Multi-task Actor-Critic with Knowledge Transfer via a Shared Critic.
Gengzhi ZhangLiang FengYaqing HouPublished in: ACML (2021)
Keyphrases
- actor critic
- multi task
- knowledge transfer
- transfer learning
- reinforcement learning
- learning tasks
- multi task learning
- function approximation
- temporal difference
- policy gradient
- approximate dynamic programming
- optimal control
- multitask learning
- reinforcement learning algorithms
- neuro fuzzy
- gradient method
- learning problems
- transfer knowledge
- state space
- policy iteration
- machine learning
- learning algorithm
- multiple tasks
- markov decision processes
- supervised learning
- model free
- knowledge sharing
- learning process
- optimal policy
- labeled data
- high dimensional
- data sets
- multi class
- text classification
- machine learning algorithms
- evaluation function
- decision trees
- feature selection
- learning experience
- neural network