Learning the Arrow of Time for Problems in Reinforcement Learning.
Nasim RahamanSteffen WolfAnirudh GoyalRoman RemmeYoshua BengioPublished in: ICLR (2020)
Keyphrases
- reinforcement learning
- learning process
- learning algorithm
- control problems
- solving problems
- online learning
- learning classifier systems
- function approximators
- optimization problems
- knowledge acquisition
- learning systems
- mobile learning
- supervised learning
- state space
- model free
- learning capabilities
- complex domains
- active learning
- robot control
- prior knowledge
- learning agents
- rl algorithms
- multi agent reinforcement learning
- eligibility traces