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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The human visual system is remarkable in learning new visual concepts fromjust a few examples. This is precisely the goal behind few-shot classincremental learning (FSCIL), where the emphasis is additionally placed onensuring the model does not suffer from "forgetting". In this paper, we pushthe boundary further for FSCIL by addressing two key questions that bottleneckits ubiquitous application (i) can the model learn from diverse modalitiesother than just photo (as humans do), and (ii) what if photos are not readilyaccessible (due to ethical and privacy constraints). Our key innovation lies inadvocating the use of sketches as a new modality for class support. The productis a "Doodle It Yourself" (DIY) FSCIL framework where the users can freelysketch a few examples of a novel class for the model to learn to recognizephotos of that class. For that, we present a framework that infuses (i)gradient consensus for domain invariant learning, (ii) knowledge distillationfor preserving old class information, and (iii) graph attention networks formessage passing between old and novel classes. We experimentally show thatsketches are better class support than text in the context of FSCIL, echoingfindings elsewhere in the sketching literature.

Doodle It Yourself: Class Incremental Learning by Drawing a Few Sketches