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

Vision-Language Models (VLMs), such as CLIP, exhibit strong image-textcomprehension abilities, facilitating advances in several downstream tasks suchas zero-shot image classification, image-text retrieval, and text-to-imagegeneration. However, the compositional reasoning abilities of existing VLMsremains subpar. The root of this limitation lies in the inadequate alignmentbetween the images and captions in the pretraining datasets. Additionally, thecurrent contrastive learning objective fails to focus on fine-grained groundingcomponents like relations, actions, and attributes, resulting in "bag-of-words"representations. We introduce a simple and effective method to improvecompositional reasoning in VLMs. Our method better leverages available datasetsby refining and expanding the standard image-text contrastive learningframework. Our approach does not require specific annotations and does notincur extra parameters. When integrated with CLIP, our technique yields notableimprovement over state-of-the-art baselines across five vision-languagecompositional benchmarks. We open-source our code athttps://github.com/lezhang7/Enhance-FineGrained.

Contrasting Intra-Modal and Ranking Cross-Modal Hard Negatives to Enhance Visio-Linguistic Compositional Understanding