Professor, University of Oxford
2 papers at NeurIPS 2025
We propose a principled taxonomy, evaluation procedure, and unified algorithm space for offline RL.
Training reinforcement learning agents from a single language instruction using vision-language models.
