Show simple item record

dc.contributor.authorZhu, Yongchun
dc.contributor.authorZhuang, Fuzhen
dc.contributor.authorZhang, Xiangliang
dc.contributor.authorQi, Zhiyuan
dc.contributor.authorShi, Zhiping
dc.contributor.authorHe, Qing
dc.date.accessioned2021-02-04T11:27:43Z
dc.date.available2021-02-04T11:27:43Z
dc.date.issued2021-01-27
dc.identifier.urihttp://hdl.handle.net/10754/667222
dc.description.abstractMany few-shot learning approaches have been designed under the meta-learning framework, which learns from a variety of learning tasks and generalizes to new tasks. These meta-learning approaches achieve the expected performance in the scenario where all samples are drawn from the same distributions (i.i.d. observations). However, in real-world applications, few-shot learning paradigm often suffers from data shift, i.e., samples in different tasks, even in the same task, could be drawn from various data distributions. Most existing few-shot learning approaches are not designed with the consideration of data shift, and thus show downgraded performance when data distribution shifts. However, it is non-trivial to address the data shift problem in few-shot learning, due to the limited number of labeled samples in each task. Targeting at addressing this problem, we propose a novel metric-based meta-learning framework to extract task-specific representations and task-shared representations with the help of knowledge graph. The data shift within/between tasks can thus be combated by the combination of task-shared and task-specific representations. The proposed model is evaluated on popular benchmarks and two constructed new challenging datasets. The evaluation results demonstrate its remarkable performance.
dc.description.sponsorshipThe research work supported by the National Key Research and Development Program of China under Grant No. 2018YFB1004300, the National Natural Science Foundation of China under Grant No. U1836206, U1811461, 61773361, the Project of Youth Innovation Promotion Association CAS under Grant No. 2017146.
dc.publisherarXiv
dc.relation.urlhttps://arxiv.org/pdf/2101.11354
dc.rightsArchived with thanks to arXiv
dc.titleCombat Data Shift in Few-shot Learning with Knowledge Graph
dc.typePreprint
dc.contributor.departmentComputer Science Program
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
dc.contributor.departmentMachine Intelligence & kNowledge Engineering Lab
dc.eprint.versionPre-print
dc.contributor.institutionKey Lab of Intelligent Information Processing of Chinese Academy of Sciences (CAS), Institute of Computing Technology, CAS, Beijing 100190, China.
dc.contributor.institutionUniversity of Chinese Academy of Sciences, Beijing 100049, China.
dc.contributor.institutionCapital Normal University.
dc.identifier.arxivid2101.11354
kaust.personZhang, Xiangliang
refterms.dateFOA2021-02-04T11:28:10Z


Files in this item

Thumbnail
Name:
Preprintfile1.pdf
Size:
812.2Kb
Format:
PDF
Description:
Pre-print

This item appears in the following Collection(s)

Show simple item record