English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Data-efficient learning has garnered significant attention, especially giventhe current trend of large multi-modal models. Recently, dataset distillationhas become an effective approach by synthesizing data samples that areessential for network training. However, it remains to be explored whichsamples are essential for the dataset distillation process itself. In thiswork, we study the data efficiency and selection for the dataset distillationtask. By re-formulating the dynamics of distillation, we provide insight intothe inherent redundancy in the real dataset, both theoretically andempirically. We propose to use the empirical loss value as a static datapruning criterion. To further compensate for the variation of the data value intraining, we find the most contributing samples based on their causal effectson the distillation. The proposed selection strategy can efficiently exploitthe training dataset, outperform the previous SOTA distillation algorithms, andconsistently enhance the distillation algorithms, even on much larger-scale andmore heterogeneous datasets, e.g., full ImageNet-1K and Kinetics-400. Webelieve this paradigm will open up new avenues in the dynamics of distillationand pave the way for efficient dataset distillation. Our code is available onhttps://github.com/silicx/GoldFromOres-BiLP.

Distill Gold from Massive Ores: Bi-level Data Pruning towards Efficient Dataset Distillation