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

Internet memes have emerged as an increasingly popular means of communicationon the Web. Although typically intended to elicit humour, they have beenincreasingly used to spread hatred, trolling, and cyberbullying, as well as totarget specific individuals, communities, or society on political,socio-cultural, and psychological grounds. While previous work has focused ondetecting harmful, hateful, and offensive memes, identifying whom they attackremains a challenging and underexplored area. Here we aim to bridge this gap.In particular, we create a dataset where we annotate each meme with itsvictim(s) such as the name of the targeted person(s), organization(s), andcommunity(ies). We then propose DISARM (Detecting vIctimS targeted by hARmfulMemes), a framework that uses named entity recognition and personidentification to detect all entities a meme is referring to, and then,incorporates a novel contextualized multimodal deep neural network to classifywhether the meme intends to harm these entities. We perform several systematicexperiments on three test setups, corresponding to entities that are (a) allseen while training, (b) not seen as a harmful target on training, and (c) notseen at all on training. The evaluation results show that DISARM significantlyoutperforms ten unimodal and multimodal systems. Finally, we show that DISARMis interpretable and comparatively more generalizable and that it can reducethe relative error rate for harmful target identification by up to 9 pointsabsolute over several strong multimodal rivals.

DISARM: Detecting the Victims Targeted by Harmful Memes