The configurable tree graph (CT-graph): measurable problems in partially observable and distal reward environments for lifelong reinforcement learning.
Andrea SoltoggioEseoghene Ben-IwhiwhuChristos PeridisPawel LadoszJeffery DickPraveen K. PillySoheil KolouriPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- partially observable
- partial observability
- decision problems
- partially observable environments
- function approximation
- spanning tree
- state space
- random walk
- markov decision processes
- partially observable markov decision processes
- dynamical systems
- reward function
- reinforcement learning algorithms
- transfer learning
- machine learning
- dynamic programming