@inproceedings{ChatzimichailidisPfreundtGaugeretal.2019,
  author    = {Avraam Chatzimichailidis and Franz-Josef Pfreundt and Nicolas R. Gauger and Janis Keuper},
  title     = {GradVis: Visualization and Second Order Analysis of Optimization Surfaces during the Training of Deep Neural Networks},
  series = {Proceedings of MLHPC 2019: 5th Workshop on Machine Learning in HPC Environments},
  organization    = {IEEE},
  isbn      = {978-1-7281-5985-0 (Online)},
  doi       = {10.1109/MLHPC49564.2019.00012},
  pages     = {66 -- 74},
  year      = {2019},
  abstract  = {Current training methods for deep neural networks boil down to very high dimensional and non-convex optimization problems which are usually solved by a wide range of stochastic gradient descent methods. While these approaches tend to work in practice, there are still many gaps in the theoretical understanding of key aspects like convergence and generalization guarantees, which are induced by the properties of the optimization surface (loss landscape). In order to gain deeper insights, a number of recent publications proposed methods to visualize and analyze the otimization surfaces. However, the computational cost of these methods are very high, making it hardly possible to use them on larger networks. In this paper, we present the GradVis Toolbox, an open source library for efficient and scalable visualization and analysis of deep neural network loss landscapes in Tesorflow and PyTorch. Introducing more efficient mathematical formulations and a novel parallelization scheme, GradVis allows to plot 2d and 3d projections of optimization surfaces and trajectories, as well as high resolution second order gradient information for large networks.},
  language  = {en}
}