arXiv subjects

Weinan Zhang

Publications and source records attributed to Weinan Zhang.

At least 217 records · Page 12Linked to original sources

arXiv2025

DriveGen: Towards Infinite Diverse Traffic Scenarios with Large Models

Microscopic traffic simulation has become an important tool for autonomous driving training and testing. Although recent data-driven approaches advance realistic behavior generation, their learning still relies primarily on a single real-world dataset, which limits their diversity and thereby hinders downstream algorithm optimization. In this paper, we propose DriveGen, a novel traffic simulation framework with large models for more diverse traffic generation that supports further customized designs. DriveGen consists of two internal stages: the initialization stage uses large language model and retrieval technique to generate map and vehicle assets; the rollout stage outputs trajectories with selected waypoint goals from visual language model and a specific designed diffusion planner. Through this two-staged process, DriveGen fully utilizes large models' high-level cognition and reasoning of driving behavior, obtaining greater diversity beyond datasets while maintaining high realism. To support effective downstream optimization, we additionally develop DriveGen-CS, an automatic corner case generation pipeline that uses failures of the driving algorithm as additional prompt knowledge for large models without the need for retraining or fine-tuning. Experiments show that our generated scenarios and corner cases have a superior performance compared to state-of-the-art baselines. Downstream experiments further verify that the synthesized traffic of DriveGen provides better optimization of the performance of typical driving algorithms, demonstrating the effectiveness of our framework.

cs.AI↗

arXiv2025

Humanoid Whole-Body Locomotion on Narrow Terrain via Dynamic Balance and Reinforcement Learning

Humans possess delicate dynamic balance mechanisms that enable them to maintain stability across diverse terrains and under extreme conditions. However, despite significant advances recently, existing locomotion algorithms for humanoid robots are still struggle to traverse extreme environments, especially in cases that lack external perception (e.g., vision or LiDAR). This is because current methods often rely on gait-based or perception-condition rewards, lacking effective mechanisms to handle unobservable obstacles and sudden balance loss. To address this challenge, we propose a novel whole-body locomotion algorithm based on dynamic balance and Reinforcement Learning (RL) that enables humanoid robots to traverse extreme terrains, particularly narrow pathways and unexpected obstacles, using only proprioception. Specifically, we introduce a dynamic balance mechanism by leveraging an extended measure of Zero-Moment Point (ZMP)-driven rewards and task-driven rewards in a whole-body actor-critic framework, aiming to achieve coordinated actions of the upper and lower limbs for robust locomotion. Experiments conducted on a full-sized Unitree H1-2 robot verify the ability of our method to maintain balance on extremely narrow terrains and under external disturbances, demonstrating its effectiveness in enhancing the robot's adaptability to complex environments. The videos are given at https://whole-body-loco.github.io.

cs.RO↗

arXiv2025

Agentic Information Retrieval

Since the 1970s, information retrieval (IR) has long been defined as the process of acquiring relevant information items from a pre-defined corpus to satisfy user information needs. Traditional IR systems, while effective in domains like web search, are constrained by their reliance on static, pre-defined information items. To this end, this paper introduces agentic information retrieval (Agentic IR), a transformative next-generation paradigm for IR driven by large language models (LLMs) and AI agents. The central shift in agentic IR is the evolving definition of ``information'' from static, pre-defined information items to dynamic, context-dependent information states. Information state refers to a particular information context that the user is right in within a dynamic environment, encompassing not only the acquired information items but also real-time user preferences, contextual factors, and decision-making processes. In such a way, traditional information retrieval, focused on acquiring relevant information items based on user queries, can be naturally extended to achieving the target information state given the user instruction, which thereby defines the agentic information retrieval. We systematically discuss agentic IR from various aspects, i.e., task formulation, architecture, evaluation, case studies, as well as challenges and future prospects. We believe that the concept of agentic IR introduced in this paper not only broadens the scope of information retrieval research but also lays the foundation for a more adaptive, interactive, and intelligent next-generation IR paradigm.

