Yi-Chen Li

REAR transforms how we achieve user preference alignment in LLMs, enabling scalable realignment without costly retraining.

SkeMex enables medical agents to evolve their reasoning capabilities by transforming raw experience into structured, reusable skills, outperforming traditional memory systems.

Instruction tuning may enhance LLMs' ability to follow commands, but it compromises their performance in critical code infilling tasks, revealing a hidden cost in AI coding assistants.

DART boosts dense retrieval reranking performance by adapting scoring functions on-the-fly, achieving significant gains without the need for extensive training data.

Unsupervised object detection can now achieve category awareness, bridging the gap with supervised methods without needing any labeled data.