Boosting Verification of Deep Reinforcement Learning via Piece-Wise Linear Decision Neural Networks.
Jiaxu TianDapeng ZhiSi LiuPeixin WangCheng ChenMin ZhangPublished in: NeurIPS (2023)
Keyphrases
- neural network
- reinforcement learning
- piece wise linear
- pattern recognition
- decision making
- learning algorithm
- function approximators
- artificial neural networks
- function approximation
- machine learning
- reinforcement learning algorithms
- decision problems
- model checking
- decision rules
- state space
- optimal policy
- back propagation
- decision makers
- fuzzy logic
- feature selection
- unsupervised feature learning
- temporal difference
- model free
- feed forward
- self organizing maps
- deep learning
- weak learners
- learning capabilities
- multi agent
- learning classifier systems
- ensemble learning
- markov decision processes
- neural network model
- transfer learning
- linear combination
- recurrent neural networks
- training data
- formal verification
- decision processes
- classification and regression trees
- fault diagnosis