Visual Superordinate Abstraction for Robust Concept Learning

doi:10.1007/s11633-022-1360-1

CORC > 自动化研究所 > 中国科学院自动化研究所 > 学术期刊 > International Journal of Automation and Computing

	Visual Superordinate Abstraction for Robust Concept Learning
	Qi Zheng 3
刊名	Machine Intelligence Research
	2023
卷号	20 期号:1 页码:79-91
关键词	Concept learning visual question answering weakly-supervised learning multi-modal learning curriculum learning
ISSN号	2731-538X
DOI	10.1007/s11633-022-1360-1
英文摘要	Concept learning constructs visual representations that are connected to linguistic semantics, which is fundamental to vision-language tasks. Although promising progress has been made, existing concept learners are still vulnerable to attribute perturbations and out-of-distribution compositions during inference. We ascribe the bottleneck to a failure to explore the intrinsic semantic hierarchy of visual concepts, e.g., {red, blue,···} “color” subspace yet cube “shape”. In this paper, we propose a visual superordinate abstraction framework for explicitly modeling semantic-aware visual subspaces (i.e., visual superordinates). With only natural visual question answering data, our model first acquires the semantic hierarchy from a linguistic view and then explores mutually exclusive visual superordinates under the guidance of linguistic hierarchy. In addition, a quasi-center visual concept clustering and superordinate shortcut learning schemes are proposed to enhance the discrimination and independence of concepts within each visual superordinate. Experiments demonstrate the superiority of the proposed framework under diverse settings, which increases the overall answering accuracy relatively by 7.5% for reasoning with perturbations and 15.6% for compositional generalization tests.
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/50901]
专题	自动化研究所_学术期刊_International Journal of Automation and Computing
作者单位	1.DATA61, Commonwealth Scientific and Industrial Research Organisation, Sydney 2122, Australia 2.JD Explore Academy, Beijing 100176, China 3.University of Sydney, Sydney 2008, Australia
推荐引用方式 GB/T 7714	Qi Zheng. Visual Superordinate Abstraction for Robust Concept Learning[J]. Machine Intelligence Research,2023,20(1):79-91.
APA	Qi Zheng.(2023).Visual Superordinate Abstraction for Robust Concept Learning.Machine Intelligence Research,20(1),79-91.
MLA	Qi Zheng."Visual Superordinate Abstraction for Robust Concept Learning".Machine Intelligence Research 20.1(2023):79-91.