About Me
I am a third-year Ph.D. student in Computer Science at Arizona State University, advised by Prof. Hua Wei. Before ASU, I received my M.Sc. in Computer Science from the University of British Columbia under the guidance of Prof. Yong Gao, and my B.Eng. in Computer Science from Beijing Jiaotong University.
My research aims to make large language models and LLM agents reliable enough for the real world:
- Uncertainty quantification in LLMs: estimating how confident a model should be in its outputs and reasoning steps, and using these estimates to improve reasoning performance and reliability.
- Learning and adaptation in LLM agents: building agents that keep learning from interaction, acquire and refine reusable skills, and coordinate well in multi-agent systems.
I am available for research internships in Winter 2027 and Summer 2027. Feel free to reach out!
News
-
A paper: “AndroidReality: How Far Are Mobile Agents from the Real World?” is accepted to NeurIPS’26 Datasets & Benchmarks Track!
-
Two papers: “Diverse Reasoning Paths Matter: Symbolic Equivalence Augmentation for Enhancing RL Exploration” and “Detect Before You Disambiguate: Perturbation-Based Input Ambiguity Detection in LLMs” are accepted to AACL’26 Main!
-
Two papers are accepted to EMNLP’26!
-
A paper: “PlantShade: Predicting Plant Shadows for Lighting-Aware Robotic Agricultural Operation” is accepted to IROS’26! Grateful for the IEEE RAS Travel Support.
-
Started my Research Scientist Internship at TikTok.
-
Two papers are accepted to ICML’26, including “Diagnosing Multi-step Reasoning Failures in Black-box LLMs via Stepwise Confidence Attribution”.
-
A paper: “The Sim-to-Real Gap of Foundation Model Agents: A Unified MDP Perspective” is accepted to KDD’26 Blue Sky Track!
-
A paper is accepted to ACL’26 Findings!
-
Two papers are accepted to PAKDD’26 (Oral) and EACL’26 Findings!
-
Our tutorial: “Uncertainty Quantification and Confidence Calibration in LLMs” was presented at ICDM’25 in Washington, DC, US.
-
Our tutorial: “Uncertainty Quantification and Confidence Calibration in LLMs” was presented at KDD’25 in Toronto, Canada.
-
A paper is accepted to COLM’25!
-
Our paper received the Best Artifact Award at ICCPS’25! 🏆
-
Two papers: “Uncertainty Quantification and Confidence Calibration in Large Language Models: A Survey” and “Is Your Explanation Reliable: Confidence-Aware Explanation on Graph Neural Networks” are accepted to KDD’25!
Selected Publications
View all →
AACL 2026
Diverse Reasoning Paths Matter: Symbolic Equivalence Augmentation for Enhancing RL Exploration
