Open-Ended Reinforcement Learning with Neural Reward Functions.
Robert MeierAsier MujikaPublished in: NeurIPS (2022)
Keyphrases
- open ended
- reward function
- reinforcement learning
- reinforcement learning algorithms
- fitted q iteration
- markov decision processes
- policy search
- state space
- optimal policy
- markov decision process
- partially observable
- learning outcomes
- inverse reinforcement learning
- function approximation
- transition model
- multiple agents
- model free
- state action
- state variables
- multi agent
- learning agent
- inquiry learning
- transition probabilities
- generative model
- learning algorithm
- learning process
- initially unknown
- affect detection
- machine learning
- markov models
- action space
- average reward
- dynamic programming
- prior knowledge
- e learning