cs.IR↗

arXiv2025

ThinkBench: Dynamic Out-of-Distribution Evaluation for Robust LLM Reasoning

Evaluating large language models (LLMs) poses significant challenges, particularly due to issues of data contamination and the leakage of correct answers. To address these challenges, we introduce ThinkBench, a novel evaluation framework designed to evaluate LLMs' reasoning capability robustly. ThinkBench proposes a dynamic data generation method for constructing out-of-distribution (OOD) datasets and offers an OOD dataset that contains 2,912 samples drawn from reasoning tasks. ThinkBench unifies the evaluation of reasoning models and non-reasoning models. We evaluate 16 LLMs and 4 PRMs under identical experimental conditions and show that most of the LLMs' performance are far from robust and they face a certain level of data leakage. By dynamically generating OOD datasets, ThinkBench effectively provides a reliable evaluation of LLMs and reduces the impact of data contamination.

cs.CL↗

arXiv2025

Score-Based Diffusion Policy Compatible with Reinforcement Learning via Optimal Transport

Diffusion policies have shown promise in learning complex behaviors from demonstrations, particularly for tasks requiring precise control and long-term planning. However, they face challenges in robustness when encountering distribution shifts. This paper explores improving diffusion-based imitation learning models through online interactions with the environment. We propose OTPR (Optimal Transport-guided score-based diffusion Policy for Reinforcement learning fine-tuning), a novel method that integrates diffusion policies with RL using optimal transport theory. OTPR leverages the Q-function as a transport cost and views the policy as an optimal transport map, enabling efficient and stable fine-tuning. Moreover, we introduce masked optimal transport to guide state-action matching using expert keypoints and a compatibility-based resampling strategy to enhance training stability. Experiments on three simulation tasks demonstrate OTPR's superior performance and robustness compared to existing methods, especially in complex and sparse-reward environments. In sum, OTPR provides an effective framework for combining IL and RL, achieving versatile and reliable policy learning. The code will be released at https://github.com/Sunmmyy/OTPR.git.

cs.LG↗

arXiv2025

Retrieval-Augmented Process Reward Model for Generalizable Mathematical Reasoning

While large language models (LLMs) have significantly advanced mathematical reasoning, Process Reward Models (PRMs) have been developed to evaluate the logical validity of reasoning steps. However, PRMs still struggle with out-of-distribution (OOD) challenges. This paper identifies key OOD issues, including step OOD, caused by differences in reasoning patterns across model types and sizes, and question OOD, which arises from dataset shifts between training data and real-world problems. To address these issues, we introduce Retrieval-Augmented Process Reward Model (RetrievalPRM), a novel framework designed to tackle these OOD issues. By utilizing a two-stage retrieval-enhanced mechanism, RetrievalPRM retrieves semantically similar questions and steps as a warmup, enhancing PRM's ability to evaluate target steps and improving generalization and reasoning consistency across different models and problem types. Our extensive experiments demonstrate that RetrievalPRM outperforms existing baselines across multiple real-world datasets. Our open-source contributions include a retrieval-enhanced dataset, a tuning framework for PRM training, and the RetrievalPRM model, establishing a new standard for PRM performance.

cs.AI↗

arXiv2025

Bursting Filter Bubble: Enhancing Serendipity Recommendations with Aligned Large Language Models

Recommender systems (RSs) often suffer from the feedback loop phenomenon, e.g., RSs are trained on data biased by their recommendations. This leads to the filter bubble effect that reinforces homogeneous content and reduces user satisfaction. To this end, serendipity recommendations, which offer unexpected yet relevant items, are proposed. Recently, large language models (LLMs) have shown potential in serendipity prediction due to their extensive world knowledge and reasoning capabilities. However, they still face challenges in aligning serendipity judgments with human assessments, handling long user behavior sequences, and meeting the latency requirements of industrial RSs. To address these issues, we propose SERAL (Serendipity Recommendations with Aligned Large Language Models), a framework comprising three stages: (1) Cognition Profile Generation to compress user behavior into multi-level profiles; (2) SerenGPT Alignment to align serendipity judgments with human preferences using enriched training data; and (3) Nearline Adaptation to integrate SerenGPT into industrial RSs pipelines efficiently. Online experiments demonstrate that SERAL improves exposure ratio (PVR), clicks, and transactions of serendipitous items by 5.7%, 29.56%, and 27.6%, enhancing user experience without much impact on overall revenue. Now, it has been fully deployed in the "Guess What You Like" of the Taobao App homepage.

cs.IR↗

arXiv2025

Boost, Disentangle, and Customize: A Robust System2-to-System1 Pipeline for Code Generation

