agents. In this paper, we present a comprehensive survey of these studies, delivering a systematic review of the field of LLM-based autonomous agents from a holistic perspective. More specifically, we first discuss the construction of LLM-based autonomous agents, for which we propose a unified framework that encompasses a majority of the previous work. Then, we present a comprehensive overview of the diverse applications of LLM-based autonomous agents in the fields of social science, natural science, and engineering. Finally, we delve into the evaluation strategies commonly used for LLM-based autonomous agents. Based on the previous studies, we also present several challenges and future directions in this field. To keep track of this field and continuously update our survey, we maintain a repository of relevant references at https://github.com/Paitesanshi/LLM-Agent-Survey.

cs.AI↗

arXiv2025

Learnable Expansion of Graph Operators for Multi-Modal Feature Fusion

In computer vision tasks, features often come from diverse representations, domains (e.g., indoor and outdoor), and modalities (e.g., text, images, and videos). Effectively fusing these features is essential for robust performance, especially with the availability of powerful pre-trained models like vision-language models. However, common fusion methods, such as concatenation, element-wise operations, and non-linear techniques, often fail to capture structural relationships, deep feature interactions, and suffer from inefficiency or misalignment of features across domains or modalities. In this paper, we shift from high-dimensional feature space to a lower-dimensional, interpretable graph space by constructing relationship graphs that encode feature relationships at different levels, e.g., clip, frame, patch, token, etc. To capture deeper interactions, we expand graphs through iterative graph relationship updates and introduce a learnable graph fusion operator to integrate these expanded relationships for more effective fusion. Our approach is relationship-centric, operates in a homogeneous space, and is mathematically principled, resembling element-wise relationship score aggregation via multilinear polynomials. We demonstrate the effectiveness of our graph-based fusion method on video anomaly detection, showing strong performance across multi-representational, multi-modal, and multi-domain feature fusion tasks.

cs.CV↗

arXiv2025

ThinK: Thinner Key Cache by Query-Driven Pruning

Large Language Models (LLMs) have revolutionized the field of natural language processing, achieving unprecedented performance across a variety of applications. However, their increased computational and memory demands present significant challenges, especially when handling long sequences. This paper focuses on the long-context scenario, addressing the inefficiencies in KV cache memory consumption during inference. Unlike existing approaches that optimize the memory based on the sequence length, we identify substantial redundancy in the channel dimension of the KV cache, as indicated by an uneven magnitude distribution and a low-rank structure in the attention weights. In response, we propose ThinK, a novel query-dependent KV cache pruning method designed to minimize attention weight loss while selectively pruning the least significant channels. Our approach not only maintains or enhances model accuracy but also achieves a reduction in KV cache memory costs by over 20% compared with vanilla KV cache eviction and quantization methods. For instance, ThinK integrated with KIVI can achieve a 2.8x reduction in peak memory usage while maintaining nearly the same quality, enabling up to a 5x increase in batch size when using a single GPU. Extensive evaluations on the LLaMA and Mistral models across various long-sequence datasets verified the efficiency of ThinK, establishing a new baseline algorithm for efficient LLM deployment without compromising performance. Our code has been made available at https://github.com/SalesforceAIResearch/ThinK.

cs.CL↗

arXiv2025

Enhancing Hepatopathy Clinical Trial Efficiency: A Secure, Large Language Model-Powered Pre-Screening Pipeline

Background: Recruitment for cohorts involving complex liver diseases, such as hepatocellular carcinoma and liver cirrhosis, often requires interpreting semantically complex criteria. Traditional manual screening methods are time-consuming and prone to errors. While AI-powered pre-screening offers potential solutions, challenges remain regarding accuracy, efficiency, and data privacy. Methods: We developed a novel patient pre-screening pipeline that leverages clinical expertise to guide the precise, safe, and efficient application of large language models. The pipeline breaks down complex criteria into a series of composite questions and then employs two strategies to perform semantic question-answering through electronic health records - (1) Pathway A, Anthropomorphized Experts' Chain of Thought strategy, and (2) Pathway B, Preset Stances within an Agent Collaboration strategy, particularly in managing complex clinical reasoning scenarios. The pipeline is evaluated on three key metrics-precision, time consumption, and counterfactual inference - at both the question and criterion levels. Results: Our pipeline achieved high precision (0.921, in criteria level) and efficiency (0.44s per task). Pathway B excelled in complex reasoning, while Pathway A was effective in precise data extraction with faster processing times. Both pathways achieved comparable precision. The pipeline showed promising results in hepatocellular carcinoma (0.878) and cirrhosis trials (0.843). Conclusions: This data-secure and time-efficient pipeline shows high precision in hepatopathy trials, providing promising solutions for streamlining clinical trial workflows. Its efficiency and adaptability make it suitable for improving patient recruitment. And its capability to function in resource-constrained environments further enhances its utility in clinical settings.

