Reinforcement Learning with Sparse Rewards using Guidance from Offline Demonstration.
Desik RengarajanGargi VaidyaAkshay SarveshDileep M. KalathilSrinivas ShakkottaiPublished in: ICLR (2022)
Keyphrases
- reinforcement learning
- state space
- function approximation
- markov decision processes
- machine learning
- reinforcement learning algorithms
- model free
- sparse data
- dynamic programming
- reward function
- real time
- high dimensional
- optimal policy
- sparse representation
- multi agent
- learning algorithm
- supervised learning
- action selection
- control policy
- function approximators
- learning problems
- reward shaping
- neural network
- compressed sensing
- partially observable
- learning process
- hidden state
- multi agent reinforcement learning