Saved in:
Bibliographic Details
Main Authors: Li, Muquan, Zhang, Dongyang, Dong, Qiang, Xie, Xiurui, Qin, Ke
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.16895
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910759186857984
author Li, Muquan
Zhang, Dongyang
Dong, Qiang
Xie, Xiurui
Qin, Ke
author_facet Li, Muquan
Zhang, Dongyang
Dong, Qiang
Xie, Xiurui
Qin, Ke
contents Contemporary deep learning, characterized by the training of cumbersome neural networks on massive datasets, confronts substantial computational hurdles. To alleviate heavy data storage burdens on limited hardware resources, numerous dataset compression methods such as dataset distillation (DD) and coreset selection have emerged to obtain a compact but informative dataset through synthesis or selection for efficient training. However, DD involves an expensive optimization procedure and exhibits limited generalization across unseen architectures, while coreset selection is limited by its low data keep ratio and reliance on heuristics, hindering its practicality and feasibility. To address these limitations, we introduce a newly versatile framework for dataset compression, namely Adaptive Dataset Quantization (ADQ). Specifically, we first identify the sub-optimal performance of naive Dataset Quantization (DQ), which relies on uniform sampling and overlooks the varying importance of each generated bin. Subsequently, we propose a novel adaptive sampling strategy through the evaluation of generated bins' representativeness score, diversity score and importance score, where the former two scores are quantified by the texture level and contrastive learning-based techniques, respectively. Extensive experiments demonstrate that our method not only exhibits superior generalization capability across different architectures, but also attains state-of-the-art results, surpassing DQ by average 3\% on various datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2412_16895
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Adaptive Dataset Quantization
Li, Muquan
Zhang, Dongyang
Dong, Qiang
Xie, Xiurui
Qin, Ke
Computer Vision and Pattern Recognition
Contemporary deep learning, characterized by the training of cumbersome neural networks on massive datasets, confronts substantial computational hurdles. To alleviate heavy data storage burdens on limited hardware resources, numerous dataset compression methods such as dataset distillation (DD) and coreset selection have emerged to obtain a compact but informative dataset through synthesis or selection for efficient training. However, DD involves an expensive optimization procedure and exhibits limited generalization across unseen architectures, while coreset selection is limited by its low data keep ratio and reliance on heuristics, hindering its practicality and feasibility. To address these limitations, we introduce a newly versatile framework for dataset compression, namely Adaptive Dataset Quantization (ADQ). Specifically, we first identify the sub-optimal performance of naive Dataset Quantization (DQ), which relies on uniform sampling and overlooks the varying importance of each generated bin. Subsequently, we propose a novel adaptive sampling strategy through the evaluation of generated bins' representativeness score, diversity score and importance score, where the former two scores are quantified by the texture level and contrastive learning-based techniques, respectively. Extensive experiments demonstrate that our method not only exhibits superior generalization capability across different architectures, but also attains state-of-the-art results, surpassing DQ by average 3\% on various datasets.
title Adaptive Dataset Quantization
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.16895