[1]张远健,赵天娜,苗夺谦.基于粒的标记增强标记分布学习[J].智能系统学报,2023,18(2):390-398.[doi:10.11992/tis.202208015]
ZHANG Yuanjian,ZHAO Tianna,MIAO Duoqian.Granule-based label enhancement in label distribution learning[J].CAAI Transactions on Intelligent Systems,2023,18(2):390-398.[doi:10.11992/tis.202208015]
点击复制
ZHANG Yuanjian,ZHAO Tianna,MIAO Duoqian.Granule-based label enhancement in label distribution learning[J].CAAI Transactions on Intelligent Systems,2023,18(2):390-398.[doi:10.11992/tis.202208015]
基于粒的标记增强标记分布学习
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
18
期数:
2023年第2期
页码:
390-398
栏目:
吴文俊人工智能科学技术奖论坛
出版日期:
2023-05-05
- Title:
- Granule-based label enhancement in label distribution learning
- Keywords:
- granular computing; label distribution learning; label enhancement; multi-label; uncertainty; local label correlation; clustering; topology
- 分类号:
- TP391
- 摘要:
- 标记分布学习能有效求解多标记学习任务，然而分类器构造以获得大规模具有更强监督信息的标注为前提，在许多应用中难以满足。一种替代的方案是以标记增强的方式从传统逻辑形式的标注中挖掘出隐含的数值型标记的重要程度。现有的标记增强方法大多假设增强后的标记需要在所有示例上保持原有逻辑标记的相关性，不能有效保持局部标记相关性。基于粒计算理论，提出了一种适用于标记分布学习的粒化标记增强学习方法。该方法通过k均值聚类构造具有局部相关性语义的信息粒，并在粒的抽象层面上，分别在图上依据逻辑标记的特性和属性空间的拓扑性质完成粒内示例的标记转化。最后，将得到的标记分布在示例层面进行融合，得到描述整个数据集标记重要程度的数值型标记。大量比较研究表明，所提出的模型可以显著地提升多标记学习的性能。
- Abstract:
- Label distribution learning can effectively deal with multilabel learning tasks. However, the construction of a classifier is based on the premise of obtaining large-scale labels with strong supervision information, which is difficult to be satisfied in many applications. An alternative solution is to mine the importance of implicit numerical labels from the traditional logical form of annotation through label enhancement. Existing label enhancement methods mainly assume that the enhanced label must maintain the relevance of the original logical label in all instances, which fails to preserve local label correlation. This paper proposes a granular-based label enhancement distribution model applicable to label distribution learning, considering the methodology of granular computing. The method constructs information granules with local correlation semantics by employing k-means clustering and completes the labeling transformation of instances in granules on the graph according to the characteristics of logical labeling and the topological properties of attribute space at the abstract level of granules. Finally, the obtained label distribution is fused at the instance level, obtaining the numerical label describing the importance of the whole data set. Extensive studies have shown that the proposed model significantly improves the accuracy of multilabel learning.
- 参考文献/References:
-
[1] ZHANG Minling, ZHOU Zhihua. A review on multi-label learning algorithms[J]. IEEE transactions on knowledge and data engineering, 2014, 26(8): 1819–1837.
[2] GIBAJA E, VENTURA S. A tutorial on multilabel learning[J]. ACM computing surveys, 2015, 47(3): 52.
[3] 耿新, 徐宁. 标记分布学习与标记增强[J]. 中国科学:信息科学, 2018, 48(5): 521–530
GENG Xin, XU Ning. Label distribution learning and label enhancement[J]. Scientia sinica (informationis), 2018, 48(5): 521–530
[4] YAO Yiyang, WANG Luo, ZHANG Luming, et al. Learning latent stable patterns for image understanding with weak and noisy labels[J]. IEEE transactions on cybernetics, 2019, 49(12): 4243–4252.
