Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

multi-image Document VQA

Multi-image document visual question answering is a multimodal artificial intelligence task in which a system answers natural language questions by analyzing, locating, and reasoning across information distributed throughout a collection of multiple document images, such as presentation slide decks or multi-page reports. Unlike standard document question answering approaches that operate on a single page, this task requires models to jointly interpret text, visual elements, and spatial layouts across several interrelated images. Effectively answering questions in this setting typically requires identifying relevant evidence pages from the broader set, performing multi-hop reasoning to connect dispersed details, and executing comparative or arithmetic operations across the document collection.

1 item

SlideVQA: A Dataset for Document Visual Question Answering on Multiple Images

SlideVQA: A Dataset for Document Visual Question Answering on Multiple Images

Ryota Tanaka, Kyosuke Nishida, Kosuke Nishida, Taku Hasegawa, Itsumi Saito, Kuniko Saito

OrganizationsNTT Corporation

Why you should read this

Introduces SlideVQA, a large-scale multi-image benchmark paired with a unified sequence-to-sequence model that challenges document visual question answering systems to perform multi-hop and numerical reasoning across multi-page slide decks.

Visual question answering on document images that contain textual, visual, and layout information, called document VQA, has received much attention recently. Although many datasets have been proposed for developing document VQA systems, most of the existing datasets focus on understanding the content relationships within a single image and not across multiple images. In this study, we propose a new multi-image document VQA dataset, SlideVQA, containing 2.6k+ slide decks composed of 52k+ slide images and 14.5k questions about a slide deck. SlideVQA requires complex reasoning, including single-hop, multi-hop, and numerical reasoning, and also provides annotated arithmetic expressions of numerical answers for enhancing the ability of numerical reasoning. Moreover, we developed a new end-to-end document VQA model that treats evidence selection and question answering in a unified sequence-to-sequence format. Experiments on SlideVQA show that our model outperformed existing state-of-the-art QA models, but that it still has a large gap behind human performance. We believe that our dataset will facilitate research on document VQA.

Added

2026-09-26