Gradient-Descent for Randomized Controllers Under Partial Observability.
Linus HeckJip SpelSebastian JungesJoshua MoermanJoost-Pieter KatoenPublished in: VMCAI (2022)
Keyphrases
- partial observability
- reinforcement learning
- planning problems
- partially observable
- belief state
- state space
- belief space
- objective function
- control system
- planning under partial observability
- learning rules
- partially observable markov decision processes
- learning algorithm
- planning domains
- markov decision process
- control strategy
- domain specific