Jin Wang

Achieve real-time, COLREGs-compliant autonomous marine navigation by formulating collision avoidance and grounding prevention as a single convex optimization problem.

Simulate-to-real gaps in maritime autonomy shrink with a new framework that realistically models how adverse weather and bathymetry degrade sensor performance.

Get better image captions without more data: reinforcement learning can train vision-language models to focus on image details by maximizing the similarity between images retrieved using the generated captions.

Forget rigid templates: RL-optimized verbalization of user logs boosts LLM-based recommendation accuracy by up to 93%.