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

Numerous visio-linguistic (V+L) representation learning methods have beendeveloped, yet existing datasets do not adequately evaluate the extent to whichthey represent visual and linguistic concepts in a unified space. We proposeseveral novel evaluation settings for V+L models, including cross-modaltransfer. Furthermore, existing V+L benchmarks often report global accuracyscores on the entire dataset, making it difficult to pinpoint the specificreasoning tasks that models fail and succeed at. We present TraVLR, a syntheticdataset comprising four V+L reasoning tasks. TraVLR's synthetic nature allowsus to constrain its training and testing distributions along task-relevantdimensions, enabling the evaluation of out-of-distribution generalisation. Eachexample in TraVLR redundantly encodes the scene in two modalities, allowingeither to be dropped or added during training or testing without losingrelevant information. We compare the performance of four state-of-the-art V+Lmodels, finding that while they perform well on test examples from the samemodality, they all fail at cross-modal transfer and have limited successaccommodating the addition or deletion of one modality. We release TraVLR as anopen challenge for the research community.

TraVLR: Now You See It, Now You Don't! A Bimodal Dataset for Evaluating Visio-Linguistic Reasoning