CST-RL: Contrastive Spatio-Temporal Representations for Reinforcement Learning.
Chi-Kai HoChung-Ta KingPublished in: IEEE Access (2023)
Keyphrases
- reinforcement learning
- spatio temporal
- reinforcement learning algorithms
- model free
- function approximation
- state space
- spatial and temporal
- optimal policy
- policy search methods
- learning algorithm
- markov decision processes
- rl algorithms
- temporal difference
- continuous state
- multi agent
- function approximators
- learning process
- transfer learning
- machine learning
- image sequences
- space time
- reinforcement learning methods
- control problems
- moving objects
- action selection
- continuous state and action spaces
- partially observable domains
- human actions
- learning classifier systems
- spatio temporal databases
- temporal difference learning
- learning agents
- learning capabilities
- partially observable
- control policy
- approximate dynamic programming
- actor critic
- policy search
- action space
- policy iteration