Impact of Mathematical Norms on Convergence of Gradient Descent Algorithms for Deep Neural Networks Learning.
Linzhe CaiXinghuo YuChaojie LiAndrew C. EberhardLien Thuy NguyenThai Doan ChuongPublished in: AI (2022)
Keyphrases
- learning algorithm
- neural network
- learning rules
- noise tolerant
- neural nets
- learning problems
- unsupervised learning
- machine learning algorithms
- convergence analysis
- computational cost
- error back propagation
- update rule
- global convergence
- deep learning
- inductive inference
- learning tasks
- loss function
- computationally efficient
- online learning
- fuzzy logic
- significant improvement
- active learning
- pattern recognition
- feature selection