Recombinant building: the ability to generate and recombine navigation structures is difficult to acquire through just reinforcement learning.
Ganesh ShindeHarshit AgrawalSanjay ChandrasekharanPublished in: CogSci (2019)
Keyphrases
- reinforcement learning
- automatically generating
- learning algorithm
- optimal policy
- neural network
- web navigation
- function approximation
- markov decision processes
- real time
- supervised learning
- dynamic programming
- search space
- information systems
- model free
- machine learning
- transition model
- policy search
- underwater vehicles
- robotic control