PID-Inspired Inductive Biases for Deep Reinforcement Learning in Partially Observable Control Tasks.
Ian CharJeff SchneiderPublished in: NeurIPS (2023)
Keyphrases
- partially observable
- reinforcement learning
- control system
- state space
- control method
- partial observability
- markov decision processes
- partially observable environments
- fuzzy pid control
- decision problems
- dynamical systems
- partially observable domains
- hidden state
- transfer learning
- infinite horizon
- markov decision problems
- reward function
- machine learning
- optimal control
- action selection
- belief state
- temperature control
- belief space
- partial observations
- function approximation
- reinforcement learning algorithms
- control strategy
- orders of magnitude
- initially unknown
- action models
- optimal policy
- multi agent
- partially observable markov decision processes
- control algorithm
- policy iteration
- evolutionary algorithm
- particle filter
- learning algorithm
- dynamic systems