LTL and Beyond: Formal Languages for Reward Function Specification in Reinforcement Learning.
Alberto CamachoRodrigo Toro IcarteToryn Q. KlassenRichard Anthony ValenzanoSheila A. McIlraithPublished in: IJCAI (2019)
Keyphrases
- formal languages
- reward function
- reinforcement learning
- reinforcement learning algorithms
- markov decision processes
- state space
- natural language
- optimal policy
- formal language
- inverse reinforcement learning
- context free grammars
- partially observable
- markov decision process
- linear temporal logic
- function approximation
- transition probabilities
- pattern languages
- temporal logic
- multiple agents
- hierarchical reinforcement learning
- temporal difference
- model free
- transition model
- machine learning
- model checking
- dynamic programming
- multi agent
- markov models
- initially unknown
- learning algorithm
- state variables
- evaluation function
- learning agent
- artificial intelligence
- data mining
- reward signal