Recent Advances in Vision and Language PreTrained Models (VL-PTMs)
Maintained by WANG Yue (yuewang@cse.cuhk.edu.hk). Last update on 2020/10/09.
Table of Contents
- Image-based VL-PTMs
- Video-based VL-PTMs
- Speech-based VL-PTMs
- Other Transformer-based multimodal networks
- Other Resources
Image-based VL-PTMs
Representation Learning
ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Tasks, NeurIPS 2019 [code]
LXMERT: Learning Cross-Modality Encoder Representations from Transformers, EMNLP 2019 [code]
VL-BERT: Pre-training of Generic Visual-Linguistic Representations, ICLR 2020 [code]
VisualBERT: A Simple and Performant Baseline for Vision and Language, arXiv 2019/08, ACL 2020 [code]
Unicoder-VL: A Universal Encoder for Vision and Language by Cross-modal Pre-training, AAAI 2020
Unified Vision-Language Pre-Training for Image Captioning and VQA, AAAI 2020, [code], (VLP)
UNITER: Learning Universal Image-text Representations, arXiv 2019/09 [code]
Weak Supervision helps Emergence of Word-Object Alignment and improves Vision-Language Tasks, arXiv 2019/12
InterBERT: Vision-and-Language Interaction for Multi-modal Pretraining, arXiv 2020/03
Oscar: Object-Semantics Aligned Pre-training for Vision-Language Tasks, arXiv 2020/04, ECCV 2020
Pixel-BERT: Aligning Image Pixels with Text by Deep Multi-Modal Transformers, arXiv 2020/04
ERNIE-VIL: KNOWLEDGE ENHANCED VISION-LANGUAGE REPRESENTATIONS THROUGH SCENE GRAPH, arXiv 2020/06
DeVLBert: Learning Deconfounded Visio-Linguistic Representations, ACM MM 2020, [code]
SEMVLP: VISION-LANGUAGE PRE-TRAINING BY ALIGNING SEMANTICS AT MULTIPLE LEVELS, ICLR 2021 submission
Task-specific
VCR: Fusion of Detected Objects in Text for Visual Question Answering, EMNLP 2019, [code], (B2T2)
TextVQA: Iterative Answer Prediction with Pointer-Augmented Multimodal Transformers for TextVQA, CVPR 2020, [code], (M4C)
VisDial: VD-BERT: A Unified Vision and Dialog Transformer with BERT, EMNLP 2020 [code], (VD-BERT)
VisDial: Large-scale Pretraining for Visual Dialog: A Simple State-of-the-Art Baseline, ECCV 2020 [code], (VisDial-BERT)
VLN: Towards Learning a Generic Agent for Vision-and-Language Navigation via Pre-training, CVPR 2020, [code], (PREVALENT)
Text-image retrieval: ImageBERT: Cross-Modal Pre-training with Large-scale Weak-supervised Image-text Data, arXiv 2020/01
Image captioning: XGPT: Cross-modal Generative Pre-Training for Image Captioning, arXiv 2020/03
Visual Question Generation: BERT Can See Out of the Box: On the Cross-modal Transferability of Text Representations, arXiv 2020/02
Text-image retrieval: CROSS-PROBE BERT FOR EFFICIENT AND EFFECTIVE CROSS-MODAL SEARCH, ICLR 2021 submission.
Other Analysis
Multi-task Learning, 12-in-1: Multi-Task Vision and Language Representation Learning, CVPR 2020, [code]
Social Bias in VL Embedding, Measuring Social Biases in Grounded Vision and Language Embeddings, arXiv 2020/02, [code]
In-depth Analysis, Are we pretraining it right? Digging deeper into visio-linguistic pretraining, arXiv 2020/04
In-depth Analysis, Behind the Scene: Revealing the Secrets of Pre-trained Vision-and-Language Models, arXiv 2020/05, ECCV 2020 Spotlight
Adversarial Training, Large-Scale Adversarial Training for Vision-and-Language Representation Learning, arXiv 2020/06
Adaptive Analysis, Adaptive Transformers for Learning Multimodal Representations, ACL SRW 2020
Neural Architecture Search, Deep Multimodal Neural Architecture Search, arXiv 2020/04
Video-based VL-PTMs
VideoBERT: A Joint Model for Video and Language Representation Learning, ICCV 2019
Learning Video Representations Using Contrastive Bidirectional Transformers, arXiv 2019/06, (CBT)
M-BERT: Injecting Multimodal Information in the BERT Structure, arXiv 2019/08
BERT for Large-scale Video Segment Classification with Test-time Augmentation, ICCV 2019 YouTube8M workshop, [code]
Bridging Text and Video: A Universal Multimodal Transformer for Video-Audio Scene-Aware Dialog, AAAI2020 DSTC8 workshop
UniViLM: A Unified Video and Language Pre-Training Model for Multimodal Understanding and Generation, arXiv 2020/02
ActBERT: Learning Global-Local Video-Text Representations, ICCV 2020
HERO: Hierarchical Encoder for Video+Language Omni-representation Pre-training, arXiv 2020/05
Video-Grounded Dialogues with Pretrained Generation Language Models, ACL 2020
Auto-captions on GIF: A Large-scale Video-sentence Dataset for Vision-language Pre-training, arXiv 2020/07
Speech-based VL-PTMs
Towards Transfer Learning for End-to-End Speech Synthesis from Deep Pre-Trained Language Models, arXiv 2019/06
Understanding Semantics from Speech Through Pre-training, arXiv 2019/09
SpeechBERT: Cross-Modal Pre-trained Language Model for End-to-end Spoken Question Answering, arXiv 2019/10
vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations, arXiv 2019/10
Effectiveness of self-supervised pre-training for speech recognition, arXiv 2019/11
Other Transformer-based multimodal networks
Multi-Modality Cross Attention Network for Image and Sentence Matching, ICCV 2020
MART: Memory-Augmented Recurrent Transformer for Coherent Video Paragraph Captioning, ACL 2020
History for Visual Dialog: Do we really need it?, ACL 2020
Cross-Modality Relevance for Reasoning on Language and Vision, ACL 2020
Other Resources
- Two recent surveys on pretrained language models
- Pre-trained Models for Natural Language Processing: A Survey, arXiv 2020/03
- A Survey on Contextual Embeddings, arXiv 2020/03
- Other surveys about multimodal research
- Trends in Integration of Vision and Language Research: A Survey of Tasks, Datasets, and Methods, arXiv 2019
- Deep Multimodal Representation Learning: A Survey, arXiv 2019
- Multimodal Machine Learning: A Survey and Taxonomy, TPAMI 2018
- A Comprehensive Survey of Deep Learning for Image Captioning, ACM Computing Surveys 2018
- Other repositories of relevant reading list