About me

Hi👋! I am a third-year PhD student in Computer and Information Science at the University of Pennsylvania, advised by Prof. Dan Roth. I received my Master degree in Economics and Computer Science from Duke univeristy advised by Prof. Sam Wiseman. I graduated from Renmin Univeristy of China (RUC), majored in Mathematics and Applied Mathematics and minor in Computer Science. I worked with Prof. Jing Zhang and Prof. Xin Zhao at RUC.

My research interests lie in Natural Language Processing and Machine Learning. I’m particularly interested in:

• Reliable LLM reasoning, especially for complex intents, including complex reasoning, decision-making, planning and interactions. I am excited about how symbolic and probabilistic reasoning tools (via solvers, constraints, uncertainty and causality) can verify and guide an LLM’s reasoning by enabling richer reward signals, more efficient synthetic supervision, better structure learning for abstraction, and stronger and more stable inference.
• LLM agents and agentic training, especially how models learn to plan, use external tools, and develop consistent world models. I am interested in how to design targeted environments, structured priors, and feedback loops that help models become more capable and grounded agents, and ultimately self-evolve through continuous interaction and reflection.

📑 Selected Research Projects

BIRD: A Trustworthy Bayesian Inference Framework for Large Language Models
Yu Feng, Ben Zhou, Weidong Lin, Dan Roth
ICLR 2025 (Oral)

VeriCoT: Neuro-symbolic Chain-of-Thought Validation via Logical Consistency Checks
Yu Feng, Nathaniel Weir, Kaj Bostrom, Sam Bayless, Darion Cassel, Sapana Chaudhary, Benjamin Kiesl-Reiter, Huzefa Rangwala
Preprint 2025

Rethinking LLM Uncertainty: A Multi-Agent Approach to Estimating Black-Box Model Uncertainty
Yu Feng, Phu Mon Htut, Zheng Qi, Wei Xiao, Manuel Mager, Nikolaos Pappas, Kishaloy Halder, Yang Li, Yassine Benajiba, Dan Roth
EMNLP 2025 Findings

Boss LLM: Adaptation via No-Regret Learning
Yu Feng∗, Avishree Khare∗, Nghia Nguyen∗, & Sikata Sengupta∗
ICLR 2025 Workshop on Scaling Self-Improving Foundation Models

BLINK: Multimodal Large Language Models Can See but Not Perceive
Xingyu Fu, Yushi Hu, Bangzheng Li, Yu Feng, Haoyu Wang, Xudong Lin, Dan Roth, Noah A. Smith, Wei-Chiu Ma, Ranjay Krishna
ECCV 2024