Accelerating Exploration with Unlabeled Prior Data
Paper
• 2311.05067 • Published
• 2
RL
-
-
Efficient Online Reinforcement Learning with Offline Data
Paper • 2302.02948 • Published • 2
Learning to (Learn at Test Time): RNNs with Expressive Hidden States
Paper • 2407.04620 • Published • 34Note Subset of parameter learnable during inference with SSL target. Great idea.
Reinforcement Pre-Training
Paper • 2506.08007 • Published • 263Note Still post-training.
-
SkillBlender: Towards Versatile Humanoid Whole-Body Loco-Manipulation via Skill Blending
Paper • 2506.09366 • Published • 8
General agents need world models
Paper • 2506.01622 • Published • 2Note Basically wrong, Markov Decision Process requires decision making invariant to history, the consideration of temporally dependent goal that's not encoded in current state itself is falacy.