Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Picture for Lukas Balles

Gradient-matching coresets for continual learning


Dec 09, 2021
Lukas Balles, Giovanni Zappella, Cédric Archambeau

* Accepted at the NeurIPS '21 Workshop on Distribution Shifts 

  Access Paper or Ask Questions

Self-Tuning Stochastic Optimization with Curvature-Aware Gradient Filtering


Nov 09, 2020
Ricky T. Q. Chen, Dami Choi, Lukas Balles, David Duvenaud, Philipp Hennig


  Access Paper or Ask Questions

The Geometry of Sign Gradient Descent


Feb 19, 2020
Lukas Balles, Fabian Pedregosa, Nicolas Le Roux


  Access Paper or Ask Questions

Limitations of the Empirical Fisher Approximation


May 29, 2019
Frederik Kunstner, Lukas Balles, Philipp Hennig


  Access Paper or Ask Questions

DeepOBS: A Deep Learning Optimizer Benchmark Suite


Mar 13, 2019
Frank Schneider, Lukas Balles, Philipp Hennig

* Accepted at ICLR 2019. 9 pages, 3 figures, 2 tables 

  Access Paper or Ask Questions

Dissecting Adam: The Sign, Magnitude and Variance of Stochastic Gradients


Jun 20, 2018
Lukas Balles, Philipp Hennig

* Presented at the 35th International Conference on Machine Learning (ICML), 2018 

  Access Paper or Ask Questions

Coupling Adaptive Batch Sizes with Learning Rates


Jun 28, 2017
Lukas Balles, Javier Romero, Philipp Hennig

* Thirty-Third Conference on Uncertainty in Artificial Intelligence (UAI), 2017, (accepted) 

  Access Paper or Ask Questions

Early Stopping without a Validation Set


Jun 06, 2017
Maren Mahsereci, Lukas Balles, Christoph Lassner, Philipp Hennig

* 16 pages, 10 figures 

  Access Paper or Ask Questions