Diagonalwise Refactorization: An Efficient Training Method for Depthwise Convolutions.
Zheng QinZhaoning ZhangDongsheng LiYiming ZhangYuxing PengPublished in: IJCNN (2018)
Keyphrases
- cost function
- synthetic data
- training process
- significant improvement
- training samples
- similarity measure
- computational complexity
- detection method
- support vector machine
- clustering method
- theoretical analysis
- high accuracy
- convergence rate
- optimization method
- support vector machine svm
- model selection
- neural network
- probabilistic model
- experimental evaluation
- dynamic programming
- objective function
- learning algorithm