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

We present a meta-learning based generative model for zero-shot learning(ZSL) towards a challenging setting when the number of training examples fromeach \emph{seen} class is very few. This setup contrasts with the conventionalZSL approaches, where training typically assumes the availability of asufficiently large number of training examples from each of the seen classes.The proposed approach leverages meta-learning to train a deep generative modelthat integrates variational autoencoder and generative adversarial networks. Wepropose a novel task distribution where meta-train and meta-validation classesare disjoint to simulate the ZSL behaviour in training. Once trained, the modelcan generate synthetic examples from seen and unseen classes. Synthesizesamples can then be used to train the ZSL framework in a supervised manner. Themeta-learner enables our model to generates high-fidelity samples using only asmall number of training examples from seen classes. We conduct extensiveexperiments and ablation studies on four benchmark datasets of ZSL and observethat the proposed model outperforms state-of-the-art approaches by asignificant margin when the number of examples per seen class is very small.

Towards Zero-Shot Learning with Fewer Seen Class Examples