Zhou Yu

About

Zhou Yu is an Associate Professor at Columbia University's Computer Science Department, co-directing DAPlab. Her research focuses on designing algorithms for real-time intelligent interactive systems that coordinate with user actions beyond spoken language, including non-verbal behaviors. She optimizes human-machine communication through multimodal sensing, speech and NLP, and HCI. She is also a co-founder of Arklex.ai and has received numerous awards including Forbes 30 under 30 in Science (2018).

Education

• Ph.D. Language Technology Institute, School of Computer Science, Carnegie Mellon University, 2017
• B.S. Computer Science, Zhejiang University, 2011
• B.A. Linguistics, English Language Major, Zhejiang University, 2011

Selected publications

• Mitigating Tail Risks in Language Model Output Distributions
• LiGen: Active Lipid Generation via a Molecular Language Model
• Budget-Aware Anytime Reasoning with LLM-Synthesized Preference Data
• FormGym: Doing Paperwork with Agents
• SAGE : A Top-Down Bottom-Up Knowledge-Grounded User Simulator for Multi-turn Agent Evaluation
• Dyna-Mind: Learning to Simulate from Experience for Better AI Agents
• Question-Answering in Asynchronous Learning Environments, Bringing Pedagogy into Focus : Evaluating Virtual Teaching Assistants
• AutoSpec: An Agentic Framework for Automatically Drafting Patent Specification
• ConFit v2: Improving Resume-Job Matching using Hypothetical Resume Embedding and Runner-Up Hard-Negative Mining
• Learning and Evaluating Factual Clarification Question Generation Without Examples