arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2212.03490cs.CV
收藏

SimVTP:使用掩码自编码器的简单视频文本预训练

SimVTP: Simple Video Text Pre-training with Masked Autoencoders

Tencent AI Lab(腾讯人工智能实验室);Tsinghua Shenzhen International Graduate School, Tsinghua University(清华大学深圳国际研究生院)

Yue Ma, Tianyu Yang, Yin Shan, Xiu Li

摘要:

本文提出了SimVTP:一个简单的视频-文本预训练框架通过掩码自编码器。我们随机屏蔽输入视频的时空管和输入文本的单词标记,然后将它们送入统一的autencoder以重建丢失的像素和单词。我们的SimVTP有几个属性:1)由于统一的自动编码器,SimVTP在另一种模态的帮助下重建一种模态的掩蔽信号,这隐含地学习了视频管和文本标记之间的跨模态对齐。2)SimVTP不仅由于视频的时间冗余而受益于高视频掩蔽比(例如90%),而且还需要高文本掩蔽比(例如75%),其远高于BERT(例如15%),以实现最佳性能。这是因为视频模态的辅助使得文本重建的挑战性降低,因此需要更高的掩模比率来使有用特征学习的借口更难。3)为SimVTP配备两种常用的跨模态训练策略:视频-文本对比学习(VTC)和视频-文本匹配(VTM),可以进一步显著提高SimVTP的可迁移性能。4)SimVTP是数据高效的,例如,仅在WebVid-2 M的10%数据上进行预训练,SimVTP在MSRVTT上实现了令人惊讶的良好结果(43.8 R@1),这远高于在CC 3 M和WebVid-2 M上预训练的最新技术水平的方法。我们将预先训练的模型转移到各种下游任务中,并实现卓越的性能。代码和模型将在https://github.com/mayuelala/SimVTP上发布。

英文摘要:

This paper presents SimVTP: a Simple Video-Text Pretraining framework via masked autoencoders. We randomly mask out the spatial-temporal tubes of input video and the word tokens of input text and then feed them into a unified autencoder to reconstruct the missing pixels and words. Our SimVTP has several properties: 1) Thanks to the unified autoencoder, SimVTP reconstructs the masked signal of one modality with the help from another modality, which implicitly learns the cross-modal alignment between video tubes and text tokens. 2) SimVTP not only benefits from a high video masking ratio (e.g. 90%) due to the temporal redundancy of video, but also needs a high text masking ratio (e.g. 75%), which is much higher than BERT (e.g. 15%), to achieve optimal performance. This is because the aid of video modality makes text reconstruction less challenging, which thus needs a higher mask ratio to make the pretext harder for useful feature learning. 3) Equipping SimVTP with video-text contrastive learning (VTC) and video-text matching (VTM), which are two commonly used cross-modal training strategies, could further improve the transferable performance significantly. 4) SimVTP is dataefficent, e.g., pre-training only on 10% data of WebVid-2M, SimVTP achieves surprisingly good results (43.8 R@1) on MSRVTT, which is far above recent state-of-the-art methods pre-trained on both CC3M and WebVid-2M. We transfer our pre-trained model to various downstream tasks and achieve superior performance. The codes and models will be released at https://github.com/mayuelala/SimVTP.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