Reinforcement Learning in the Era of LLMs: What is Essential? What is needed? An RL Perspective on RLHF, Prompting, and Beyond.
Hao SunPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- function approximation
- state space
- reinforcement learning algorithms
- model free
- rl algorithms
- optimal policy
- machine learning
- control problems
- markov decision processes
- transfer learning
- policy search
- optimal control
- action selection
- multi agent
- temporal difference learning
- learning algorithm
- supervised learning
- direct policy search
- state and action spaces
- autonomous learning
- neural network
- reinforcement learning methods
- partially observable markov decision processes
- partially observable
- dynamic programming
- real valued
- learning problems
- learning agents
- complex domains
- learning process
- continuous state
- learning classifier systems
- approximate dynamic programming
- multi agent reinforcement learning
- actor critic
- reinforcement learning agents
- continuous state and action spaces