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

Recent methods for visual question answering rely on large-scale annotateddatasets. Manual annotation of questions and answers for videos, however, istedious, expensive and prevents scalability. In this work, we propose to avoidmanual annotation and generate a large-scale training dataset for videoquestion answering making use of automatic cross-modal supervision. We leveragea question generation transformer trained on text data and use it to generatequestion-answer pairs from transcribed video narrations. Given narrated videos,we then automatically generate the HowToVQA69M dataset with 69Mvideo-question-answer triplets. To handle the open vocabulary of diverseanswers in this dataset, we propose a training procedure based on a contrastiveloss between a video-question multi-modal transformer and an answertransformer. We introduce the zero-shot VideoQA task and show excellentresults, in particular for rare answers. Furthermore, we demonstrate our methodto significantly outperform the state of the art on MSRVTT-QA, MSVD-QA,ActivityNet-QA and How2QA. Finally, for a detailed evaluation we introduceiVQA, a new VideoQA dataset with reduced language biases and high-qualityredundant manual annotations. Our code, datasets and trained models areavailable at https://antoyang.github.io/just-ask.html.

Just Ask: Learning to Answer Questions from Millions of Narrated Videos