Scholar

Changwang Zhang (张长旺)

Google Scholar ID: 4j8dHOkAAAAJ

CCF Theoretical Computer Science Technical Committee

AI AgentLarge Reasoning ModelMultimodal Large Language ModelInformation Retrieval

Citations & Impact

All-time

Citations

796

H-index

i10-index

Publications

Co-authors

list available

Contact

Publications

20 items

Browse publications on Google Scholar (top-right) ↗

Resume (English only)

Academic Achievements

Proposed OThink-MR1, a multimodal reasoning model and framework that significantly enhances generalization and reasoning in multimodal tasks
Developed GRPO-D, a dynamic reinforcement learning algorithm achieving over 61.63% average relative performance improvement over supervised fine-tuning (SFT) with strong cross-task generalization
Introduced EulerFormer (SIGIR’24), which enhances Transformer expressiveness and robustness via complex attention networks and adaptive rotational position encoding
Presented PoseCrafter (ECCV’24), a novel method for personalized video generation with precise pose control, outperforming baselines across 8 standard metrics
Research featured in The Guardian and Daily Mail
Recipient of Tencent Gold Award for Excellence in R&D, Operation Excellence Award, Open Source Collaboration Award, and Micro Innovation Award

Co-authors

10 total