Landmark based guidance for reinforcement learning agents under partial observability.
Alper DemirErkin ÇildenFaruk PolatPublished in: Int. J. Mach. Learn. Cybern. (2023)
Keyphrases
- partial observability
- reinforcement learning agents
- reinforcement learning
- dynamic environments
- partially observable
- belief space
- state abstraction
- partially observable markov decision processes
- function approximation
- markov decision process
- state space
- planning under partial observability
- learning algorithm
- multi agent environments
- belief state
- transfer learning
- reinforcement learning algorithms
- multi agent
- model free
- learning agent
- optimal policy
- planning problems
- markov decision processes
- initial state
- supervised learning
- path planning
- dynamical systems
- search algorithm