TVLT: Textless Vision-Language Transformer
This work addresses the need for efficient and compact multimodal models that bypass text processing, potentially benefiting applications in low-resource or real-time settings, though it builds on existing masked autoencoding and contrastive learning techniques.
The paper tackles the problem of learning vision-and-language representations without relying on text, using raw visual and audio inputs in a homogeneous transformer architecture, and achieves performance comparable to text-based methods on tasks like visual question answering and retrieval with 28x faster inference and 1/3 the parameters.
In this work, we present the Textless Vision-Language Transformer (TVLT), where homogeneous transformer blocks take raw visual and audio inputs for vision-and-language representation learning with minimal modality-specific design, and do not use text-specific modules such as tokenization or automatic speech recognition (ASR). TVLT is trained by reconstructing masked patches of continuous video frames and audio spectrograms (masked autoencoding) and contrastive modeling to align video and audio. TVLT attains performance comparable to its text-based counterpart on various multimodal tasks, such as visual question answering, image retrieval, video retrieval, and multimodal sentiment analysis, with 28x faster inference speed and only 1/3 of the parameters. Our findings suggest the possibility of learning compact and efficient visual-linguistic representations from low-level visual and audio signals without assuming the prior existence of text. Our code and checkpoints are available at: https://github.com/zinengtang/TVLT