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

Augmenting pretrained language models (LMs) with a vision encoder (e.g.,Flamingo) has obtained the state-of-the-art results in image-to-textgeneration. However, these models store all the knowledge within theirparameters, thus often requiring enormous model parameters to model theabundant visual concepts and very rich textual descriptions. Additionally, theyare inefficient in incorporating new data, requiring a computational-expensivefine-tuning process. In this work, we introduce a Retrieval-augmented VisualLanguage Model, Re-ViLM, built upon the Flamingo, that supports retrieving therelevant knowledge from the external database for zero and in-context few-shotimage-to-text generations. By storing certain knowledge explicitly in theexternal database, our approach reduces the number of model parameters and caneasily accommodate new data during evaluation by simply updating the database.We also construct an interleaved image and text data that facilitatesin-context few-shot learning capabilities. We demonstrate that Re-ViLMsignificantly boosts performance for image-to-text generation tasks, especiallyfor zero-shot and few-shot generation in out-of-domain settings with 4 timesless parameters compared with baseline methods.

Re-ViLM: Retrieval-Augmented Visual Language Model for Zero and Few-Shot Image Captioning