[5] 曾雪强, 华鑫, 刘平生, 等. 基于情感轮和情感词典的文本情感分布标记增强方法[J]. 计算机学报, 2021, 44(6): 1080–1094
ZENG Xueqiang, HUA Xin, LIU Pingsheng, et al. Emotion wheel and lexicon based text emotion distribution label enhancement method[J]. Chinese journal of computers, 2021, 44(6): 1080–1094
[6] GENG Xin. Label distribution learning[J]. IEEE transactions on knowledge and data engineering, 2016, 28(7): 1734–1748.
[7] SHAO Ruifeng, XU Ning, GENG Xin. Multi-label learning with label enhancement[C]//IEEE International Conference on Data Mining. Singapore: IEEE, 2018: 437-446.
[8] 熊传镇，钱文彬，王映龙. 基于标记增强和模糊辨识度的标记分布特征选择[J]. 数据采集与处理, 2021, 36(3): 529-543.
XIONG Chuanzhen, QIAN Wenbin, WANG Yinglong. Label enhancement and fuzzy discernibility based label distribution feature selection[J]. Journal of data acquisition and processing, 2021, 36(3): 529-543.
[9] XU Ning, LIU Yunpeng, GENG Xin. Label enhancement for label distribution learning[J]. IEEE transactions on knowledge and data engineering, 2021, 33(4): 1632–1643.
[10] EL GAYAR N, SCHWENKER F, PALM G. A study of the robustness of KNN classifiers trained using soft labels[M]//Artificial Neural Networks in Pattern Recognition. Berlin: Springer Berlin, 2006: 67?80.
[11] JIANG Xiufeng, YI Zhang, LYU Jiancheng. Fuzzy SVM with a new fuzzy membership function[J]. Neural computing and applications, 2006, 15(3/4): 268–276.
[12] LI Yukun, ZHANG Minling, GENG Xin. Leveraging implicit relative labeling-importance information for effective multi-label learning[J]. 2015 IEEE international conference on data mining, 2015: 251?260.
[13] HOU Peng, GENG Xin, ZHANG Minling. Multi-label manifold learning[J]. Proceedings of the AAAI conference on artificial intelligence, 2016, 30(1): 1680–1686.
[14] YAO Jing tao, VASILAKOS A V, PEDRYCZ W. Granular computing: perspectives and challenges[J]. IEEE transactions on cybernetics, 2013, 43(6): 1977–1989.
[15] 刘清，邱桃荣，刘澜. 基于非标准分析的粒计算研究[J]. 计算机学报, 2015, 38(8): 1618?1627.
LIU Qing, QIU Taorong, LIU Lan. The research of granular computing based on nonstandard analysis[J]. Chinese journal of computers, 2015, 38(8): 1618?1627.
[16] PEDRYCZ W. Granular computing for data analytics: a manifesto of human-centric computing[J]. IEEE/CAA journal of automatica sinica, 2018, 5(6): 1025–1034.
[17] 徐健锋,苗夺谦,张远健. 分段延迟代价敏感三支决策[J]. 软件学报, 2022, 33(10): 3754–3775
XU Jianfeng, MIAO Duoqian, ZHANG Yuanjian. Piece-wise delay cost-sensitive three-way decisions[J]. Journal of software, 2022, 33(10): 3754–3775
[18] 苗夺谦，高阳，吴伟志，等. 粒计算与知识发现白皮书[R]. 北京：中国人工智能学会, 2022.
MIAO Duoqian, GAO Yang, WU Weizhi, et al. White paper for granular computing and knowledge discovery[R]. Beijing: Chinese Association for Artificial Intelligence, 2022.
[19] ZHANG Yuanjian, MIAO Duoqian, ZHANG Zhifei, et al. A three-way selective ensemble model for multi-label classification[J]. International journal of approximate reasoning, 2018, 103: 394–413.
[20] ZHANG Yuanjian, MIAO Duoqian, PEDRYCZ W, et al. Granular structure-based incremental updating for multi-label classification[J]. Knowledge-based systems, 2020, 189: 105066.
