Enhancing Dataset Distillation via Non-Critical Region Refinement

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tran, Minh-Tuan, Le, Trung, Le, Xuan-May, Do, Thanh-Toan, Phung, Dinh
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913754050985984
author Tran, Minh-Tuan
Le, Trung
Le, Xuan-May
Do, Thanh-Toan
Phung, Dinh
author_facet Tran, Minh-Tuan
Le, Trung
Le, Xuan-May
Do, Thanh-Toan
Phung, Dinh
contents Dataset distillation has become a popular method for compressing large datasets into smaller, more efficient representations while preserving critical information for model training. Data features are broadly categorized into two types: instance-specific features, which capture unique, fine-grained details of individual examples, and class-general features, which represent shared, broad patterns across a class. However, previous approaches often struggle to balance these features-some focus solely on class-general patterns, neglecting finer instance details, while others prioritize instance-specific features, overlooking the shared characteristics essential for class-level understanding. In this paper, we introduce the Non-Critical Region Refinement Dataset Distillation (NRR-DD) method, which preserves instance-specific details and fine-grained regions in synthetic data while enriching non-critical regions with class-general information. This approach enables models to leverage all pixel information, capturing both feature types and enhancing overall performance. Additionally, we present Distance-Based Representative (DBR) knowledge transfer, which eliminates the need for soft labels in training by relying on the distance between synthetic data predictions and one-hot encoded labels. Experimental results show that NRR-DD achieves state-of-the-art performance on both small- and large-scale datasets. Furthermore, by storing only two distances per instance, our method delivers comparable results across various settings. The code is available at https://github.com/tmtuan1307/NRR-DD.
format Preprint
id arxiv_https___arxiv_org_abs_2503_18267
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing Dataset Distillation via Non-Critical Region Refinement
Tran, Minh-Tuan
Le, Trung
Le, Xuan-May
Do, Thanh-Toan
Phung, Dinh
Computer Vision and Pattern Recognition
Dataset distillation has become a popular method for compressing large datasets into smaller, more efficient representations while preserving critical information for model training. Data features are broadly categorized into two types: instance-specific features, which capture unique, fine-grained details of individual examples, and class-general features, which represent shared, broad patterns across a class. However, previous approaches often struggle to balance these features-some focus solely on class-general patterns, neglecting finer instance details, while others prioritize instance-specific features, overlooking the shared characteristics essential for class-level understanding. In this paper, we introduce the Non-Critical Region Refinement Dataset Distillation (NRR-DD) method, which preserves instance-specific details and fine-grained regions in synthetic data while enriching non-critical regions with class-general information. This approach enables models to leverage all pixel information, capturing both feature types and enhancing overall performance. Additionally, we present Distance-Based Representative (DBR) knowledge transfer, which eliminates the need for soft labels in training by relying on the distance between synthetic data predictions and one-hot encoded labels. Experimental results show that NRR-DD achieves state-of-the-art performance on both small- and large-scale datasets. Furthermore, by storing only two distances per instance, our method delivers comparable results across various settings. The code is available at https://github.com/tmtuan1307/NRR-DD.
title Enhancing Dataset Distillation via Non-Critical Region Refinement
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.18267