Computer Vision / Multimodal Learning
I am a first-year CS PhD student in the Multimodal Vision Research Laboratory at Washington University, advised by Dr. Nathan Jacobs. I work on computer vision and multimodal learning.
Before joining WashU, I obtained my Bachelor's and Master's degree from Wuhan University. I also spent a wonderful time at Microsoft Research Asia in 2024, working on MLLMs for scientific diagrams.
Currently, I am working on multimodal learning, spatial reasoning and world model-related problems. If you share similar interests, feel free to reach out for collaboration!
My past research focused on learning representations for few-shot scenarios and building MLLMs for geospatial tasks. I have also applied AI to domains including games, agriculture, geoscience, urban traffic, and environmental monitoring, where I believe domain-grounded problems can motivate, shape, and co-evolve with methodology.
We build SeqGeo-VL, a benchmark of ~39K paired ground observations, trajectories, route descriptions, and overhead satellite maps, and train TrajLoc, a VLM embedding model for cross-view matching with overhead maps. TrajLoc supports video, language, or hybrid queries with optional trajectory conditioning, and serves as an MLLM tool for outdoor spatial reasoning.
Developed with industry collaborators Microsoft and Wherobots. A scalable model recipe and data pipeline for planetary-scale field delineation.
Designed the first benchmark for evaluating multimodal large language models on geologic map interpretation. Proposed an agentic baseline achieving state-of-the-art performance.
