Transfer Learning Between RTS Combat Scenarios Using Component-Action Deep Reinforcement Learning.
Richard KellyDavid ChurchillPublished in: AIIDE Workshops (2020)
Keyphrases
- transfer learning
- reinforcement learning
- learning tasks
- action selection
- action models
- knowledge transfer
- cross domain
- machine learning
- action space
- active learning
- target domain
- function approximation
- transfer knowledge
- manifold alignment
- labeled data
- multi task learning
- learning algorithm
- multi task
- collaborative filtering
- reinforcement learning algorithms
- state space
- domain adaptation
- temporal difference
- learning problems
- model free
- text classification
- text mining
- transferring knowledge
- solve complex tasks
- markov decision processes
- optimal policy
- domain specific
- reinforcement learning agents