Large language models (LLMs) have demonstrated remarkable capabilities in various domains, particularly in system 1 tasks, yet the intricacies of their problem-solving mechanisms in system 2 tasks are not sufficiently explored. Recent research on System2-to-System1 methods surge, exploring the System 2 reasoning knowledge via inference-time computation and compressing the explored knowledge into System 1 process. In this paper, we focus on code generation, which is a representative System 2 task, and identify two primary challenges: (1) the complex hidden reasoning processes and (2) the heterogeneous data distributions that complicate the exploration and training of robust LLM solvers. To tackle these issues, we propose a novel BDC framework that explores insightful System 2 knowledge of LLMs using a MC-Tree-Of-Agents algorithm with mutual \textbf{B}oosting, \textbf{D}isentangles the heterogeneous training data for composable LoRA-experts, and obtain \textbf{C}ustomized problem solver for each data instance with an input-aware hypernetwork to weight over the LoRA-experts, offering effectiveness, flexibility, and robustness. This framework leverages multiple LLMs through mutual verification and boosting, integrated into a Monte-Carlo Tree Search process enhanced by reflection-based pruning and refinement. Additionally, we introduce the DisenLora algorithm, which clusters heterogeneous data to fine-tune LLMs into composable Lora experts, enabling the adaptive generation of customized problem solvers through an input-aware hypernetwork. This work lays the groundwork for advancing LLM capabilities in complex reasoning tasks, offering a novel System2-to-System1 solution.

cs.AI↗

arXiv2025

RHINO: Learning Real-Time Humanoid-Human-Object Interaction from Human Demonstrations

Humanoid robots have shown success in locomotion and manipulation. Despite these basic abilities, humanoids are still required to quickly understand human instructions and react based on human interaction signals to become valuable assistants in human daily life. Unfortunately, most existing works only focus on multi-stage interactions, treating each task separately, and neglecting real-time feedback. In this work, we aim to empower humanoid robots with real-time reaction abilities to achieve various tasks, allowing human to interrupt robots at any time, and making robots respond to humans immediately. To support such abilities, we propose a general humanoid-human-object interaction framework, named RHINO, i.e., Real-time Humanoid-human Interaction and Object manipulation. RHINO provides a unified view of reactive motion, instruction-based manipulation, and safety concerns, over multiple human signal modalities, such as languages, images, and motions. RHINO is a hierarchical learning framework, enabling humanoids to learn reaction skills from human-human-object demonstrations and teleoperation data. In particular, it decouples the interaction process into two levels: 1) a high-level planner inferring human intentions from real-time human behaviors; and 2) a low-level controller achieving reactive motion behaviors and object manipulation skills based on the predicted intentions. We evaluate the proposed framework on a real humanoid robot and demonstrate its effectiveness, flexibility, and safety in various scenarios.

cs.RO↗

arXiv2025

HammerBench: Fine-Grained Function-Calling Evaluation in Real Mobile Device Scenarios

Evaluating the performance of LLMs in multi-turn human-agent interactions presents significant challenges, particularly due to the complexity and variability of user behavior. In this paper, we introduce HammerBench, a novel benchmark framework for assessing LLMs' function-calling capabilities in real-world, multi-turn dialogues. HammerBench simulates diverse mobile assistant use cases, incorporating imperfect instructions, dynamic question-answer trajectories, intent and argument shifts, and the indirect use of external information through pronouns. To construct this benchmark, we curate a comprehensive dataset derived from popular mobile app functionalities and anonymized user logs, complemented by a cost-effective data generation pipeline leveraging open-source models. HammerBench is further augmented with fine-grained interaction snapshots and metrics, enabling detailed evaluation of function-calling performance across individual conversational turns. We demonstrate the effectiveness of HammerBench by evaluating several leading LLMs and uncovering key performance trends. Our experiments reveal that different types of parameter name errors are a significant source of failure across different interaction scenarios, highlighting critical areas for further improvement in LLM robustness for mobile assistant applications.

cs.CL↗

arXiv2025

Affine $\imath$quantum groups and twisted Yangians in Drinfeld presentations

We formulate a family of algebras, twisted Yangians (of split type) in current generators and relations, via a degeneration of the Drinfeld presentation of affine $\imath$quantum groups (associated with split Satake diagrams). These new algebras admit PBW type bases and are shown to be a deformation of twisted current algebras; presentations for twisted current algebras are also provided. For type AI, it matches with the Drinfeld presentation of twisted Yangian obtained via Gauss decomposition. We conjecture that our split twisted Yangians are isomorphic to the corresponding ones in RTT presentation.