cs.AI↗

arXiv2025

CrystalFlow: A Flow-Based Generative Model for Crystalline Materials

Deep learning-based generative models have emerged as powerful tools for modeling complex data distributions and generating high-fidelity samples, offering a transformative approach to efficiently explore the configuration space of crystalline materials. In this work, we present CrystalFlow, a flow-based generative model specifically developed for the generation of crystalline materials. CrystalFlow constructs Continuous Normalizing Flows to model lattice parameters, atomic coordinates, and/or atom types, which are trained using Conditional Flow Matching techniques. Through an appropriate choice of data representation and the integration of a graph-based equivariant neural network, the model effectively captures the fundamental symmetries of crystalline materials, which ensures data-efficient learning and enables high-quality sampling. Our experiments demonstrate that CrystalFlow achieves state-of-the-art performance across standard generation benchmarks, and exhibits versatile conditional generation capabilities including producing structures optimized for specific external pressures or desired material properties. These features highlight the model's potential to address realistic crystal structure prediction challenges, offering a robust and efficient framework for advancing data-driven research in condensed matter physics and material science.

cond-mat.mtrl-sci↗

arXiv2025

Evolving Skeletons: Motion Dynamics in Action Recognition

Skeleton-based action recognition has gained significant attention for its ability to efficiently represent spatiotemporal information in a lightweight format. Most existing approaches use graph-based models to process skeleton sequences, where each pose is represented as a skeletal graph structured around human physical connectivity. Among these, the Spatiotemporal Graph Convolutional Network (ST-GCN) has become a widely used framework. Alternatively, hypergraph-based models, such as the Hyperformer, capture higher-order correlations, offering a more expressive representation of complex joint interactions. A recent advancement, termed Taylor Videos, introduces motion-enhanced skeleton sequences by embedding motion concepts, providing a fresh perspective on interpreting human actions in skeleton-based action recognition. In this paper, we conduct a comprehensive evaluation of both traditional skeleton sequences and Taylor-transformed skeletons using ST-GCN and Hyperformer models on the NTU-60 and NTU-120 datasets. We compare skeletal graph and hypergraph representations, analyzing static poses against motion-injected poses. Our findings highlight the strengths and limitations of Taylor-transformed skeletons, demonstrating their potential to enhance motion dynamics while exposing current challenges in fully using their benefits. This study underscores the need for innovative skeletal modelling techniques to effectively handle motion-rich data and advance the field of action recognition.

cs.CV↗

arXiv2025

Do Language Models Understand Time?

Large language models (LLMs) have revolutionized video-based computer vision applications, including action recognition, anomaly detection, and video summarization. Videos inherently pose unique challenges, combining spatial complexity with temporal dynamics that are absent in static images or textual data. Current approaches to video understanding with LLMs often rely on pretrained video encoders to extract spatiotemporal features and text encoders to capture semantic meaning. These representations are integrated within LLM frameworks, enabling multimodal reasoning across diverse video tasks. However, the critical question persists: Can LLMs truly understand the concept of time, and how effectively can they reason about temporal relationships in videos? This work critically examines the role of LLMs in video processing, with a specific focus on their temporal reasoning capabilities. We identify key limitations in the interaction between LLMs and pretrained encoders, revealing gaps in their ability to model long-term dependencies and abstract temporal concepts such as causality and event progression. Furthermore, we analyze challenges posed by existing video datasets, including biases, lack of temporal annotations, and domain-specific limitations that constrain the temporal understanding of LLMs. To address these gaps, we explore promising future directions, including the co-evolution of LLMs and encoders, the development of enriched datasets with explicit temporal labels, and innovative architectures for integrating spatial, temporal, and semantic reasoning. By addressing these challenges, we aim to advance the temporal comprehension of LLMs, unlocking their full potential in video analysis and beyond. Our paper's GitHub repository can be found at https://github.com/Darcyddx/Video-LLM.

