Relative stability toward diffeomorphisms indicates performance in deep nets

Abstract

Understanding why deep nets can classify data in large dimensions remains a challenge. It has been proposed that they do so by becoming stable to diffeomorphisms, yet existing empirical measurements support that it is often not the case. We revisit this question by defining a maximum-entropy distribution on diffeomorphisms, that allows to study typical diffeomorphisms of a given norm. We confirm that stability toward diffeomorphisms does not strongly correlate to performance on benchmark data sets of images. By contrast, we find that the stability toward diffeomorphisms relative to that of generic transformations $R_f$ correlates remarkably with the test error ${\epsilon }_t$. It is of order unity at initialization but decreases by several decades during training for state-of-the-art architectures. For CIFAR10 and 15 known architectures, we find ${\epsilon }_t\approx 0.2\sqrt{R_f}$, suggesting that obtaining a small $R_f$ is important to achieve good performance. We study how $R_f$ depends on the size of the training set and compare it to a simple model of invariant learning.

NeurIPS 2021 Conference

References in corpus (12)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• An Empirical Evaluation of Deep Learning on Highway Driving
• Making Convolutional Networks Shift-Invariant Again
• Spatially Transformed Adversarial Examples
• When Do Neural Networks Outperform Kernel Methods?
• Generalization Guarantees for Neural Networks via Harnessing the Low-rank Structure of the Jacobian
• Finite Versus Infinite Neural Networks: an Empirical Study
• Implicit Bias of Gradient Descent for Wide Two-layer Neural Networks Trained with the Logistic Loss
• Classifying high-dimensional Gaussian mixtures: Where kernel methods fail and neural networks succeed
• Geometric compression of invariant manifolds in neural nets
• Perspective: A Phase Diagram for Deep Learning unifying Jamming, Feature Learning and Lazy Training