math.QA↗

arXiv2025

Beyond Prior Limits: Addressing Distribution Misalignment in Particle Filtering

Particle filtering is a Bayesian inference method and a fundamental tool in state estimation for dynamic systems, but its effectiveness is often limited by the constraints of the initial prior distribution, a phenomenon we define as the Prior Boundary Phenomenon. This challenge arises when target states lie outside the prior's support, rendering traditional particle filtering methods inadequate for accurate estimation. Although techniques like unbounded priors and larger particle sets have been proposed, they remain computationally prohibitive and lack adaptability in dynamic scenarios. To systematically overcome these limitations, we propose the Diffusion-Enhanced Particle Filtering Framework, which introduces three key innovations: adaptive diffusion through exploratory particles, entropy-driven regularisation to prevent weight collapse, and kernel-based perturbations for dynamic support expansion. These mechanisms collectively enable particle filtering to explore beyond prior boundaries, ensuring robust state estimation for out-of-boundary targets. Theoretical analysis and extensive experiments validate framework's effectiveness, indicating significant improvements in success rates and estimation accuracy across high-dimensional and non-convex scenarios.

stat.ML↗

arXiv2025

Full-Stack Optimized Large Language Models for Lifelong Sequential Behavior Comprehension in Recommendation

In this paper, we address the lifelong sequential behavior incomprehension problem in large language models (LLMs) for recommendation, where LLMs struggle to extract useful information from long user behavior sequences, even within their context limits. To tackle this, we propose ReLLaX (Retrieval-enhanced Large Language models Plus), a framework offering optimization across data, prompt, and parameter levels. At the data level, we introduce Semantic User Behavior Retrieval (SUBR) to reduce sequence heterogeneity, making it easier for LLMs to extract key information. For prompt-level enhancement, we employ Soft Prompt Augmentation (SPA) to inject collaborative knowledge, aligning item representations with recommendation tasks and improving LLMs's exploration of item relationships. Finally, at the parameter level, we propose Component Fully-interactive LoRA (CFLoRA), which enhances LoRA's expressiveness by enabling interactions between its components, allowing better capture of sequential information. Moreover, we present new perspectives to compare current LoRA-based LLM4Rec methods, i.e. from both a composite and a decomposed view. We theoretically demonstrate that the ways they employ LoRA for recommendation are degraded versions of our CFLoRA, with different constraints on atom component interactions. Extensive experiments on three public datasets demonstrate ReLLaX's superiority over existing baselines and its ability to mitigate lifelong sequential behavior incomprehension effectively.

cs.IR↗

arXiv2025

A Survey on LLM-based Multi-Agent System: Recent Advances and New Frontiers in Application

LLM-based Multi-Agent Systems ( LLM-MAS ) have become a research hotspot since the rise of large language models (LLMs). However, with the continuous influx of new related works, the existing reviews struggle to capture them comprehensively. This paper presents a comprehensive survey of these studies. We first discuss the definition of LLM-MAS, a framework encompassing much of previous work. We provide an overview of the various applications of LLM-MAS in (i) solving complex tasks, (ii) simulating specific scenarios, and (iii) evaluating generative agents. Building on previous studies, we also highlight several challenges and propose future directions for research in this field.

cs.CL↗

arXiv2025

MADiff: Offline Multi-agent Learning with Diffusion Models

Offline reinforcement learning (RL) aims to learn policies from pre-existing datasets without further interactions, making it a challenging task. Q-learning algorithms struggle with extrapolation errors in offline settings, while supervised learning methods are constrained by model expressiveness. Recently, diffusion models (DMs) have shown promise in overcoming these limitations in single-agent learning, but their application in multi-agent scenarios remains unclear. Generating trajectories for each agent with independent DMs may impede coordination, while concatenating all agents' information can lead to low sample efficiency. Accordingly, we propose MADiff, which is realized with an attention-based diffusion model to model the complex coordination among behaviors of multiple agents. To our knowledge, MADiff is the first diffusion-based multi-agent learning framework, functioning as both a decentralized policy and a centralized controller. During decentralized executions, MADiff simultaneously performs teammate modeling, and the centralized controller can also be applied in multi-agent trajectory predictions. Our experiments demonstrate that MADiff outperforms baseline algorithms across various multi-agent learning tasks, highlighting its effectiveness in modeling complex multi-agent interactions. Our code is available at https://github.com/zbzhu99/madiff.

