Scholar
Ming Yin
Google Scholar ID: OLo70zgAAAAJ
PhD Student at Duke University
Citations & Impact
Citations
87
H-index
4
i10-index
3
Publications
8
Co-authors
1
list available
Publications
Resume
Academic Achievements
- Published multiple papers including 'LiveMCP-101: Stress Testing and Diagnosing MCP-enabled Agents on Challenging Queries' (preprint), 'Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems' (ICML 2025, Spotlight), 'Poisoning Federated Recommender Systems with Fake Users' (WWW 2024), 'LoBAM: LoRA-Based Backdoor Attack on Model Merging' (ICLR 2025 Workshop), 'Keyframe-oriented Vision Token Pruning: Enhancing Efficiency of Large Vision Language Models on Long-Form Video Processing' (ICCV 2025), 'Robust Federated Learning Mitigates Client-side Training Data Distribution Inference Attacks' (WWW 2024), 'Complex Logical Instruction Generation' (preprint), and 'MMTok: Multimodal Coverage Maximization for Efficient Inference of VLMs' (preprint).
Research Experience
- Served as a GenAI Research Intern at Zoom (May 2025 - Aug 2025); worked as a Teaching Assistant for ECE 661 Computer Engineering Machine Learning and Deep Neural Nets at Duke University (Fall 2025). Also provided reviewer service for EMNLP 2025, ICLR 2025, ICCV 2025, NeurIPS 2025, and TDSC.
Education
- Received a Bachelor's degree from the School of the Gifted Young, USTC in 2024 at the age of 20; currently a second-year PhD student at Duke University, supervised by Prof. Yiran Chen.
Background
- Currently interested in LLM Agents, LLM Reasoning, and Trustworthy AI.
