Saved in:
Bibliographic Details
Main Authors: Li, Mang, Lyu, Wei
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.06374
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917224485224448
author Li, Mang
Lyu, Wei
author_facet Li, Mang
Lyu, Wei
contents The one-epoch overfitting problem has drawn widespread attention, especially in CTR and CVR estimation models in search, advertising, and recommendation domains. These models which rely heavily on large-scale sparse categorical features, often suffer a significant decline in performance when trained for multiple epochs. Although recent studies have proposed heuristic solutions, the fundamental cause of this phenomenon remains unclear. In this work, we present a theoretical explanation grounded in Rademacher complexity, supported by empirical experiments, to explain why overfitting occurs in models with large-scale sparse categorical features. Based on this analysis, we propose a regularization method that constrains the norm budget of embedding layers adaptively. Our approach not only prevents the severe performance degradation observed during multi-epoch training, but also improves model performance within a single epoch. This method has already been deployed in online production systems.
format Preprint
id arxiv_https___arxiv_org_abs_2511_06374
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Adaptive Regularization for Large-Scale Sparse Feature Embedding Models
Li, Mang
Lyu, Wei
Machine Learning
The one-epoch overfitting problem has drawn widespread attention, especially in CTR and CVR estimation models in search, advertising, and recommendation domains. These models which rely heavily on large-scale sparse categorical features, often suffer a significant decline in performance when trained for multiple epochs. Although recent studies have proposed heuristic solutions, the fundamental cause of this phenomenon remains unclear. In this work, we present a theoretical explanation grounded in Rademacher complexity, supported by empirical experiments, to explain why overfitting occurs in models with large-scale sparse categorical features. Based on this analysis, we propose a regularization method that constrains the norm budget of embedding layers adaptively. Our approach not only prevents the severe performance degradation observed during multi-epoch training, but also improves model performance within a single epoch. This method has already been deployed in online production systems.
title Adaptive Regularization for Large-Scale Sparse Feature Embedding Models
topic Machine Learning
url https://arxiv.org/abs/2511.06374