WikiWeb2M: A Page-Level Multimodal Wikipedia Dataset
This dataset addresses a gap for researchers in multimodal AI by providing structured resources for webpage tasks, though it is incremental as it builds on existing Wikipedia data.
The authors tackled the lack of a comprehensive multimodal dataset for webpage understanding by introducing WikiWeb2M, which retains full page-level images, text, and structure from Wikipedia, enabling tasks like page description generation and contextual image captioning.
Webpages have been a rich resource for language and vision-language tasks. Yet only pieces of webpages are kept: image-caption pairs, long text articles, or raw HTML, never all in one place. Webpage tasks have resultingly received little attention and structured image-text data underused. To study multimodal webpage understanding, we introduce the Wikipedia Webpage 2M (WikiWeb2M) suite; the first to retain the full set of images, text, and structure data available in a page. WikiWeb2M can be used for tasks like page description generation, section summarization, and contextual image captioning.