Searcharxiv⌕ Search

arXiv subjects

Shengqin Wang

Publications and source records attributed to Shengqin Wang.

7 recordsLinked to original sources

MRC-MMAgent: Reinforcement Learning for Multi-Role Collaboration in Multimodal Tool-Using Agents

Multimodal tool-using agents augment vision-language models with access to external information and capabilities, enabling them to solve knowledge-intensive and visually complex tasks that cannot be addressed using visual context alone. However, existing reinforcement learning pipelines largely train single-role agents to search and answer using terminal rewards, overlooking how humans inspect retrieved evidence, reformulate queries, and revise conclusions during search. Consequently, they fail to optimize this self-corrective process over complete interaction trajectories. We propose MRC-MMAgent, a multi-role collaborative reinforcement learning framework that incorporates trajectory assessment and continuation directly into the executable agent loop. First, the framework establishes a multi-role collaborative loop that integrates solving, inspection, and continuation, enabling the agent to acquire evidence through tools, evaluate the preceding trajectory, and resume problem solving when the current answer is insufficiently supported. Second, we introduce a role-coupled learning objective that combines task success, interaction validity, and successful correction with full-trajectory policy optimization across the decisions generated by all roles. Extensive experiments show that, using only the original baseline training set without constructing additional datasets, our method consistently outperforms the corresponding base models and achieves leading or highly competitive performance across benchmarks.

cs.CV↗

ECOD: Experience-Calibrated On-Policy Distillation for Multi-Turn Agents

Sparse and delayed rewards make reinforcement learning challenging for long-horizon language agents. On-policy distillation (OPD) complements environmental rewards with dense teacher feedback on student-generated contexts. Standard OPD can misalign supervision with relative teacher competence, student trajectory outcomes, and the temporal structure of task execution. Recent variants adjust guidance through divergence-based weighting and temporal schedules, while joint calibration from task-completion experience remains underexplored. We introduce Experience-Calibrated On-Policy Distillation (ECOD), which allocates teacher supervision at the task, trajectory, and turn levels. ECOD compares teacher and student success-rate posteriors under matched task conditions, restricts direct supervision to failed student trajectories, and derives a temporal prior from successful teacher lengths. The resulting weight applies to teacher losses or teacher-derived advantages. We evaluate ECOD on ALFWorld and WebShop with Qwen2.5-1.5B-Instruct and Qwen2.5-7B-Instruct. Under matched teacher settings, ECOD improves standard OPD by up to 4.06% in relative ALFWorld unseen success. Across both model scales, ECOD-assisted SOD, TCOD-F2B, and ATOD configurations yield average paired relative gains of 6.20% on ALFWorld unseen and 7.05% on WebShop, with WebShop gains ranging from 2.29% to 21.60%. These results support experience-calibrated supervision allocation as a reusable mechanism across distinct distillation objectives.

cs.LG↗

TMCS: Tool-Grounded Multi-Agent Reasoning for Compositional Chemical Problem Solving

Despite the promise of Large Language Models (LLMs) in computational chemistry, rigorous combinatorial chemistry problems remain difficult because they require quantitatively constrained molecular modification, candidate validation, and systematic revision after failed attempts. Existing tool-augmented chemical agents demonstrate useful planning and tool use, but they rarely provide a unified loop for property-driven molecular optimization and workflow-level composition. To bridge this gap, we propose Tool-Grounded Multi-Agent Reasoning for Compositional Chemical Problem Solving (TMCS), a step-by-step multi-agent framework that formalizes chemical problem solving as an interpretable, tool-augmented workflow. At the task level, specialized agents leverage external tools, few-shot trajectory memory, and structured reflection to iteratively refine solutions. At the workflow level, TMCS chains generation, understanding, editing, description, and optimization into a closed-loop pipeline. Evaluations across multiple chemical tasks demonstrate that TMCS consistently enhances chemical reasoning across both open- and closed-source base models, achieving state-of-the-art performance.

