Generalizing Across Multi-Objective Reward Functions in Deep Reinforcement Learning.
Eli FriedmanFred FontainePublished in: CoRR (2018)
Keyphrases
- reward function
- multi objective
- reinforcement learning
- reinforcement learning algorithms
- state space
- markov decision processes
- evolutionary algorithm
- multi objective optimization
- policy search
- partially observable
- inverse reinforcement learning
- optimal policy
- markov decision process
- genetic algorithm
- multiple agents
- function approximation
- objective function
- learning agent
- transition model
- machine learning
- dynamic programming
- initially unknown
- learning algorithm
- state variables
- action selection
- multi agent
- function approximators
- model free
- temporal difference
- reward signal
- control policies
- average reward
- multi agent systems
- transition probabilities