Generalization Guarantees for Neural Networks via Harnessing the Low-rank Structure of the Jacobian

Abstract

Modern neural network architectures often generalize well despite containing many more parameters than the size of the training dataset. This paper explores the generalization capabilities of neural networks trained via gradient descent. We develop a data-dependent optimization and generalization theory which leverages the low-rank structure of the Jacobian matrix associated with the network. Our results help demystify why training and generalization is easier on clean and structured datasets and harder on noisy and unstructured datasets as well as how the network size affects the evolution of the train and test errors during training. Specifically, we use a control knob to split the Jacobian spectum into "information" and "nuisance" spaces associated with the large and small singular values. We show that over the information space learning is fast and one can quickly train a model with zero training loss that can also generalize well. Over the nuisance space training is slower and early stopping can help with generalization at the expense of some bias. We also show that the overall generalization capability of the network is controlled by how well the label vector is aligned with the information space. A key feature of our results is that even constant width neural nets can provably generalize for sufficiently nice datasets. We conduct various numerical experiments on deep networks that corroborate our theoretical findings and demonstrate that: (i) the Jacobian of typical neural networks exhibit low-rank structure with a few large singular values and many small ones leading to a low-dimensional information space, (ii) over the information space learning is fast and most of the label vector falls on this space, and (iii) label noise falls on the nuisance space and impedes optimization/generalization.

References in corpus (3)

• Fine-Grained Analysis of Optimization and Generalization for Overparameterized Two-Layer Neural Networks
• Gradient Descent with Early Stopping is Provably Robust to Label Noise for Overparameterized Neural Networks
• Deterministic PAC-Bayesian generalization bounds for deep networks via generalizing noise-resilience

Cited by in corpus (13)

• LoRA: Low-Rank Adaptation of Large Language Models
• Gradient Descent with Early Stopping is Provably Robust to Label Noise for Overparameterized Neural Networks
• Denoising and Regularization via Exploiting the Structural Bias of Convolutional Generators
• Compressive sensing with un-trained neural networks: Gradient descent finds the smoothest approximation
• Understanding Overparameterization in Generative Adversarial Networks
• Kernelized Heterogeneous Risk Minimization
• Traces of Class/Cross-Class Structure Pervade Deep Learning Spectra
• Early Stopping in Deep Networks: Double Descent and How to Eliminate it
• Perspective: A Phase Diagram for Deep Learning unifying Jamming, Feature Learning and Lazy Training
• Edge of chaos as a guiding principle for modern neural network training
• Deep Networks Provably Classify Data on Curves
• On the Impact of Stable Ranks in Deep Nets
• A Revision of Neural Tangent Kernel-based Approaches for Neural Networks