Tight Risk Bounds for Gradient Descent on Separable Data

Neural Information Processing Systems 

Recently, there has been a marked increase in interest regarding the generalization capabilities of unregularized gradient-based learning methods.

Similar Docs  Excel Report  more

TitleSimilaritySource
None found