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/

Global: Zendy Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In order to appropriately filter multi-modality data sets on a web-scale, itbecomes crucial to employ suitable filtering methods to boost performance andreduce training costs. For instance, LAION papers employs the CLIP score filterto select data with CLIP scores surpassing a certain threshold. On the otherhand, T-MARS achieves high-quality data filtering by detecting and masking textwithin images and then filtering by CLIP score. Through analyzing the dataset,we observe a significant proportion of redundant information, such as numbers,present in the textual content. Our experiments on a subset of the data unveilthe profound impact of these redundant elements on the CLIP scores. A logicalapproach would involve reevaluating the CLIP scores after eliminating theseinfluences. Experimentally, our text-based CLIP filter outperforms thetop-ranked method on the ``small scale" of DataComp (a data filteringbenchmark) on ImageNet distribution shifts, achieving a 3.6% performanceimprovement. The results also demonstrate that our proposed text-masked filteroutperforms the original CLIP score filter when selecting the top 40% of thedata. The impact of numbers on CLIP and their handling provide valuableinsights for improving the effectiveness of CLIP training, including languagerewrite techniques.

Devil in the Number: Towards Robust Multi-modality Data Filter