A PAC Learning Algorithm for LTL and Omega-Regular Objectives in MDPs.
Mateo PerezFabio SomenziAshutosh TrivediPublished in: AAAI (2024)
Keyphrases
- concept classes
- sample complexity
- learning algorithm
- vc dimension
- markov decision processes
- reinforcement learning
- learning problems
- theoretical analysis
- upper bound
- linear threshold
- generalization error
- supervised learning
- active learning
- training examples
- pac model
- state space
- temporal logic
- model checking
- linear temporal logic
- learning models
- machine learning algorithms
- markov decision process
- optimal policy
- machine learning
- bounded model checking
- factored mdps
- training data
- linear time temporal logic
- finite horizon
- reinforcement learning algorithms
- dynamic programming
- policy iteration
- partially observable
- infinite horizon
- linear programming
- search space
- learning tasks
- heuristic search
- real time dynamic programming