Efficient verification of neural networks based on neuron branching and LP abstraction.
Liang ZhaoXinmin DuanChenglong YangYuehao LiuYansong DongXiaobing WangWensheng WangPublished in: Neurocomputing (2024)
Keyphrases
- neural network
- pattern recognition
- high level
- artificial neural networks
- linear programming
- neural network model
- genetic algorithm
- activation function
- training algorithm
- fuzzy logic
- back propagation
- cost effective
- computationally expensive
- feedforward neural networks
- network architecture
- feasible solution
- model checking
- dynamic programming
- np hard
- special case
- machine learning