2021

Scaling Up Vision-Language Pre-training for Image Captioning

Hu, Xiaowei, Gan, Zhe, Wang, Jianfeng et al.

Understand

In recent years, we have witnessed significant performance boost in the image captioning task based on vision-language pre-training (VLP).

Scale is believed to be an important factor for this advance.
However, most existing work only focuses on pre-training transformers with moderate sizes (e.g., 12 or 24 layers) on roughly 4 million images.
In this paper, we present LEMON, a LargE-scale iMage captiONer, and provide the first empirical study on the scaling behavior of VLP for image captioning.

Built on

Im2text: Describing images using 1 million captioned photographs
Vicente Ordonez, Girish Kulkarni, and Tamara Berg · 2011
Earlier work this paper cites.
Microsoft coco captions: Data collection and evaluation server
Original
Xinlei Chen, Hao Fang, Tsung-Yi Lin, Ramakrishna Vedantam, Saurabh Gupta, Piotr Dollár, and C Lawrence Zitnick · 2015
Earlier work this paper cites.
Visual genome: Connecting language and vision using crowdsourced dense image annotations
Ranjay Krishna, Yuke Zhu, Oliver Groth, Justin Johnson, Kenji Hata, Joshua Kravitz, Stephanie Chen, Yannis Kalantidis, Li-Jia Li, David A Shamma, et al · 2017
Earlier work this paper cites.
Boosting image captioning with attributes
Ting Yao, Yingwei Pan, Yehao Li, Zhaofan Qiu, and Tao Mei · 2017
Earlier work this paper cites.
Conceptual captions: A cleaned, hypernymed, image alt-text dataset for automatic image captioning
Piyush Sharma, Nan Ding, Sebastian Goodman, and Radu Soricut · 2018
Earlier work this paper cites.
Exploring visual relationship for image captioning
Ting Yao, Yingwei Pan, Yehao Li, and Tao Mei · 2018
Earlier work this paper cites.
nocaps: novel object captioning at scale
Harsh Agrawal, Karan Desai, Yufei Wang, Xinlei Chen, Rishabh Jain, Mark Johnson, Dhruv Batra, Devi Parikh, Stefan Lee, and Peter Anderson · 2019
Earlier work this paper cites.
Bert: Pre-training of deep bidirectional transformers for language understanding
Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova · 2019
Earlier work this paper cites.
Image captioning: Transforming objects into words
Original
Simao Herdade, Armin Kappeler, Kofi Boakye, and Joao Soares · 2019
Earlier work this paper cites.
Attention on attention for image captioning
Lun Huang, Wenmin Wang, Jie Chen, and Xiao-Yong Wei · 2019
Earlier work this paper cites.
Entangled transformer for image captioning
Guang Li, Linchao Zhu, Ping Liu, and Yi Yang · 2019
Earlier work this paper cites.
Visualbert: A simple and performant baseline for vision and language
Original
Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, and Kai-Wei Chang · 2019
Earlier work this paper cites.
Vilbert: Pretraining task-agnostic visiolinguistic representations for vision-and-language tasks
Jiasen Lu, Dhruv Batra, Devi Parikh, and Stefan Lee · 2019
Earlier work this paper cites.
Contrastive bidirectional transformer for temporal representation learning
Original
Chen Sun, Fabien Baradel, Kevin Murphy, and Cordelia Schmid · 2019
Earlier work this paper cites.
Videobert: A joint model for video and language representation learning
Chen Sun, Austin Myers, Carl Vondrick, Kevin Murphy, and Cordelia Schmid · 2019
Earlier work this paper cites.
Lxmert: Learning cross-modality encoder representations from transformers
Hao Tan and Mohit Bansal · 2019
Earlier work this paper cites.
Auto-encoding scene graphs for image captioning
Xu Yang, Kaihua Tang, Hanwang Zhang, and Jianfei Cai · 2019
Earlier work this paper cites.
Language models are few-shot learners
Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al · 2020
Earlier work this paper cites.
Behind the scene: Revealing the secrets of pre-trained vision-and-language models
Jize Cao, Zhe Gan, Yu Cheng, Licheng Yu, Yen-Chun Chen, and Jingjing Liu · 2020
Earlier work this paper cites.
Uniter: Universal image-text representation learning
Yen-Chun Chen, Linjie Li, Licheng Yu, Ahmed El Kholy, Faisal Ahmed, Zhe Gan, Yu Cheng, and Jingjing Liu · 2020
Earlier work this paper cites.

Similar

Meshed-memory transformer for image captioning
Marcella Cornia, Matteo Stefanini, Lorenzo Baraldi, and Rita Cucchiara · 2020
Cited alongside, same era.
Large-scale adversarial training for vision-and-language representation learning
Zhe Gan, Yen-Chun Chen, Linjie Li, Chen Zhu, Yu Cheng, and Jingjing Liu · 2020
Cited alongside, same era.
Scaling laws for autoregressive generative modeling
Original
Tom Henighan, Jared Kaplan, Mor Katz, Mark Chen, Christopher Hesse, Jacob Jackson, Heewoo Jun, Tom B Brown, Prafulla Dhariwal, Scott Gray, et al · 2020
Cited alongside, same era.
Pixel-bert: Aligning image pixels with text by deep multi-modal transformers
Original
Zhicheng Huang, Zhaoyang Zeng, Bei Liu, Dongmei Fu, and Jianlong Fu · 2020
Cited alongside, same era.
Scaling laws for neural language models
Original
Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei · 2020
Cited alongside, same era.
Unicoder-vl: A universal encoder for vision and language by cross-modal pre-training
Gen Li, Nan Duan, Yuejian Fang, Daxin Jiang, and Ming Zhou · 2020
Cited alongside, same era.
Hero: Hierarchical encoder for video+ language omni-representation pre-training
Linjie Li, Yen-Chun Chen, Yu Cheng, Zhe Gan, Licheng Yu, and Jingjing Liu · 2020
Cited alongside, same era.
A closer look at the robustness of vision-and-language pre-trained models
Original
Linjie Li, Zhe Gan, and Jingjing Liu · 2020
Cited alongside, same era.
What does bert with vision look at?
Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, and Kai-Wei Chang · 2020
Cited alongside, same era.
Oscar: Object-semantics aligned pre-training for vision-language tasks
Xiujun Li, Xi Yin, Chunyuan Li, Pengchuan Zhang, Xiaowei Hu, Lei Zhang, Lijuan Wang, Houdong Hu, Li Dong, Furu Wei, et al · 2020
Cited alongside, same era.
12-in-1: Multi-task vision and language representation learning
Jiasen Lu, Vedanuj Goswami, Marcus Rohrbach, Devi Parikh, and Stefan Lee · 2020
Cited alongside, same era.
X-linear attention networks for image captioning
Yingwei Pan, Ting Yao, Yehao Li, and Tao Mei · 2020
Cited alongside, same era.

