PID-Inspired Inductive Biases for Deep Reinforcement Learning in Partially Observable Control Tasks.
Ian CharJeff G. SchneiderPublished in: CoRR (2023)
Keyphrases
- partially observable
- reinforcement learning
- control system
- control method
- partial observability
- decision problems
- markov decision processes
- state space
- partially observable environments
- dynamical systems
- optimal control
- fuzzy pid control
- partially observable domains
- infinite horizon
- transfer learning
- temperature control
- reward function
- hidden state
- markov decision problems
- belief space
- pid controller
- machine learning
- belief state
- function approximation
- reinforcement learning algorithms
- action selection
- optimal policy
- control algorithm
- action models
- control strategy
- partial observations
- planning domains
- partially observable markov decision process
- dynamic programming
- fully observable
- autonomous robots
- domain specific
- learning algorithm