News

Our paper Cognitive Episodes in LLM Reasoning Traces Enable Interpretable Human Item Difficulty Prediction has been accepted by Transactions on Machine Learning Research!

Our paper A Missing Piece for Trustworthy AI Reviewers: From Benchmarking Rhetorical Robustness to SciCore Review has been selected for a Best Paper Honorable Mention at the NeurIPS 2026 Workshop on AI-Native Academia!

A high school research project I supervised, “Learning When to Revise: Conflict Memory for Medical Multi-Agent Reasoning,” has been selected for a Best Paper Finalist Award at the IEEE BigData 2026 High School Symposium!

Three new papers are now available!
1 A Missing Piece for Trustworthy AI Reviewers: From Benchmarking Rhetorical Robustness to SciCore Review;
2 The Emerging AI Paper-Review Arms Race: Adversarial Co-Evolution in Scholarly Publishing;
3 How Can Rhetoric Reward-Hack AI Reviewers? Dissecting Rhetorical Sensitivity in AI-Based Peer Review.

Two new papers are now available!
1 Cognitive Episodes in LLM Reasoning Traces Enable Interpretable Human Item Difficulty Prediction;
2 LLMs Struggle to Measure What Distinguishes Students of Different Proficiency Levels: A Study of Item Discrimination in Reading Comprehension Assessment.

I received the Big Contribution grant from the Virginia Tech Student Engineers Council to support my research project.

I have joined Virginia Tech as a Virginia Tech Presidential Postdoctoral Fellow, working with Prof. Dawei Zhou. Looking forward to the next stage of my research journey.

🎓 I received my Ph.D. from Stony Brook University.

One paper was selected as a LAW Workshop @ NeurIPS 2025 Spotlight paper!
CaughtCheating: Is Your MLLM a Good Cheating Detective? Exploring the Boundary of Visual Perception and Reasoning.

One paper was accepted by UrbanAI Workshop @ NeurIPS 2025!
From Image Generation to Infrastructure Design: a Multi-agent Pipeline for Street Design Generation.

Our hurricane damage detection research was featured by NOAA NCCOS: New Study Advances Detection of Building Damage After Hurricanes.

One paper was put on the arXiv: From Image Generation to Infrastructure Design: a Multi-agent Pipeline for Street Design Generation.

One paper was put on the arXiv: CaughtCheating: Is Your MLLM a Good Cheating Detective? Exploring the Boundary of Visual Perception and Reasoning.

Two papers were accepted by ACL 2025!
1 From Perceptions to Decisions: Wildfire Evacuation Decision Prediction with Behavioral Theory-informed LLMs;
2 Mosaic-IT: Free Compositional Data Augmentation Improves Instruction Tuning;

I successfully passed the qualifying examination and advanced to Ph.D. candidacy at Stony Brook University.

One paper was accepted by NAACL 2025!
RuleR: Improving LLM Controllability by Rule-based Data Recycling

I am excited to join Johns Hopkins University as a Visiting Scholar for a one-year fixed term.

One paper was accepted by International Journal of Disaster Risk Reduction!
Near-real-time earthquake-induced fatality estimation using crowdsourced data and large-language models

One paper was accepted by International Journal of Disaster Risk Reduction!
“Scalable and rapid building damage detection after hurricane Ian using causal Bayesian networks and InSAR imagery”

I arrived at Stony Brook University, officially beginning my journey as a Ph.D. student.

I completed my Master’s in Computer Science at Stevens Institute of Technology.

I received the Provost Doctoral Fellowship at Stevens Institute of Technology in recognition of my academic excellence.

My research project advised by Prof. Xiaojiang Du has won the ECE Research Scholarship Award.

I started my Master’s at Stevens Institute of Technology.