cs.CV↗

arXiv2025

Pay Attention to What You Need

Although large language models (LLMs) have achieved significant success in natural language processing, they still struggle with long-context comprehension. Traditional approaches to mitigating this issue typically rely on fine-tuning or retraining, which is both resource-intensive and challenging to deploy in lightweight industrial settings. In this paper, we investigate the potential to accomplish this without any additional resources. Through an in-depth study of the attention mechanism in LLMs, we propose a method called Scaled ReAttention (SRA) to strengthen LLMs' ability to interpret and retrieve information by strategically manipulating their attention scores during inference. Through extensive experiments, we demonstrate that integrating SRA significantly boosts LLMs' performance on a variety of downstream tasks, highlighting its practical potential for enhancing language understanding without incurring the overhead of traditional training.

cs.CL↗

arXiv2025

Non-Abelian interlayer coherent fractional quantum Hall states

We study non-Abelian fractional quantum Hall state in double layer systems at total filling factor $1/2$. Recent progresses in two-dimensional van der Waals materials made it possible to explore the regime with very small interlayer distance. Numerical calculations suggests interlayer phase coherence can develop between the layers such that the electrons may redistribute between them without changing the Hall response. It corresponds to spontaneous breaking of the U(1) symmetry associated with the particle number difference in the layers. This state manifests itself as superfluid in counterflow measurement and has characteristic Hall response when current is passed through one layer and voltages in both layers are measured. As the interlayer distance increases, a phase transition to the Halperin 331 state occurs. We also discuss similar physics for bosonic systems with specially designed interactions.

cond-mat.str-el↗

arXiv2025

AttentionEngine: A Versatile Framework for Efficient Attention Mechanisms on Diverse Hardware Platforms

Transformers and large language models (LLMs) have revolutionized machine learning, with attention mechanisms at the core of their success. As the landscape of attention variants expands, so too do the challenges of optimizing their performance, particularly across different hardware platforms. Current optimization strategies are often narrowly focused, requiring extensive manual intervention to accommodate changes in model configurations or hardware environments. In this paper, we introduce AttentionEngine, a comprehensive framework designed to streamline the optimization of attention mechanisms across heterogeneous hardware backends. By decomposing attention computation into modular operations with customizable components, AttentionEngine enables flexible adaptation to diverse algorithmic requirements. The framework further automates kernel optimization through a combination of programmable templates and a robust cross-platform scheduling strategy. Empirical results reveal performance gains of up to 10x on configurations beyond the reach of existing methods. AttentionEngine offers a scalable, efficient foundation for developing and deploying attention mechanisms with minimal manual tuning. Our code has been open-sourced and is available at https://github.com/microsoft/AttentionEngine.

cs.CL↗

arXiv2025

Thermal Broadening of Phonon Spectral Function in Classical Lattice Models: Projective Truncation Approximation

Thermal broadening of the quasi-particle peak in the spectral function is an important physical feature in many statistical systems, but it is difficult to calculate. To tackle this problem, we propose the $H$-expanded basis within the projective truncation approximation (PTA) of the Green's function equation of motion. A zeros-removing technique is introduced to stabilize the iterative solution of the PTA equations. Benchmarking calculations on the classical one-variable anharmonic oscillator model and the one-dimensional $ϕ^4$ lattice model show that the thermal broadened quasi-particle peak in the spectral function can be produced on a semi-quantitative level. Using this method, we discuss the low- and high- temperature power-law behaviors of the spectral width $Γ_k(T)$ of the one-dimensional $ϕ^4$ model, finding it in contradiction with the assumption of effective phonon theory. A short-chain limit of this model is also discovered. Issues of extending the $H$-expanded basis to quantum systems and of the applicability of the Debye formula for thermal conductivity are discussed.

