METRA: Scalable Unsupervised RL with Metric-Aware Abstraction.
Seohong ParkOleh RybkinSergey LevinePublished in: ICLR (2024)
Keyphrases
- completely unsupervised
- reinforcement learning
- high level abstractions
- unsupervised learning
- metric space
- distance measure
- decision theoretic planning
- web scale
- semi supervised
- distance function
- lightweight
- machine learning
- highly scalable
- data sets
- image quality
- supervised learning
- multi agent
- transfer learning
- markov decision processes
- state space
- function approximation
- similarity metric
- learning process
- pairwise
- unsupervised manner
- learning algorithm
- objective function