Can I Trust Your Answer? Visually Grounded Video Question Answering

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xiao, Junbin, Yao, Angela, Li, Yicong, Chua, Tat Seng
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913291355291648
author Xiao, Junbin
Yao, Angela
Li, Yicong
Chua, Tat Seng
author_facet Xiao, Junbin
Yao, Angela
Li, Yicong
Chua, Tat Seng
contents We study visually grounded VideoQA in response to the emerging trends of utilizing pretraining techniques for video-language understanding. Specifically, by forcing vision-language models (VLMs) to answer questions and simultaneously provide visual evidence, we seek to ascertain the extent to which the predictions of such techniques are genuinely anchored in relevant video content, versus spurious correlations from language or irrelevant visual context. Towards this, we construct NExT-GQA -- an extension of NExT-QA with 10.5$K$ temporal grounding (or location) labels tied to the original QA pairs. With NExT-GQA, we scrutinize a series of state-of-the-art VLMs. Through post-hoc attention analysis, we find that these models are extremely weak in substantiating the answers despite their strong QA performance. This exposes the limitation of current VLMs in making reliable predictions. As a remedy, we further explore and propose a grounded-QA method via Gaussian mask optimization and cross-modal learning. Experiments with different backbones demonstrate that this grounding mechanism improves both grounding and QA. With these efforts, we aim to push towards trustworthy VLMs in VQA systems. Our dataset and code are available at https://github.com/doc-doc/NExT-GQA.
format Preprint
id arxiv_https___arxiv_org_abs_2309_01327
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Can I Trust Your Answer? Visually Grounded Video Question Answering
Xiao, Junbin
Yao, Angela
Li, Yicong
Chua, Tat Seng
Computer Vision and Pattern Recognition
Artificial Intelligence
Multimedia
We study visually grounded VideoQA in response to the emerging trends of utilizing pretraining techniques for video-language understanding. Specifically, by forcing vision-language models (VLMs) to answer questions and simultaneously provide visual evidence, we seek to ascertain the extent to which the predictions of such techniques are genuinely anchored in relevant video content, versus spurious correlations from language or irrelevant visual context. Towards this, we construct NExT-GQA -- an extension of NExT-QA with 10.5$K$ temporal grounding (or location) labels tied to the original QA pairs. With NExT-GQA, we scrutinize a series of state-of-the-art VLMs. Through post-hoc attention analysis, we find that these models are extremely weak in substantiating the answers despite their strong QA performance. This exposes the limitation of current VLMs in making reliable predictions. As a remedy, we further explore and propose a grounded-QA method via Gaussian mask optimization and cross-modal learning. Experiments with different backbones demonstrate that this grounding mechanism improves both grounding and QA. With these efforts, we aim to push towards trustworthy VLMs in VQA systems. Our dataset and code are available at https://github.com/doc-doc/NExT-GQA.
title Can I Trust Your Answer? Visually Grounded Video Question Answering
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Multimedia
url https://arxiv.org/abs/2309.01327