Online Active Learning for Drifting Data Streams

doi:10.1109/TNNLS.2021.3091681

CORC > 数学与系统科学研究院 > 中国科学院数学与系统科学研究院 > 应用数学研究所

	Online Active Learning for Drifting Data Streams
	Liu, Sanmin 4,5; Xue, Shan 5; Wu, Jia 5; Zhou, Chuan 1; Yang, Jian 5; Li, Zhao 3; Cao, Jie 2
刊名	IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
	2021-07-20
页码	15
关键词	Labeling Data models Uncertainty Biological system modeling Computational modeling Cognition Adaptation models Active learning concept drift data stream classification online incremental learning
ISSN号	2162-237X
DOI	10.1109/TNNLS.2021.3091681
英文摘要	Classification methods for streaming data are not new, but very few current frameworks address all three of the most common problems with these tasks: concept drift, noise, and the exorbitant costs associated with labeling the unlabeled instances in data streams. Motivated by this gap in the field, we developed an active learning framework based on a dual-query strategy and Ebbinghaus's law of human memory cognition. Called CogDQS, the query strategy samples only the most representative instances for manual annotation based on local density and uncertainty, thus significantly reducing the cost of labeling. The policy for discerning drift from noise and replacing outdated instances with new concepts is based on the three criteria of the Ebbinghaus forgetting curve: recall, the fading period, and the memory strength. Simulations comparing CogDQS with baselines on six different data streams containing gradual drift or abrupt drift with and without noise show that our approach produces accurate, stable models with good generalization ability at minimal labeling, storage, and computation costs.
资助项目	Nature Science Foundation of Anhui Province[1608085MF147] ; Humanities and Social Science Foundation of the Ministry of Education[18YJA630114] ; Major Project of Natural Science Research in Colleges and Universities of Anhui Province[KJ2019ZD15] ; National Natural Science Foundation of China[92046026] ; National Natural Science Foundation of China[71701089] ; International Innovation Cooperation Project of Jiangsu Province[BZ2020008]
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000733532500001
内容类型	期刊论文
源URL	[http://ir.amss.ac.cn/handle/2S8OKBNM/59709]
专题	应用数学研究所
通讯作者	Cao, Jie
作者单位	1.Chinese Acad Sci, Acad Math & Syst Sci, Beijing 100190, Peoples R China 2.Nanjing Univ Finance & Econ, Jiangsu Prov Key Lab E Business, Nanjing 210023, Peoples R China 3.Alibaba Grp, Hangzhou 310000, Peoples R China 4.Anhui Polytech Univ, Sch Comp & Informat, Wuhu 241000, Peoples R China 5.Macquarie Univ, Dept Comp, Sydney, NSW 2109, Australia
推荐引用方式 GB/T 7714	Liu, Sanmin,Xue, Shan,Wu, Jia,et al. Online Active Learning for Drifting Data Streams[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,2021:15.
APA	Liu, Sanmin.,Xue, Shan.,Wu, Jia.,Zhou, Chuan.,Yang, Jian.,...&Cao, Jie.(2021).Online Active Learning for Drifting Data Streams.IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,15.
MLA	Liu, Sanmin,et al."Online Active Learning for Drifting Data Streams".IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2021):15.