arrow
返回

Combat data shift in few-shot learning with knowledge graph

delete2022-08-08
delete3
delete
OA
AI
Y
Yongchun Zhu
F
Fuzhen Zhuang *
X
Xiangliang Zhang
Z
Zhiyuan Qi
Z
Zhiping Shi
J
Juan Cao
何清 (Qing He)
DOI:10.1007/s11704-022-1339-7delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
Many 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 nontrivial 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.
Keyword:
few-shot
data shift
knowledge graph

期刊

Frontiers of Computer Science 封面图
Frontiers of Computer Science
IF:
4.6
论文数:
1.6K
被引数:
2.8K

机构

B
Beihang University
学者数:
5.2W
论文数: 4.1W
被引数: 37
University of California System 封面图
University of California System
学者数:
37.5W
论文数: 33.7W
被引数: 6.6K
C
chinese academy of sciences
学者数:
56.7W
论文数: 44.9W
被引数: 704
学者 查看更多机构
引用论文

引用论文

DC discharge plasma studies for nanostructured carbon CVD
err2003-03-01
err0
PREAI
errA.N. Obraztsov; A.A. Zolotukhin; A.O. Ustinov; A.P. Volkov; Yu. Svirko; K. Jefimovs
err分享
err收藏
Quantitative Evaluation of Collagen Crosslinks and Corresponding Tensile Mechanical Properties in Mouse Cervical Tissue during Normal Pregnancy
err2014-11-14
err0
errOAAI
errKyoko Yoshida; Hongfeng Jiang; MiJung Kim; Joy Vink; Serge Cremers; David Paik; Ronald Wapner; Mala Mahendroo; Kristin Myers
err分享
err收藏
Degenerative Cervical Spondylosis
err2007-06-01
err0
PREAI
errRaj D. Rao; Bradford L. Currier; Todd J. Albert; Christopher M. Bono; Satyajit V. Marawar; Kornelis A. Poelstra; Jason C. Eck
err分享
err收藏
学者 查看更多内容