cs.AI↗

ProMMSearchAgent: A Generalizable Multimodal Search Agent Trained with Process-Oriented Rewards

Training multimodal agents via reinforcement learning for knowledge-intensive visual reasoning is fundamentally hindered by the extreme sparsity of outcome-based supervision and the unpredictability of live web environments. To resolve these algorithmic and environmental bottlenecks, we introduce ProMMSearchAgent, establishing a novel Sim-to-Real training paradigm for multimodal search. We decouple policy learning into a deterministic, local static sandbox. Crucially, to learn effectively within this constrained environment, we propose an introspective process-oriented reward. By probing the agent's own parametric knowledge boundaries, we generate dense behavioral metadata that explicitly rewards the correct cognitive decision, initiating a multimodal or text search only when visually or factually uncertain. Extensive experiments demonstrate that our locally-trained policy transfers zero-shot to the live Google Search API. ProMMSearchAgent achieves new SOTA performance, outperforming MMSearch-R1 by +5.1% on FVQA-test, +6.3% on InfoSeek, and +11.3% on MMSearch.

cs.CV↗

DR-MMSearchAgent: Deepening Reasoning in Multimodal Search Agents

Agentic multimodal models have garnered significant attention for their ability to leverage external tools to tackle complex tasks. However, it is observed that such agents often meet premature interaction collapse, caused by two primary reasons: 1) the terminal reward often appending on the last token prevents the advantage from distinguishing trajectories with exploratory behavior; 2) excessively redundant context hinders the agent from absorbing useful feedback. To address these issues, we propose the Deepening Reasoning MMSearchAgent, the framework leverages the structural proximity to derive advantage signals from the whole rollout trajectories in an entire batch, such that trajectories of different lengths are further encouraged to be generated, even when containing the same correct answer. Additionally, differentiated gaussian rewards are employed to dynamically calibrate interaction tolerance, thereby ensuring information reliability and reduce redundancy. To support multi-turn interaction training, we have constructed a multi-step deep-reasoning dataset including 3602 high-quality QA pair with at least 3 reasonning steps. Extensive experiments demonstrate that our method achieves state-of-the-art performance, outperforming the MMSearch-R1 by 8.4$\%$ on FVQA-test.

cs.CV↗

Dynamic Spatial-temporal Hypergraph Convolutional Network for Skeleton-based Action Recognition

Skeleton-based action recognition relies on the extraction of spatial-temporal topological information. Hypergraphs can establish prior unnatural dependencies for the skeleton. However, the existing methods only focus on the construction of spatial topology and ignore the time-point dependence. This paper proposes a dynamic spatial-temporal hypergraph convolutional network (DST-HCN) to capture spatial-temporal information for skeleton-based action recognition. DST-HCN introduces a time-point hypergraph (TPH) to learn relationships at time points. With multiple spatial static hypergraphs and dynamic TPH, our network can learn more complete spatial-temporal features. In addition, we use the high-order information fusion module (HIF) to fuse spatial-temporal information synchronously. Extensive experiments on NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets show that our model achieves state-of-the-art, especially compared with hypergraph methods.

cs.CV↗

Skeleton-based Action Recognition via Temporal-Channel Aggregation

Skeleton-based action recognition methods are limited by the semantic extraction of spatio-temporal skeletal maps. However, current methods have difficulty in effectively combining features from both temporal and spatial graph dimensions and tend to be thick on one side and thin on the other. In this paper, we propose a Temporal-Channel Aggregation Graph Convolutional Networks (TCA-GCN) to learn spatial and temporal topologies dynamically and efficiently aggregate topological features in different temporal and channel dimensions for skeleton-based action recognition. We use the Temporal Aggregation module to learn temporal dimensional features and the Channel Aggregation module to efficiently combine spatial dynamic channel-wise topological features with temporal dynamic topological features. In addition, we extract multi-scale skeletal features on temporal modeling and fuse them with an attention mechanism. Extensive experiments show that our model results outperform state-of-the-art methods on the NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets.

cs.CV↗