Latent Plans for Task-Agnostic Offline Reinforcement Learning.
Erick Rosete-BeasOier MeesGabriel KalweitJoschka BoedeckerWolfram BurgardPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- deterministic domains
- plan recognition
- function approximation
- real time
- machine learning
- model free
- latent variables
- optimal policy
- state space
- planning problems
- plan generation
- learning algorithm
- macro actions
- robotic control
- function approximators
- markov decision processes
- multi agent
- partially ordered
- learning process
- dynamic programming
- temporal difference learning
- reinforcement learning methods
- stochastic approximation
- multi agent reinforcement learning
- transition model
- domain independent
- collaborative filtering