cs.AI↗

arXiv2024

LLM-based Multi-Agent Systems: Techniques and Business Perspectives

In the era of (multi-modal) large language models, most operational processes can be reformulated and reproduced using LLM agents. The LLM agents can perceive, control, and get feedback from the environment so as to accomplish the given tasks in an autonomous manner. Besides the environment-interaction property, the LLM agents can call various external tools to ease the task completion process. The tools can be regarded as a predefined operational process with private or real-time knowledge that does not exist in the parameters of LLMs. As a natural trend of development, the tools for calling are becoming autonomous agents, thus the full intelligent system turns out to be a LLM-based Multi-Agent System (LaMAS). Compared to the previous single-LLM-agent system, LaMAS has the advantages of i) dynamic task decomposition and organic specialization, ii) higher flexibility for system changing, iii) proprietary data preserving for each participating entity, and iv) feasibility of monetization for each entity. This paper discusses the technical and business landscapes of LaMAS. To support the ecosystem of LaMAS, we provide a preliminary version of such LaMAS protocol considering technical requirements, data privacy, and business incentives. As such, LaMAS would be a practical solution to achieve artificial collective intelligence in the near future.

cs.AI↗

arXiv2024

Diffusion-based Reinforcement Learning via Q-weighted Variational Policy Optimization

Diffusion models have garnered widespread attention in Reinforcement Learning (RL) for their powerful expressiveness and multimodality. It has been verified that utilizing diffusion policies can significantly improve the performance of RL algorithms in continuous control tasks by overcoming the limitations of unimodal policies, such as Gaussian policies, and providing the agent with enhanced exploration capabilities. However, existing works mainly focus on the application of diffusion policies in offline RL, while their incorporation into online RL is less investigated. The training objective of the diffusion model, known as the variational lower bound, cannot be optimized directly in online RL due to the unavailability of 'good' actions. This leads to difficulties in conducting diffusion policy improvement. To overcome this, we propose a novel model-free diffusion-based online RL algorithm, Q-weighted Variational Policy Optimization (QVPO). Specifically, we introduce the Q-weighted variational loss, which can be proved to be a tight lower bound of the policy objective in online RL under certain conditions. To fulfill these conditions, the Q-weight transformation functions are introduced for general scenarios. Additionally, to further enhance the exploration capability of the diffusion policy, we design a special entropy regularization term. We also develop an efficient behavior policy to enhance sample efficiency by reducing the variance of the diffusion policy during online interactions. Consequently, the QVPO algorithm leverages the exploration capabilities and multimodality of diffusion policies, preventing the RL agent from converging to a sub-optimal policy. To verify the effectiveness of QVPO, we conduct comprehensive experiments on MuJoCo benchmarks. The final results demonstrate that QVPO achieves state-of-the-art performance on both cumulative reward and sample efficiency.

cs.LG↗

arXiv2024

Beyond Graph Convolution: Multimodal Recommendation with Topology-aware MLPs

Given the large volume of side information from different modalities, multimodal recommender systems have become increasingly vital, as they exploit richer semantic information beyond user-item interactions. Recent works highlight that leveraging Graph Convolutional Networks (GCNs) to explicitly model multimodal item-item relations can significantly enhance recommendation performance. However, due to the inherent over-smoothing issue of GCNs, existing models benefit only from shallow GCNs with limited representation power. This drawback is especially pronounced when facing complex and high-dimensional patterns such as multimodal data, as it requires large-capacity models to accommodate complicated correlations. To this end, in this paper, we investigate bypassing GCNs when modeling multimodal item-item relationship. More specifically, we propose a Topology-aware Multi-Layer Perceptron (TMLP), which uses MLPs instead of GCNs to model the relationships between items. TMLP enhances MLPs with topological pruning to denoise item-item relations and intra (inter)-modality learning to integrate higher-order modality correlations. Extensive experiments on three real-world datasets verify TMLP's superiority over nine baselines. We also find that by discarding the internal message passing in GCNs, which is sensitive to node connections, TMLP achieves significant improvements in both training efficiency and robustness against existing models.

cs.IR↗