Then

Vivo: Surpassing human performance in novel object captioning with visual vocabulary pre-training
Xiaowei Hu, Xi Yin, Kevin Lin, Lijuan Wang, Lei Zhang, Jianfeng Gao, and Zicheng Liu · 2021
Closest in time.
Seeing out of the box: End-to-end pre-training for vision-language representation learning
Zhicheng Huang, Zhaoyang Zeng, Yupan Huang, Bei Liu, Dongmei Fu, and Jianlong Fu · 2021
Closest in time.
Wenlan: Bridging vision and language by large-scale multi-modal pre-training
Original
Yuqi Huo, Manli Zhang, Guangzhen Liu, Haoyu Lu, Yizhao Gao, Guoxing Yang, Jingyuan Wen, Heng Zhang, Baogui Xu, Weihao Zheng, et al · 2021
Closest in time.
Scaling up visual and vision-language representation learning with noisy text supervision
Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc V Le, Yunhsuan Sung, Zhen Li, and Tom Duerig · 2021
Closest in time.
Vilt: Vision-and-language transformer without convolution or region supervision
Wonjae Kim, Bokyung Son, and Ildoo Kim · 2021
Closest in time.
Less is more: Clipbert for video-and-language learning via sparse sampling
Jie Lei, Linjie Li, Luowei Zhou, Zhe Gan, Tamara L Berg, Mohit Bansal, and Jingjing Liu · 2021
Closest in time.
Align before fuse: Vision and language representation learning with momentum distillation
Junnan Li, Ramprasaath R Selvaraju, Akhilesh Deepak Gotmare, Shafiq Joty, Caiming Xiong, and Steven Hoi · 2021
Closest in time.
Unimo: Towards unified-modal understanding and generation via cross-modal contrastive learning
Wei Li, Can Gao, Guocheng Niu, Xinyan Xiao, Hao Liu, Jiachen Liu, Hua Wu, and Haifeng Wang · 2021
Closest in time.
Learning transferable visual models from natural language supervision
Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al · 2021
Closest in time.
Laion-400m: Open dataset of clip-filtered 400 million image-text pairs
Original
Christoph Schuhmann, Richard Vencu, Romain Beaumont, Robert Kaczmarczyk, Clayton Mullis, Aarush Katta, Theo Coombes, Jenia Jitsev, and Aran Komatsuzaki · 2021
Closest in time.
Wit: Wikipedia-based image text dataset for multimodal multilingual machine learning
Original
Krishna Srinivasan, Karthik Raman, Jiecao Chen, Michael Bendersky, and Marc Najork · 2021
Closest in time.
Scale efficiently: Insights from pre-training and fine-tuning transformers
Original
Yi Tay, Mostafa Dehghani, Jinfeng Rao, William Fedus, Samira Abnar, Hyung Won Chung, Sharan Narang, Dani Yogatama, Ashish Vaswani, and Donald Metzler · 2021
Closest in time.
Simvlm: Simple visual language model pretraining with weak supervision
Original
Zirui Wang, Jiahui Yu, Adams Wei Yu, Zihang Dai, Yulia Tsvetkov, and Yuan Cao · 2021
Closest in time.
Probing inter-modality: Visual parsing with self-attention for vision-language pre-training
Hongwei Xue, Yupan Huang, Bei Liu, Houwen Peng, Jianlong Fu, Houqiang Li, and Jiebo Luo · 2021
Closest in time.
Tap: Text-aware pre-training for text-vqa and text-caption
Zhengyuan Yang, Yijuan Lu, Jianfeng Wang, Xi Yin, Dinei Florencio, Lijuan Wang, Cha Zhang, Lei Zhang, and Jiebo Luo · 2021
Closest in time.
Ernie-vil: Knowledge enhanced vision-language representations through scene graph
Fei Yu, Jiji Tang, Weichong Yin, Yu Sun, Hao Tian, Hua Wu, and Haifeng Wang · 2021
Closest in time.
Scaling vision transformers
Original
Xiaohua Zhai, Alexander Kolesnikov, Neil Houlsby, and Lucas Beyer · 2021
Closest in time.
Vinvl: Revisiting visual representations in vision-language models
Pengchuan Zhang, Xiujun Li, Xiaowei Hu, Jianwei Yang, Lei Zhang, Lijuan Wang, Yejin Choi, and Jianfeng Gao · 2021
Closest in time.
Rstnet: Captioning with adaptive attention on visual and non-visual words
Xuying Zhang, Xiaoshuai Sun, Yunpeng Luo, Jiayi Ji, Yiyi Zhou, Yongjian Wu, Feiyue Huang, and Rongrong Ji · 2021
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…