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

The field of Few-Shot Learning (FSL), or learning from very few (typically$1$ or $5$) examples per novel class (unseen during training), has received alot of attention and significant performance advances in the recent literature.While number of techniques have been proposed for FSL, several factors haveemerged as most important for FSL performance, awarding SOTA even to thesimplest of techniques. These are: the backbone architecture (bigger isbetter), type of pre-training on the base classes (meta-training vs regularmulti-class, currently regular wins), quantity and diversity of the baseclasses set (the more the merrier, resulting in richer and better adaptivefeatures), and the use of self-supervised tasks during pre-training (serving asa proxy for increasing the diversity of the base set). In this paper we proposeyet another simple technique that is important for the few shot learningperformance - a search for a compact feature sub-space that is discriminativefor a given few-shot test task. We show that the Task-Adaptive FeatureSub-Space Learning (TAFSSL) can significantly boost the performance in FSLscenarios when some additional unlabeled data accompanies the novel few-shottask, be it either the set of unlabeled queries (transductive FSL) or someadditional set of unlabeled data samples (semi-supervised FSL). Specifically,we show that on the challenging miniImageNet and tieredImageNet benchmarks,TAFSSL can improve the current state-of-the-art in both transductive andsemi-supervised FSL settings by more than $5\%$, while increasing the benefitof using unlabeled data in FSL to above $10\%$ performance gain.

TAFSSL: Task-Adaptive Feature Sub-Space Learning for few-shot classification