Mutual Information Based Knowledge Transfer Under State-Action Dimension Mismatch.
Michael WanTanmay GangwaniJian PengPublished in: CoRR (2020)
Keyphrases
- knowledge transfer
- state action
- reinforcement learning
- evaluation function
- transfer learning
- knowledge sharing
- stochastic games
- markov decision process
- action space
- average reward
- state transitions
- belief state
- neural network
- model free
- function approximators
- stochastic processes
- function approximation
- markov decision processes
- information sharing
- learning algorithm