# The Low-Rank Simplicity Bias in Deep Networks

Source: [The Low-Rank Simplicity Bias in Deep Networks](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d8>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-27T06:21:27.168Z

Authors: Minyoung Huh, Hossein Mobahi, Richard Zhang, Brian Cheung, Pulkit Agrawal, Phillip Isola

Published: 2021-03-18T17:58:02.000Z

## Abstract

Modern deep neural networks are highly over-parameterized compared to the data on which they are trained, yet they often generalize remarkably well. A flurry of recent work has asked: why do deep networks not overfit to their training data? In this work, we make a series of empirical observations that investigate and extend the hypothesis that deeper networks are inductively biased to find solutions with lower effective rank embeddings. We conjecture that this bias exists because the volume of functions that maps to low effective rank embedding increases with depth. We show empirically that our claim holds true on finite width linear and non-linear models on practical learning paradigms and show that on natural data, these are often the solutions that generalize well. We then show that the simplicity bias exists at both initialization and after training and is resilient to hyper-parameters and learning methods. We further demonstrate how linear over-parameterization of deep non-linear models can be used to induce low-rank bias, improving generalization performance on CIFAR and ImageNet without changing the modeling capacity.

## 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.

- [Image Generation on ImageNet](<https://www.sota2.com/research/sota/image-generation-on-imagenet.md>): Image Generation; ImageNet; FID: 48.07.

- [Language Modeling on WikiText-103](<https://www.sota2.com/research/sota/language-modeling-on-wikitext-103.md>): Language Modeling; WikiText-103; PPL: 21.7.

- [Zero-shot Common Sense Reasoning on Common Sense Reasoning](<https://www.sota2.com/research/sota/zero-shot-common-sense-reasoning-on-common-sense-reasoning.md>): Zero-shot Common Sense Reasoning; Common Sense Reasoning; Zero-shot Accuracy: 48.4.

- [Language Modeling on WikiText-2](<https://www.sota2.com/research/sota/language-modeling-on-wikitext-2.md>): Language Modeling; WikiText-2; Perplexity (PPL): 21.92.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2103.10427>)

[PDF](<https://arxiv.org/pdf/2103.10427.pdf>)
