Learning to Answer Visual Questions from Web Videos.

Abstract

Recent methods for visual question answering rely on large-scale annotated datasets. Manual annotation of questions and answers for videos, however, is tedious, expensive and prevents scalability. In this work, we propose to avoid manual annotation and generate a large-scale training dataset for video question answering making use of automatic cross-modal supervision. We leverage a question generation transformer trained on text data and use it to generate question-answer pairs from transcribed video narrations. Given narrated videos, we then automatically generate the HowToVQA69M dataset with 69M video-question-answer triplets. To handle the open vocabulary of diverse answers in this dataset, we propose a training procedure based on a contrastive loss between a video-question multi-modal transformer and an answer transformer. We introduce the zero-shot VideoQA task and the VideoQA feature probe evaluation setting and show excellent results. Furthermore, our method achieves competitive results on MSRVTT-QA, ActivityNet-QA, MSVD-QA and How2QA datasets. We also show that our approach generalizes to another source of web video and text data. We generate the WebVidVQA3M dataset from videos with alt-text annotations, and show its benefits for training VideoQA models. Finally, for a detailed evaluation we introduce iVQA, a new VideoQA dataset with reduced language bias and high-quality manual annotations.