Summary of Elucidating the Design Space Of Dataset Condensation, by Shitong Shao et al.
Elucidating the Design Space of Dataset Condensation
by Shitong Shao, Zikai Zhou, Huanran Chen, Zhiqiang Shen
First submitted to arxiv on: 21 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed Elucidate Dataset Condensation (EDC) framework efficiently transfers critical attributes from an original dataset to a synthetic version, maintaining diversity and realism while improving model training efficiency. EDC addresses limitations in previous methods by implementing soft category-aware matching and adjusting the learning rate schedule, achieving state-of-the-art accuracy on ImageNet-1k with a ResNet-18 model at a compression ratio of 0.78%. This performance exceeds SRe2L, G-VBSM, and RDED by margins of 27.3%, 17.2%, and 6.6%, respectively. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper introduces a new way to shrink big datasets into smaller ones while keeping the important information. This helps train AI models faster and better. The method, called Elucidate Dataset Condensation (EDC), works by matching similar things in the original dataset with similar things in the synthetic version. It also adjusts how fast the model learns new things. EDC is a big improvement over previous methods and can even be used on very large datasets. |
Keywords
» Artificial intelligence » Resnet
