Using Natural Language for Reward Shaping in Reinforcement Learning.
Prasoon GoyalScott NiekumRaymond J. MooneyPublished in: CoRR (2019)
Keyphrases
- reward shaping
- reinforcement learning
- natural language
- reinforcement learning algorithms
- complex domains
- machine learning
- state space
- knowledge representation
- function approximation
- markov decision problems
- learning algorithm
- partially observable
- markov decision processes
- model free
- reward function
- bayesian networks
- dialogue system
- learning process
- function approximators
- multi agent
- monte carlo
- sufficient conditions
- temporal difference
- supervised learning