Poster
Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNs
Timur Garipov · Pavel Izmailov · Dmitrii Podoprikhin · Dmitry Vetrov · Andrew Wilson
Room 517 AB #162
Keywords: [ Visualization or Exposition Techniques for Deep Networks ] [ Boosting and Ensemble Methods ] [ Computer Vision ] [ Supervised Deep Networks ]
The loss functions of deep neural networks are complex and their geometric properties are not well understood. We show that the optima of these complex loss functions are in fact connected by simple curves, over which training and test accuracy are nearly constant. We introduce a training procedure to discover these high-accuracy pathways between modes. Inspired by this new geometric insight, we also propose a new ensembling method entitled Fast Geometric Ensembling (FGE). Using FGE we can train high-performing ensembles in the time required to train a single model. We achieve improved performance compared to the recent state-of-the-art Snapshot Ensembles, on CIFAR-10, CIFAR-100, and ImageNet.
Live content is unavailable. Log in and register to view live content