SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1712.09913 · 2017

Visualizing the Loss Landscape of Neural Nets

Tom Goldstein, Hao Li, Christoph Studer, Gavin Taylor, Zheng Xu

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 8 functions out of this paper's own repositories and ran 3 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

FunctionStatusWhere it lives
ModelParameters Ran marcellodebernardi/loss-landscapes/loss_landscapes/model_interface/model_parameters.py
code served (permissive licence) · get_code("de98a260f327adb1")
NormalizeVector Ran StephenThacker/Visualiation-of-Loss-Function/PlotLossFunction.py
pointer only (licence: NONE) · get_code("ea488312c43ec2de")
perturb_params Ran activatedgeek/function-space-map/experiments/evaluate_landscape.py
code served (permissive licence) · get_code("d50731bae3d21b5a")
filter_normalize Not yet run marcellodebernardi/loss-landscapes/loss_landscapes/model_interface/model_parameters.py
code served (permissive licence) · get_code("a58cf2b132df8df9")
normalize_direction Not yet run Westlake-AI/openmixup/openmixup/utils/loss_landscape_utils.py
code served (permissive licence) · get_code("465e856d3ec6e818")
normalize_direction Not yet run JoelNiklaus/loss_landscape/net_plotter.py
code served (permissive licence) · get_code("8fda4d9878d66f73")
normalize_direction Not yet run okn-yu/Visualizing-the-Loss-Landscape-of-Neural-Nets/src/directions.py
pointer only (licence: NONE) · get_code("47f610cf6e37def7")
normalize_directions_for_weights Not yet run okn-yu/Visualizing-the-Loss-Landscape-of-Neural-Nets/src/directions.py
pointer only (licence: NONE) · get_code("89b9484af736d895")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

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 wellchosen 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.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1712.09913")
get_code_for_paper("1712.09913")
have("1712.09913")

Connect an agent — have() is free.