Intrinsic Rewards from Self-Organizing Feature Maps for Exploration in Reinforcement Learning.
Marius LindegaardHjalmar Jacob VinjeOdin Aleksander SeverinsenPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- exploration strategy
- exploration exploitation
- action selection
- model based reinforcement learning
- markov decision processes
- active exploration
- reinforcement learning algorithms
- state space
- function approximation
- model free
- reward function
- machine learning
- temporal difference
- exploration exploitation tradeoff
- reward shaping
- autonomous learning
- learning problems
- optimal policy
- multi agent
- learning algorithm
- robotic control
- transfer learning
- control policy
- reinforcement learning methods
- hidden state
- neural network
- partially observable
- function approximators
- stochastic approximation
- supervised learning
- expected reward
- balancing exploration and exploitation