Towards a Unified Benchmark for Reinforcement Learning in Sparse Reward Environments.
Yongxin KangEnmin ZhaoYifan ZangKai LiJunliang XingPublished in: ICONIP (4) (2022)
Keyphrases
- reinforcement learning
- state space
- function approximation
- multi agent environments
- real world
- temporal difference
- eligibility traces
- reinforcement learning algorithms
- reward function
- machine learning
- high dimensional
- learning agent
- sparse data
- model free
- dynamic environments
- unified model
- partially observable environments
- optimal policy
- optimal control
- sparse representation
- markov decision processes
- multi agent
- learning process
- total reward
- robotic control
- sparse coding
- compressive sensing
- markov decision process
- supervised learning
- policy gradient
- multi agent reinforcement learning
- inverse reinforcement learning
- neural network
- reward shaping