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

Vision-language models such as CLIP learn a generic text-image embedding fromlarge-scale training data. A vision-language model can be adapted to a newclassification task through few-shot prompt tuning. We find that such a prompttuning process is highly robust to label noises. This intrigues us to study thekey reasons contributing to the robustness of the prompt tuning paradigm. Weconducted extensive experiments to explore this property and find the keyfactors are: 1) the fixed classname tokens provide a strong regularization tothe optimization of the model, reducing gradients induced by the noisy samples;2) the powerful pre-trained image-text embedding that is learned from diverseand generic web data provides strong prior knowledge for image classification.Further, we demonstrate that noisy zero-shot predictions from CLIP can be usedto tune its own prompt, significantly enhancing prediction accuracy in theunsupervised setting. The code is available at https://github.com/CEWu/PTNL.

Why Is Prompt Tuning for Vision-Language Models Robust to Noisy Labels?