On Assessing The Safety of Reinforcement Learning algorithms Using Formal Methods.
Paulina Stevia Nouwou MindomAmin NikanjamFoutse KhomhJohn MullinsPublished in: CoRR (2021)
Keyphrases
- formal methods
- reinforcement learning algorithms
- safety critical
- reinforcement learning
- state space
- markov decision processes
- model free
- reinforcement learning problems
- formal specification
- eligibility traces
- software engineering
- reinforcement learning methods
- temporal difference
- knowledge based systems
- model checking
- learning algorithm
- reward function
- artificial intelligence
- partially observable environments
- function approximation
- dynamic environments
- relational databases
- development lifecycle
- neural network
- tabula rasa
- modeling language
- expert systems