Dueling RL: Reinforcement Learning with Trajectory Preferences.
Aadirupa SahaAldo PacchianoJonathan LeePublished in: AISTATS (2023)
Keyphrases
- reinforcement learning
- function approximation
- state space
- reinforcement learning algorithms
- control problems
- model free
- markov decision processes
- user preferences
- learning algorithm
- temporal difference learning
- temporal difference
- direct policy search
- machine learning
- optimal policy
- partially observable domains
- multi agent
- supervised learning
- transfer learning
- autonomous learning
- partially observable
- trajectory data
- dynamic programming
- markov decision problems
- rl algorithms
- reinforcement learning methods
- multi attribute
- learning agents
- function approximators
- policy iteration
- markov decision process
- learning problems
- decision making
- continuous state
- learning process
- approximate dynamic programming
- policy search
- reward shaping
- action selection
- continuous state and action spaces
- optimal control