%0 Generic %1 li2017visualizing %A Li, Hao %A Xu, Zheng %A Taylor, Gavin %A Studer, Christoph %A Goldstein, Tom %D 2017 %K machinelearning neuralnetwork visualization %T Visualizing the Loss Landscape of Neural Nets %U http://arxiv.org/abs/1712.09913 %X Neural network training relies on our ability to find "good" minimizers of highly non-convex loss functions. It is well-known that certain network architecture designs (e.g., skip connections) produce loss functions that train easier, and well-chosen training parameters (batch size, learning rate, optimizer) produce minimizers that generalize better. However, the reasons for these differences, and their effects on the underlying loss landscape, are not well understood. In this paper, we explore the structure of neural loss functions, and the effect of loss landscapes on generalization, using a range of visualization methods. First, we introduce a simple "filter normalization" method that helps us visualize loss function curvature and make meaningful side-by-side comparisons between loss functions. Then, using a variety of visualizations, we explore how network architecture affects the loss landscape, and how training parameters affect the shape of minimizers.

@misc{li2017visualizing, abstract = {Neural network training relies on our ability to find "good" minimizers of highly non-convex loss functions. It is well-known that certain network architecture designs (e.g., skip connections) produce loss functions that train easier, and well-chosen training parameters (batch size, learning rate, optimizer) produce minimizers that generalize better. However, the reasons for these differences, and their effects on the underlying loss landscape, are not well understood. In this paper, we explore the structure of neural loss functions, and the effect of loss landscapes on generalization, using a range of visualization methods. First, we introduce a simple "filter normalization" method that helps us visualize loss function curvature and make meaningful side-by-side comparisons between loss functions. Then, using a variety of visualizations, we explore how network architecture affects the loss landscape, and how training parameters affect the shape of minimizers.}, added-at = {2020-01-23T18:20:45.000+0100}, author = {Li, Hao and Xu, Zheng and Taylor, Gavin and Studer, Christoph and Goldstein, Tom}, biburl = {https://www.bibsonomy.org/bibtex/2a2e4c6ef1ee6521656e8b2ed05e5b284/cpankow}, description = {[1712.09913] Visualizing the Loss Landscape of Neural Nets}, interhash = {11b08ffd9f6ee740a9a17da8e6e4c46e}, intrahash = {a2e4c6ef1ee6521656e8b2ed05e5b284}, keywords = {machinelearning neuralnetwork visualization}, note = {cite arxiv:1712.09913Comment: NIPS 2018 (extended version, 10.5 pages), code is available at https://github.com/tomgoldstein/loss-landscape}, timestamp = {2020-01-23T18:20:45.000+0100}, title = {Visualizing the Loss Landscape of Neural Nets}, url = {http://arxiv.org/abs/1712.09913}, year = 2017 }