Saved in:
Bibliographic Details
Main Authors: Mohanty, Saumyaranjan, Reddy, Aravind, Mopuri, Konda Reddy
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.13083
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911346410389504
author Mohanty, Saumyaranjan
Reddy, Aravind
Mopuri, Konda Reddy
author_facet Mohanty, Saumyaranjan
Reddy, Aravind
Mopuri, Konda Reddy
contents In Dataset Condensation, the goal is to synthesize a small dataset that replicates the training utility of a large original dataset. Existing condensation methods synthesize datasets with significant redundancy, so there is a dire need to reduce redundancy and improve the diversity of the synthesized datasets. To tackle this, we propose an intuitive Diversity Regularizer (DiRe) composed of cosine similarity and Euclidean distance, which can be applied off-the-shelf to various state-of-the-art condensation methods. Through extensive experiments, we demonstrate that the addition of our regularizer improves state-of-the-art condensation methods on various benchmark datasets from CIFAR-10 to ImageNet-1K with respect to generalization and diversity metrics.
format Preprint
id arxiv_https___arxiv_org_abs_2512_13083
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DiRe: Diversity-promoting Regularization for Dataset Condensation
Mohanty, Saumyaranjan
Reddy, Aravind
Mopuri, Konda Reddy
Computer Vision and Pattern Recognition
Machine Learning
In Dataset Condensation, the goal is to synthesize a small dataset that replicates the training utility of a large original dataset. Existing condensation methods synthesize datasets with significant redundancy, so there is a dire need to reduce redundancy and improve the diversity of the synthesized datasets. To tackle this, we propose an intuitive Diversity Regularizer (DiRe) composed of cosine similarity and Euclidean distance, which can be applied off-the-shelf to various state-of-the-art condensation methods. Through extensive experiments, we demonstrate that the addition of our regularizer improves state-of-the-art condensation methods on various benchmark datasets from CIFAR-10 to ImageNet-1K with respect to generalization and diversity metrics.
title DiRe: Diversity-promoting Regularization for Dataset Condensation
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2512.13083