0

VideoCon: Robust Video-Language Alignment via Contrast Captions

A video-language alignment model finetuned with VideoCon, a dataset of plausible contrastive video captions, achieves superior performance in video-language tasks compared to existing models.

Year
2023
Venue
CVPR 2024 1
Authors
5
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2311.10111ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

Abstract

Despite being (pre)trained on a massive amount of data, state-of-the-art video-language alignment models are not robust to semantically-plausible contrastive changes in the video captions. Our work addresses this by identifying a broad spectrum of contrast misalignments, such as replacing entities, actions, and flipping event order, which alignment models should be robust against. To this end, we introduce the VideoCon, a video-language alignment dataset constructed by a large language model that generates plausible contrast video captions and explanations for differences between original and contrast video captions. Then, a generative video-language model is finetuned with VideoCon to assess video-language entailment and generate explanations. Our VideoCon-based alignment model significantly outperforms current models. It exhibits a 12-point increase in AUC for the video-language alignment task on human-generated contrast captions. Finally, our model sets new state of the art zero-shot performance in temporally-extensive video-language tasks such as text-to-video retrieval (SSv2-Temporal) and video question answering (ATP-Hard). Moreover, our model shows superior performance on novel videos and human-crafted captions and explanations. Our code and data are available at https://github.com/Hritikbansal/videocon.

Authors

5