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

Knowledge-based visual question answering (VQA) is a vision-language taskthat requires an agent to correctly answer image-related questions usingknowledge that is not presented in the given image. It is not only a morechallenging task than regular VQA but also a vital step towards building ageneral VQA system. Most existing knowledge-based VQA systems process knowledgeand image information similarly and ignore the fact that the knowledge base(KB) contains complete information about a triplet, while the extracted imageinformation might be incomplete as the relations between two objects aremissing or wrongly detected. In this paper, we propose a novel model nameddynamic knowledge memory enhanced multi-step graph reasoning (DMMGR), whichperforms explicit and implicit reasoning over a key-value knowledge memorymodule and a spatial-aware image graph, respectively. Specifically, the memorymodule learns a dynamic knowledge representation and generates aknowledge-aware question representation at each reasoning step. Then, thisrepresentation is used to guide a graph attention operator over thespatial-aware image graph. Our model achieves new state-of-the-art accuracy onthe KRVQR and FVQA datasets. We also conduct ablation experiments to prove theeffectiveness of each component of the proposed model.

Dynamic Key-value Memory Enhanced Multi-step Graph Reasoning for Knowledge-based Visual Question Answering