Dataset Condensation with Distribution Matching

Bo Zhao, Hakan Bilen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

Computational cost of training state-of-the-art deep models in many learning problems is rapidly increasing due to more sophisticated models and larger datasets. A recent promising direction for reducing training cost is dataset condensation that aims to replace the original large training set with a significantly smaller learned synthetic set while preserving the original information. While training deep models on the small set of condensed images can be extremely fast, their synthesis remains computationally expensive due to the complex bi-level optimization and second-order derivative computation. In this work, we propose a simple yet effective method that synthesizes condensed images by matching feature distributions of the synthetic and original training images in many sampled embedding spaces. Our method significantly reduces the synthesis cost while achieving comparable or better performance. Thanks to its efficiency, we apply our method to more realistic and larger datasets with sophisticated neural architectures and obtain a significant performance boost. We also show promising practical benefits of our method in continual learning and neural architecture search.
Original languageEnglish
Title of host publicationProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision 2023 (WACV)
Number of pages13
ISBN (Electronic)9781665493468
ISBN (Print)9781665493475
Publication statusPublished - 1 Jan 2023
EventIEEE/CVF Winter Conference on Applications of Computer Vision, 2023 - Waikoloa, United States
Duration: 3 Jan 20237 Jan 2023

Publication series

NameIEEE Winter Conference on Applications of Computer Vision
ISSN (Print)2472-6737
ISSN (Electronic)2642-9381


ConferenceIEEE/CVF Winter Conference on Applications of Computer Vision, 2023
Abbreviated titleWACV 2023
Country/TerritoryUnited States
Internet address


Dive into the research topics of 'Dataset Condensation with Distribution Matching'. Together they form a unique fingerprint.

Cite this