Jun Yan

Static environments can be transformed on-the-fly to better suit agent learning, resulting in up to a 9.0-point performance boost with fewer execution steps.

Fine-grained metrics reveal that robots can recover from failures more effectively than previously thought, reshaping our understanding of their capabilities.

High binary recognition performance in AMP models fails to predict real-world assay outcomes, revealing critical gaps in current evaluation methods.

Forget hand-annotated data: Magnet distills multi-turn tool-use skills into LLMs by automatically generating training trajectories that outperform even Gemini 1.5 Pro.