Yaoqing Yang

Yaoqing Yang is an Assistant Professor at the Department of Computer Science, Dartmouth College. He earned his PhD in Electrical and Computer Engineering (ECE) from Carnegie Mellon University (CMU) and completed postdoctoral research at UC Berkeley's RISE Lab. His work focuses on robustness in machine learning systems, spectral analysis of neural networks, and algorithm design for structured data like graphs and point clouds.

• PhD in ECE, Carnegie Mellon University
• Postdoc, RISE Lab, UC Berkeley
• BS in Electrical Engineering, Tsinghua University

Research interests include diagnosing and mitigating model failures through heavy-tailed spectral analysis, decision boundary studies, and loss landscape visualization. He develops methods such as AlphaPruning and SharpBalance to enhance large language models and ensemble learning.

Recent work spans 2025 publications on spectral evolution of neural networks, agentic AI for science, and LLM safety. Key collaborations include Michael W. Mahoney and other researchers.

• Burke Research Initiation Award, Dartmouth (2024)
• DOE grant for scientific foundation models (2024)
• DARPA grant for AI robustness (2024)

He serves as Area Chair at NeurIPS 2025 and ICLR 2026, and has presented at Google Research, Lawrence Berkeley National Laboratory, and leading universities worldwide.

His lab at Dartmouth engages in theoretical and applied research, with connections to UC Berkeley's RISE Lab and collaborations across institutions like CMU and Tsinghua University.