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

Self-supervision has shown outstanding results for natural languageprocessing, and more recently, for image recognition. Simultaneously, visiontransformers and its variants have emerged as a promising and scalablealternative to convolutions on various computer vision tasks. In this paper, weare the first to question if self-supervised vision transformers (SSL-ViTs) canbe adapted to two important computer vision tasks in the low-label, high-dataregime: few-shot image classification and zero-shot image retrieval. Themotivation is to reduce the number of manual annotations required to train avisual embedder, and to produce generalizable and semantically meaningfulembeddings. For few-shot image classification we train SSL-ViTs without anysupervision, on external data, and use this trained embedder to adapt quicklyto novel classes with limited number of labels. For zero-shot image retrieval,we use SSL-ViTs pre-trained on a large dataset without any labels and fine-tunethem with several metric learning objectives. Our self-supervised attentionrepresentations outperforms the state-of-the-art on several public benchmarksfor both tasks, namely miniImageNet and CUB200 for few-shot imageclassification by up-to 6%-10%, and Stanford Online Products, Cars196 andCUB200 for zero-shot image retrieval by up-to 4%-11%. Code is available at\url{https://github.com/AutoVision-cloud/SSL-ViT-lowlabel-highdata}.

Visual Representation Learning with Self-Supervised Attention for Low-Label High-data Regime