Saved in:
Bibliographic Details
Main Authors: Magen, Roey, Shang, Shuning, Xu, Zhiwei, Frei, Spencer, Hu, Wei, Vardi, Gal
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.07746
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915147534041088
author Magen, Roey
Shang, Shuning
Xu, Zhiwei
Frei, Spencer
Hu, Wei
Vardi, Gal
author_facet Magen, Roey
Shang, Shuning
Xu, Zhiwei
Frei, Spencer
Hu, Wei
Vardi, Gal
contents The phenomenon of benign overfitting, where a trained neural network perfectly fits noisy training data but still achieves near-optimal test performance, has been extensively studied in recent years for linear models and fully-connected/convolutional networks. In this work, we study benign overfitting in a single-head softmax attention model, which is the fundamental building block of Transformers. We prove that under appropriate conditions, the model exhibits benign overfitting in a classification setting already after two steps of gradient descent. Moreover, we show conditions where a minimum-norm/maximum-margin interpolator exhibits benign overfitting. We study how the overfitting behavior depends on the signal-to-noise ratio (SNR) of the data distribution, namely, the ratio between norms of signal and noise tokens, and prove that a sufficiently large SNR is both necessary and sufficient for benign overfitting.
format Preprint
id arxiv_https___arxiv_org_abs_2410_07746
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Benign Overfitting in Single-Head Attention
Magen, Roey
Shang, Shuning
Xu, Zhiwei
Frei, Spencer
Hu, Wei
Vardi, Gal
Machine Learning
The phenomenon of benign overfitting, where a trained neural network perfectly fits noisy training data but still achieves near-optimal test performance, has been extensively studied in recent years for linear models and fully-connected/convolutional networks. In this work, we study benign overfitting in a single-head softmax attention model, which is the fundamental building block of Transformers. We prove that under appropriate conditions, the model exhibits benign overfitting in a classification setting already after two steps of gradient descent. Moreover, we show conditions where a minimum-norm/maximum-margin interpolator exhibits benign overfitting. We study how the overfitting behavior depends on the signal-to-noise ratio (SNR) of the data distribution, namely, the ratio between norms of signal and noise tokens, and prove that a sufficiently large SNR is both necessary and sufficient for benign overfitting.
title Benign Overfitting in Single-Head Attention
topic Machine Learning
url https://arxiv.org/abs/2410.07746