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

Recent advancements in AI have led to the development of large multimodalmodels (LMMs) capable of processing complex tasks involving joint reasoningover text and visual content in the image (e.g., navigating maps in publicplaces). This paper introduces ConTextual, a novel benchmark comprisinginstructions designed explicitly to evaluate LMMs' ability to performcontext-sensitive text-rich visual reasoning. ConTextual emphasizes diversereal-world scenarios (e.g., time-reading, navigation, shopping and more)demanding a deeper understanding of the interactions between textual and visualelements. Our findings reveal a significant performance gap of 30.8% betweenthe best-performing LMM, GPT-4V(ision), and human capabilities using humanevaluation indicating substantial room for improvement in context-sensitivetext-rich visual reasoning. Notably, while GPT-4V excelled in abstractcategories like meme and quote interpretation, its overall performance stilllagged behind humans. In addition to human evaluations, we also employedautomatic evaluation metrics using GPT-4, uncovering similar trends inperformance disparities. We also perform a fine-grained evaluation acrossdiverse visual contexts and provide qualitative analysis which provides arobust framework for future advancements in the LMM design.https://con-textual.github.io/

ConTextual: Evaluating Context-Sensitive Text-Rich Visual Reasoning in Large Multimodal Models