CAN-PINN: A Fast Physics-Informed Neural Network Based on Coupled-Automatic-Numerical Differentiation Method.
Pao-Hsiung ChiuJian Cheng WongChinchun OoiMy Ha DaoYew-Soon OngPublished in: CoRR (2021)
Keyphrases
- detection method
- fully automatic
- synthetic data
- high precision
- sensitivity analysis
- cost function
- experimental evaluation
- neural network
- classification accuracy
- clustering method
- segmentation method
- main contribution
- support vector machine
- computational cost
- objective function
- similarity measure
- feature set
- classification method
- optimization method
- semi automatic
- evaluation method
- numerical methods
- optimization algorithm
- support vector machine svm
- edge detection
- high accuracy
- probabilistic model
- significant improvement
- pairwise
- computer science
- genetic algorithm