PPL
Byeongchan Kim
Papers (4)
Latent Representation Alignment for Offline Goal-Conditioned Reinforcement LearningMulti-Step Likelihood-Ratio Correction for Reinforcement Learning with Verifiable RewardsPeng's Q($λ$) for Conservative Value Estimation in Offline Reinforcement LearningRelFlexformer: Efficient Attention 3D-Transformers for Integrable Relative Positional Encodings