Large-scale Dataset Pruning with Dynamic Uncertainty

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Muyang, Yang, Shuo, Huang, Tiejun, Zhao, Bo
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914833974165504
author He, Muyang
Yang, Shuo
Huang, Tiejun
Zhao, Bo
author_facet He, Muyang
Yang, Shuo
Huang, Tiejun
Zhao, Bo
contents The state of the art of many learning tasks, e.g., image classification, is advanced by collecting larger datasets and then training larger models on them. As the outcome, the increasing computational cost is becoming unaffordable. In this paper, we investigate how to prune the large-scale datasets, and thus produce an informative subset for training sophisticated deep models with negligible performance drop. We propose a simple yet effective dataset pruning method by exploring both the prediction uncertainty and training dynamics. We study dataset pruning by measuring the variation of predictions during the whole training process on large-scale datasets, i.e., ImageNet-1K and ImageNet-21K, and advanced models, i.e., Swin Transformer and ConvNeXt. Extensive experimental results indicate that our method outperforms the state of the art and achieves 25% lossless pruning ratio on both ImageNet-1K and ImageNet-21K. The code and pruned datasets are available at https://github.com/BAAI-DCAI/Dataset-Pruning.
format Preprint
id arxiv_https___arxiv_org_abs_2306_05175
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Large-scale Dataset Pruning with Dynamic Uncertainty
He, Muyang
Yang, Shuo
Huang, Tiejun
Zhao, Bo
Machine Learning
Computer Vision and Pattern Recognition
The state of the art of many learning tasks, e.g., image classification, is advanced by collecting larger datasets and then training larger models on them. As the outcome, the increasing computational cost is becoming unaffordable. In this paper, we investigate how to prune the large-scale datasets, and thus produce an informative subset for training sophisticated deep models with negligible performance drop. We propose a simple yet effective dataset pruning method by exploring both the prediction uncertainty and training dynamics. We study dataset pruning by measuring the variation of predictions during the whole training process on large-scale datasets, i.e., ImageNet-1K and ImageNet-21K, and advanced models, i.e., Swin Transformer and ConvNeXt. Extensive experimental results indicate that our method outperforms the state of the art and achieves 25% lossless pruning ratio on both ImageNet-1K and ImageNet-21K. The code and pruned datasets are available at https://github.com/BAAI-DCAI/Dataset-Pruning.
title Large-scale Dataset Pruning with Dynamic Uncertainty
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2306.05175