Balancing Multiple Sources of Reward in Reinforcement Learning.
Christian R. SheltonPublished in: NIPS (2000)
Keyphrases
- multiple sources
- reinforcement learning
- function approximation
- reinforcement learning algorithms
- multi source
- data sources
- optimal policy
- reward function
- state space
- average reward
- model free
- markov decision processes
- machine learning
- total reward
- eligibility traces
- heterogeneous sources
- temporal difference
- multi agent
- multiple information sources
- reward shaping
- multiple data sources
- data from multiple sources
- transfer learning
- dynamic programming
- learning algorithm
- policy gradient
- partially observable environments
- user interests
- data sets
- policy evaluation
- website
- partially observable
- social media
- action selection
- information extraction