cond-mat.str-el↗

arXiv2025

Reward Models Identify Consistency, Not Causality

Reward models (RMs) play a crucial role in aligning large language models (LLMs) with human preferences and enhancing reasoning quality. Traditionally, RMs are trained to rank candidate outputs based on their correctness and coherence. However, in this work, we present several surprising findings that challenge common assumptions about RM behavior. Our analysis reveals that state-of-the-art reward models prioritize structural consistency over causal correctness. Specifically, removing the problem statement has minimal impact on reward scores, whereas altering numerical values or disrupting the reasoning flow significantly affects RM outputs. Furthermore, RMs exhibit a strong dependence on complete reasoning trajectories truncated or incomplete steps lead to significant variations in reward assignments, indicating that RMs primarily rely on learned reasoning patterns rather than explicit problem comprehension. These findings hold across multiple architectures, datasets, and tasks, leading to three key insights: (1) RMs primarily assess coherence rather than true reasoning quality; (2) The role of explicit problem comprehension in reward assignment is overstated; (3) Current RMs may be more effective at ranking responses than verifying logical validity. Our results suggest a fundamental limitation in existing reward modeling approaches, emphasizing the need for a shift toward causality-aware reward models that go beyond consistency-driven evaluation.

cs.LG↗

arXiv2025

Extracting the Luttinger parameter from a single wave function

The low-energy physics of Tomonaga-Luttinger liquids (TLLs) is controlled by the Luttinger parameter. We demonstrate that this parameter can be extracted from a single wave function for one-component TLLs with periodic boundary condition. This method relies on the fact that TLLs are described by conformal field theory in which crosscap states can be constructed. The overlaps between the crosscap states and the ground state as well as some excited states are proved to be universal numbers that directly reveal the Luttinger parameter. In microscopic lattice models, crosscap states are formed by putting each pair of antipodal sites into a maximally entangled state. Analytical and numerical calculations are performed in a few representative models to substantiate the conformal field theory prediction. The extracted Luttinger parameters are generally quite accurate in finite-size systems with moderate lengths, so there is no need to perform data fitting and/or finite-size scaling.

cond-mat.str-el↗

arXiv2025

PRIV-QA: Privacy-Preserving Question Answering for Cloud Large Language Models

The rapid development of large language models (LLMs) is redefining the landscape of human-computer interaction, and their integration into various user-service applications is becoming increasingly prevalent. However, transmitting user data to cloud-based LLMs presents significant risks of data breaches and unauthorized access to personal identification information. In this paper, we propose a privacy preservation pipeline for protecting privacy and sensitive information during interactions between users and LLMs in practical LLM usage scenarios. We construct SensitiveQA, the first privacy open-ended question-answering dataset. It comprises 57k interactions in Chinese and English, encompassing a diverse range of user-sensitive information within the conversations. Our proposed solution employs a multi-stage strategy aimed at preemptively securing user information while simultaneously preserving the response quality of cloud-based LLMs. Experimental validation underscores our method's efficacy in balancing privacy protection with maintaining robust interaction quality. The code and dataset are available at https://github.com/ligw1998/PRIV-QA.

cs.CL↗

arXiv2025

SessionRec: Next Session Prediction Paradigm For Generative Sequential Recommendation

We introduce SessionRec, a novel next-session prediction paradigm (NSPP) for generative sequential recommendation, addressing the fundamental misalignment between conventional next-item prediction paradigm (NIPP) and real-world recommendation scenarios. Unlike NIPP's item-level autoregressive generation that contradicts actual session-based user interactions, our framework introduces a session-aware representation learning through hierarchical sequence aggregation (intra/inter-session), reducing attention computation complexity while enabling implicit modeling of massive negative interactions, and a session-based prediction objective that better captures users' diverse interests through multi-item recommendation in next sessions. Moreover, we found that incorporating a rank loss for items within the session under the next session prediction paradigm can significantly improve the ranking effectiveness of generative sequence recommendation models. We also verified that SessionRec exhibits clear power-law scaling laws similar to those observed in LLMs. Extensive experiments conducted on public datasets and online A/B test in Meituan App demonstrate the effectiveness of SessionRec. The proposed paradigm establishes new foundations for developing industrial-scale generative recommendation systems through its model-agnostic architecture and computational efficiency.

cs.IR↗