Wenwei Zhang

Research Scientist@Shanghai AI Laboratory·中国·OpenReview

研究方向

Large Language Model · Scalable Oversight/Self-improvement of LLMs · Language Agents · Self/Semi-supervised learning · 2D/3D Object Detection · Segmentation · and Tracking · Representation Learning · CNNs · Neural Architecture Search · AutoML

Wenwei Zhang

Achieving Olympia-Level Geometry Large Language Model Agent via Complexity Boosting Reinforcement Learning

RIG: Synergizing Reasoning and Imagination in End-to-End Generalist Policy

MM-HELIX: Boosting Multimodal Long-Chain Reflective Reasoning with Holistic Platform and Adaptive Hybrid Policy Optimization

LATTE: Learner-Adaptive Teacher-Forced Reflection for Advancing Deep Search

OPV: Outcome-based Process Verifier for Efficient Long Chain-of-Thought Verification

LONG-HORIZON REASONING AGENT FOR OLYMPIAD- LEVEL MATHEMATICAL PROBLEM SOLVING

The Imitation Game: Turing Machine Imitator is Length Generalizable Reasoner

UnifiedVerifier: Unifying Paradigms in Automated LLM Evaluation

Pre-Trained Policy Discriminators are General Reward Models

Rethinking Verification for LLM Code Generation: From Generation to Testing

Semi-off-Policy Reinforcement Learning for Vision-Language Slow-Thinking Reasoning

Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning

Mask-DPO: Generalizable Fine-grained Factuality Alignment of LLMs

Training Language Models to Critique with Multi-Agent Feedback

MindSearch: Mimicking Human Minds Elicits Deep AI Searcher