On Neural Consolidation for Transfer in Reinforcement Learning.
Valentin GuilletDennis WilsonCarlos Aguilar MelchorEmmanuel RachelsonPublished in: SSCI (2022)
Keyphrases
- reinforcement learning
- fitted q iteration
- transfer learning
- network architecture
- neural network
- function approximation
- state space
- model free
- reinforcement learning algorithms
- previously learned
- multi agent
- temporal difference
- knowledge transfer
- optimal policy
- sensory inputs
- markov decision processes
- learning process
- cross domain
- action selection
- multi agent reinforcement learning
- nonlinear predictive control
- neural computation
- artificial neural
- hebbian learning
- computational neuroscience
- stochastic approximation
- real robot
- neural model
- data center
- dynamical systems
- mobile robot
- machine learning