MTCL learns multi-scale temporal correlations in videos via contrastive learning to produce more informative representations that improve sample efficiency and performance in downstream RL tasks.
Improving transformer world models for data-efficient rl
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.LG 2verdicts
UNVERDICTED 2representative citing papers
Simulus integrates flexible tokenization, intrinsic motivation, prioritized world model replay, and regression-as-classification to achieve state-of-the-art sample efficiency for planning-free world model agents on visual Atari 100K, DMC Proprioception 500K, and symbolic Craftax-1M benchmarks.
citing papers explorer
-
From Pixels to Temporal Correlations: Learning Informative Representations for Reinforcement Learning Pre-training
MTCL learns multi-scale temporal correlations in videos via contrastive learning to produce more informative representations that improve sample efficiency and performance in downstream RL tasks.
-
Simulus: Combining Improvements in Sample-Efficient World Model Agents
Simulus integrates flexible tokenization, intrinsic motivation, prioritized world model replay, and regression-as-classification to achieve state-of-the-art sample efficiency for planning-free world model agents on visual Atari 100K, DMC Proprioception 500K, and symbolic Craftax-1M benchmarks.