Yuejiang Liu

Continuous scoring from LLM-as-a-Verifier leads to state-of-the-art verification accuracy and improved sample efficiency in reinforcement learning tasks.

DREAM-Chunk transforms action chunking by leveraging latent world models to enhance robustness against stochastic dynamics without the need for policy retraining.

World models can now self-improve by identifying their own prediction errors, thanks to a clever decomposition of action-conditioned prediction into easier-to-verify components.

Turns out, the best memory design for robotic manipulation depends heavily on the task, with no single architecture dominating across the board.