Self-Generation of Reward by Sensor Input in Reinforcement Learning.
Kaoru NikaidoKentarou KurashigePublished in: RVSP (2013)
Keyphrases
- reinforcement learning
- function approximation
- eligibility traces
- sensor data
- learning algorithm
- markov decision processes
- temporal difference
- reinforcement learning algorithms
- generation process
- sensor networks
- state space
- optimal policy
- reward function
- internal state
- data acquisition
- multi agent
- multi sensor
- real time
- learning capabilities
- partially observable environments
- action selection
- model free
- optimal control
- transfer learning
- sufficient conditions
- input data
- machine learning