Publications

Should Under-parameterized Student Networks Copy or Average Teacher Weights?

Berfin Şimşek, Amire Bendjeddou, Wulfram Gerstner, Johanni Brea

PDF Talk (5 mins) Poster Slides

Understanding Out-of-Distribution Accuracies through Quantifying Difficulty of Test Samples

Berfin Şimşek, Melissa Hall, Levent Sagun

Deep Linear Networks Dynamics: Low-Rank Biases Induced by Initialization Scale and L2 Regularization

Arthur Jacot, François Ged, Berfin Şimşek, Franck Gabriel, Clément Hongler

Geometry of the Loss Landscape in Overparameterized Neural Networks: Symmetries and Invariances

Berfin Şimşek, François Ged, Arthur Jacot, Francesco Spadaro, Clément Hongler, Wulfram Gerstner, Johanni Brea

PDF Poster Talk (5 mins) Talk (20 mins) Talk (1 hour) Slides

Kernel Alignment Risk Estimator: Risk Prediction from Training Data

Arthur Jacot, Berfin Şimşek, Francesco Spadaro, Clément Hongler, Franck Gabriel

Implicit Regularization of Random Feature Models

Arthur Jacot*, Berfin Şimşek*, Francesco Spadaro, Clément Hongler, Franck Gabriel

Online Bounded Component Analysis: A Simple Recurrent Neural Network with Local Update Rule for Unsupervised Separation of Dependent and Independent Sources

Berfin Şimşek, Alper T. Erdogan

Weight-Space Symmetry in Deep Networks Gives Rise to Permutation Saddles, Connected by Equal-Loss Valleys Across the Loss Landscape

Johanni Brea*, Berfin Şimşek*, Bernd Illing, Wulfram Gerstner