Unsupervised Vision-and-Language Pre-training Without Parallel Images and Captions
arXiv:2010.12831
Abstract
Pre-trained contextual vision-and-language (V&L) models have achieved impressive performance on various benchmarks. However, existing models require a large amount of parallel image-caption data for pre-training. Such data are costly to collect and require cumbersome curation. Inspired by unsupervised machine translation, we investigate if a strong V&L representation model can be learned through unsupervised pre-training without image-caption corpora. In particular, we propose to conduct ``mask-and-predict'' pre-training on text-only and image-only corpora and introduce the object tags detected by an object recognition model as anchor points to bridge two modalities. We find that such a simple approach achieves performance close to a model pre-trained with aligned data, on four English V&L benchmarks. Our work challenges the widely held notion that aligned data is necessary for V&L pre-training, while significantly reducing the amount of supervision needed for V&L models.
arXiv commentNAACL 2021 Camera Ready
References in corpus (24)
- An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- A Simple Framework for Contrastive Learning of Visual Representations
- Microsoft COCO Captions: Data Collection and Evaluation Server
- VisualBERT: A Simple and Performant Baseline for Vision and Language
- VL-BERT: Pre-training of Generic Visual-Linguistic Representations
- A Simple Method for Commonsense Reasoning
- Revisiting Unreasonable Effectiveness of Data in Deep Learning Era
- Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books
- Pixel-BERT: Aligning Image Pixels with Text by Deep Multi-Modal Transformers
- Large-Scale Adversarial Training for Vision-and-Language Representation Learning
- Unsupervised Visual Representation Learning by Context Prediction
- LXMERT: Learning Cross-Modality Encoder Representations from Transformers
- The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes
- How multilingual is Multilingual BERT?
- ERNIE-ViL: Knowledge Enhanced Vision-Language Representations Through Scene Graph
- Unicoder-VL: A Universal Encoder for Vision and Language by Cross-modal Pre-training
- MAttNet: Modular Attention Network for Referring Expression Comprehension
- Revisiting Self-Supervised Visual Representation Learning
- Modeling Context in Referring Expressions
- Unpaired Image Captioning via Scene Graph Alignments
- Weakly-supervised Visual Grounding of Phrases with Linguistic Structures
- VIVO: Visual Vocabulary Pre-Training for Novel Object Captioning
- Vokenization: Improving Language Understanding with Contextualized, Visual-Grounded Supervision
- MAF: Multimodal Alignment Framework for Weakly-Supervised Phrase Grounding