A Generalization Theory of Gradient Descent for Learning Over-parameterized Deep ReLU Networks.
Yuan CaoQuanquan GuPublished in: CoRR (2019)
Keyphrases
- learning algorithm
- connectionist networks
- reinforcement learning
- learning process
- supervised learning
- learning systems
- knowledge acquisition
- active learning
- unsupervised learning
- objective function
- cost function
- data sets
- learning tasks
- learning problems
- efficient learning
- artificial intelligence
- explanation based generalization