Provably Correct Training of Neural Network Controllers Using Reachability Analysis.
Xiaowu SunYasser ShoukryPublished in: CoRR (2021)
Keyphrases
- provably correct
- reachability analysis
- neural network
- feed forward neural networks
- training algorithm
- training process
- markov decision processes
- model checking
- feedforward neural networks
- back propagation
- timed automata
- backpropagation algorithm
- formal methods
- artificial neural networks
- situation calculus
- multi layer perceptron
- training set
- recurrent neural networks
- neural network model
- hidden layer
- reinforcement learning
- state space
- control system
- machine learning
- neural network structure
- real time