# Learning Invariances in Neural Networks Source: [Learning Invariances in Neural Networks]() Collection index: [llms.txt]() Data updated: 2025-10-25T22:46:52.640Z Authors: Gregory Benton, Marc Finzi, Pavel Izmailov, Andrew Gordon Wilson Published: 2020-10-22T17:18:48.000Z ## Abstract Invariances to translations have imbued convolutional neural networks with powerful generalization properties. However, we often do not know a priori what invariances are present in the data, or to what extent a model should be invariant to a given symmetry group. We show how to \\emph{learn} invariances and equivariances by parameterizing a distribution over augmentations and optimizing the training loss simultaneously with respect to the network parameters and augmentation parameters. With this simple procedure we can recover the correct set and extent of invariances on image classification, regression, segmentation, and molecular property prediction from a large space of augmentations, on training data alone. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Predictive accuracy and symmetry recovery on Double Pendulum 6D (test)](): Predictive accuracy and symmetry recovery; Double Pendulum 6D (test); MSE: 0.01053. - [Top Quark Tagging on Top Quark Tagging Elliptic Regime](): Top Quark Tagging; Top Quark Tagging Elliptic Regime; Accuracy: 74.9. ## Original sources and implementation [Original paper]() [PDF]()