Qi Lu

Skip the expensive sensors: a low-cost drone and some clever video processing can reliably detect structural damage by tracking subtle vibration changes.

Reinforcement learning can teach robot swarms to bid more effectively in task allocation auctions, outperforming hand-crafted heuristics without sacrificing decentralized control.

Robot swarms can achieve significantly faster and more complete resource collection by using a simple visitation grid to guide exploration, even with limited onboard resources.

Vision-language models falter at the fine-grained temporal recognition crucial for surgical video understanding, while SurgRec excels.