SSGD: A safe and efficient method of gradient descent.
Jinhuan DuanXianxian LiShiqi GaoJinyan WangZili ZhongPublished in: CoRR (2020)
Keyphrases
- cost function
- synthetic data
- computationally efficient
- classification method
- theoretical analysis
- significant improvement
- high accuracy
- clustering method
- detection method
- support vector machine svm
- probabilistic model
- dynamic programming
- fully automatic
- evaluation method
- data structure
- segmentation method
- objective function
- highly efficient
- detection algorithm
- input data
- feature set
- classification accuracy
- artificial neural networks
- preprocessing
- computational complexity