Generalizing to generalize: Humans flexibly switch between compositional and conjunctive structures during reinforcement learning.
Nicholas T. FranklinMichael J. FrankPublished in: PLoS Comput. Biol. (2020)
Keyphrases
- reinforcement learning
- high speed
- reinforcement learning algorithms
- robotic control
- multi agent
- function approximation
- database
- real world
- temporal difference
- model free
- human behavior
- computer vision
- learning algorithm
- transfer learning
- markov decision processes
- information systems
- state space
- dynamic programming
- artificial intelligence
- robot control
- body movements
- multi agent reinforcement learning
- bayesian networks