← Search

Stanisław Jastrzębski

2 accepted papers

2019

On the Relation Between the Sharpest Directions of DNN Loss and the SGD Step Length

ICLR 2019poster

The training of deep neural networks with Stochastic Gradient Descent (SGD) with a large learning rate or a small batch-size typically ends in flat regions of the weight space, as indicated by small eigenvalues of the Hessian of the training loss. This was found to correlate with a good final genera…

2017

A Closer Look at Memorization in Deep Networks

ICML 2017poster

We examine the role of memorization in deep learning, drawing connections to capacity, generalization, and adversarial robustness. While deep networks are capable of memorizing noise data, our results suggest that they tend to prioritize learning simple patterns first. In our experiments, we expose…

Cited by 2324SourcePDFScholar