1. One paper put on arXiv on Sep 6, 2025: From Image Generation to Infrastructure Design: a Multi-agent Pipeline for Street Design Generation.
2. One paper put on arXiv on Jun 23, 2025: CaughtCheating: Is Your MLLM a Good Cheating Detective? Exploring the Boundary of Visual Perception and Reasoning.
3. Two papers accepted by ACL 2025 on May 16, 2025: From Perceptions to Decisions: Wildfire Evacuation Decision Prediction with Behavioral Theory-informed LLMs; Mosaic-IT: Free Compositional Data Augmentation Improves Instruction Tuning.
4. Successfully passed the qualifying examination and advanced to Ph.D. candidacy at Stony Brook University on Jan 24, 2025.
5. One paper accepted by NAACL 2025 on Jan 22, 2025: RuleR: Improving LLM Controllability by Rule-based Data Recycling.
6. One paper accepted by International Journal of Disaster Risk Reduction on Jul 14, 2024: Near-real-time earthquake-induced fatality estimation using crowdsourced data and large-language models.
7. One paper accepted by International Journal of Disaster Risk Reduction on Mar 5, 2024: Scalable and rapid building damage detection after hurricane Ian using causal Bayesian networks and InSAR imagery.
Research Experience
Graduate Visiting Scholar at Johns Hopkins University from 2024 to 2025.
Education
Bachelor of Science in Computer Science from Xi’an Jiaotong University in 2020; Master of Science in Computer Science from Stevens Institute of Technology in 2023, advised by Prof. Xiaojiang Du; Currently a Ph.D. candidate at Stony Brook University, advised by Prof. Susu Xu.
Background
Research Interests: Machine Learning (ML), Natural Language Processing (NLP), and Large Language Models (LLM). Specific Focus: LLM applications in Disaster Management, Post-training for LLMs, Multi-Modal LLMs (MLLMs).
Miscellany
Personal Interests: Interested in RAG, Agent, and RL.