Most existing video-and-language (VidL) research focuses on a single dataset, or multiple datasets of a single task. In reality, a truly useful VidL system is expected to be easily generalizable to diverse tasks, domains, and datasets. To facilitate the evaluation of such systems, we introduce Video-And-Language Understanding Evaluation (VALUE) benchmark, an assemblage of 11 VidL datasets over 3 popular tasks: (i) text-to-video retrieval; (ii) video question answering; and (iii) video captioning. VALUE benchmark aims to cover a broad range of video genres, video lengths, data volumes, and task difficulty levels. Rather than focusing on single-channel videos with visual information only, VALUE promotes models that leverage information from both video frames and their associated subtitles, as well as models that share knowledge across multiple tasks. We evaluate various baseline methods with and without large-scale VidL pre-training, and systematically investigate the impact of video input channels, fusion methods, and different video representations. We also study the transferability between tasks, and conduct multi-task learning under different settings. The significant gap between our best model and human performance calls for future study for advanced VidL models. VALUE is available at https://value-benchmark.github.io/.
@article{arxiv.2106.04632,
title = {VALUE: A Multi-Task Benchmark for Video-and-Language Understanding Evaluation},
author = {Linjie Li and Jie Lei and Zhe Gan and Licheng Yu and Yen-Chun Chen and Rohit Pillai and Yu Cheng and Luowei Zhou and Xin Eric Wang and William Yang Wang and Tamara Lee Berg and Mohit Bansal and Jingjing Liu and Lijuan Wang and Zicheng Liu},
journal= {arXiv preprint arXiv:2106.04632},
year = {2021}
}
Comments
To appear in 35th Conference on Neural Information Processing Systems (NeurIPS 2021) Track on Datasets and Benchmarks