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

Without explicit feedback, humans can rapidly learn the meaning of words.Children can acquire a new word after just a few passive exposures, a processknown as fast mapping. This word learning capability is believed to be the mostfundamental building block of multimodal understanding and reasoning. Despiterecent advancements in multimodal learning, a systematic and rigorousevaluation is still missing for human-like word learning in machines. To fillin this gap, we introduce the MachinE Word Learning (MEWL) benchmark to assesshow machines learn word meaning in grounded visual scenes. MEWL covers human'score cognitive toolkits in word learning: cross-situational reasoning,bootstrapping, and pragmatic learning. Specifically, MEWL is a few-shotbenchmark suite consisting of nine tasks for probing various word learningcapabilities. These tasks are carefully designed to be aligned with thechildren's core abilities in word learning and echo the theories in thedevelopmental literature. By evaluating multimodal and unimodal agents'performance with a comparative analysis of human performance, we notice a sharpdivergence in human and machine word learning. We further discuss thesedifferences between humans and machines and call for human-like few-shot wordlearning in machines.

MEWL: Few-shot multimodal word learning with referential uncertainty