Retrieval-enhanced adversarial training with dynamic memory-augmented attention for image paragraph captioning

doi:10.1016/j.knosys.2020.106730

	Retrieval-enhanced adversarial training with dynamic memory-augmented attention for image paragraph captioning
	Xu, Chunpu 1; Yang, Min 1; Ao, Xiang 2; Shen, Ying 3; Xu, Ruifeng 4; Tian, Jinwen 5
刊名	KNOWLEDGE-BASED SYSTEMS
	2021-02-28
卷号	214 页码:10
关键词	Image paragraph captioning Key-value memory network Adversarial training
ISSN号	0950-7051
DOI	10.1016/j.knosys.2020.106730
英文摘要	Existing image paragraph captioning methods generate long paragraph captions solely from input images, relying on insufficient information. In this paper, we propose a retrieval-enhanced adversarial training with dynamic memory-augmented attention for image paragraph captioning (RAMP), which makes full use of the R-best retrieved candidate captions to enhance the image paragraph captioning via adversarial training. Concretely, RAMP treats the retrieved captions as reference captions to augment the discriminator during adversarial training, encouraging the image captioning model (generator) to incorporate informative content in retrieved captions into the generated caption. In addition, a retrieval-enhanced dynamic memory-augmented attention network is devised to keep track of the coverage information and attention history along with the update-chain of the decoder state, and therefore avoiding generating repetitive or incomplete image descriptions. Finally, a copying mechanism is applied to select words from the retrieved candidate captions, which are then put into the proper positions of the target caption so as to improve the fluency and informativeness of the generated caption. Extensive experiments on a benchmark dataset (i.e., Stanford) demonstrate that the proposed RAMP model significantly outperforms the state-of-the-art methods across multiple evaluation metrics. For reproducibility, we submit the code and data at https://github.com/anonymous-caption/RAMP. (C) 2020 Elsevier B.V. All rights reserved.
资助项目	National Natural Science Foundation of China[61906185] ; Natural Science Foundation of Guangdong Province of China[2019A1515011705] ; Shenzhen Science and Technology Innovation Program, China[KQTD20190929172835662] ; Youth Innovation Promotion Association of CAS China, China ; Shenzhen Basic Research Foundation, China[JCYJ20200109113441941]
WOS研究方向	Computer Science
语种	英语
出版者	ELSEVIER
WOS记录号	WOS:000618605200010
内容类型	期刊论文
源URL	[http://119.78.100.204/handle/2XEOYT63/16170]
专题	中国科学院计算技术研究所
通讯作者	Yang, Min
作者单位	1.Chinese Acad Sci, Shenzhen Inst Adv Technol, Shenzhen Key Lab High Performance Data Min, Shenzhen, Guangdong, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Beijing, Peoples R China 3.Sun Yat Sen Univ, Sch Intelligent Engn, Guangzhou, Guangdong, Peoples R China 4.Harbin Inst Technol, Shenzhen, Peoples R China 5.Huazhong Univ Sci & Technol, Wuhan, Hubei, Peoples R China
推荐引用方式 GB/T 7714	Xu, Chunpu,Yang, Min,Ao, Xiang,et al. Retrieval-enhanced adversarial training with dynamic memory-augmented attention for image paragraph captioning[J]. KNOWLEDGE-BASED SYSTEMS,2021,214:10.
APA	Xu, Chunpu,Yang, Min,Ao, Xiang,Shen, Ying,Xu, Ruifeng,&Tian, Jinwen.(2021).Retrieval-enhanced adversarial training with dynamic memory-augmented attention for image paragraph captioning.KNOWLEDGE-BASED SYSTEMS,214,10.
MLA	Xu, Chunpu,et al."Retrieval-enhanced adversarial training with dynamic memory-augmented attention for image paragraph captioning".KNOWLEDGE-BASED SYSTEMS 214(2021):10.