[21] ZHANG Yuanjian, ZHAO Tianna, MIAO Duoqian, et al. Granular multilabel batch active learning with pairwise label correlation[J]. IEEE transactions on systems, man, and cybernetics:systems, 2022, 52(5): 3079–3091.
[22] HUANG Shengjun, ZHOU Zhihua. Multi-label learning by exploiting label correlations locally[C]//AAAI’12: Proceedings of the Twenty-Sixth AAAI Conference on Artificial Intelligence. New York: ACM, 2012: 949?955.
[23] ZHENG Xiang, JIA Xiuyi, LI Weiwei. Label distribution learning by exploiting sample correlations locally[J]. Proceedings of the AAAI conference on artificial intelligence, 2018, 32(1): 3310–3317.
[24] JIA Xiuyi, LI Zechao, ZHENG Xiang, et al. Label distribution learning with label correlations on local samples[J]. IEEE transactions on knowledge and data engineering, 2021, 33(4): 1619–1631.
[25] ZHANG Jujie, FANG Min, LI Xiao. Clustered intrinsic label correlations for multi-label classification[J]. Expert systems with applications, 2017, 81: 134–146.
[26] GENG Xin, YIN Chao, ZHOU Zhihua. Facial age estimation by learning from label distributions[J]. IEEE transactions on pattern analysis and machine intelligence, 2013, 35(10): 2401–2412.
- 相似文献/References:
-
[1]王国胤,张清华,胡? 军.粒计算研究综述[J].智能系统学报,2007,2(6):8.
WANG Guo-yin,ZHANG Qing-hua,HU Jun.An overview of granular computing[J].CAAI Transactions on Intelligent Systems,2007,2():8.
[2]周丹晨.采用粒计算的属性权重确定方法[J].智能系统学报,2015,10(2):273.[doi:10.3969/j.issn.1673-4785.201312008]
ZHOU Danchen.A method for ascertaining the weight of attributes based on granular computing[J].CAAI Transactions on Intelligent Systems,2015,10():273.[doi:10.3969/j.issn.1673-4785.201312008]
[3]李峰,苗夺谦,刘财辉,等.基于决策粗糙集的图像分割[J].智能系统学报,2014,9(2):143.[doi:10.3969/j.issn.1673-4785.201307022]
LI Feng,MIAO Duoqian,LIU Caihui,et al.Image segmentation algorithm based on the decision-theoretic rough set model[J].CAAI Transactions on Intelligent Systems,2014,9():143.[doi:10.3969/j.issn.1673-4785.201307022]
[4]汤建国,汪江桦,韩莉英,等.基于覆盖粗糙集的语言动力系统[J].智能系统学报,2014,9(2):229.[doi:10.3969/j.issn.1673-4785.201307018]
TANG Jianguo,WANG Jianghua,HAN Liying,et al.Linguistic dynamic systems based on covering-based rough sets[J].CAAI Transactions on Intelligent Systems,2014,9():229.[doi:10.3969/j.issn.1673-4785.201307018]
[5]刘翠君,赵才荣,苗夺谦,等.粒化的Mean Shift行人跟踪算法[J].智能系统学报,2016,11(4):433.[doi:10.11992/tis.201605033]
LIU Cuijun,ZHAO Cairong,MIAO Duoqian,et al.Granular mean shift pedestrian tracking algorithm[J].CAAI Transactions on Intelligent Systems,2016,11():433.[doi:10.11992/tis.201605033]
[6]苗夺谦,张清华,钱宇华,等.从人类智能到机器实现模型——粒计算理论与方法[J].智能系统学报,2016,11(6):743.[doi:10.11992/tis.201612014]
MIAO Duoqian,ZHANG Qinghua,QIAN Yuhua,et al.From human intelligence to machine implementation model: theories and applications based on granular computing[J].CAAI Transactions on Intelligent Systems,2016,11():743.[doi:10.11992/tis.201612014]
[7]王映龙,曾淇,钱文彬,等.变精度下不完备邻域决策系统的属性约简算法[J].智能系统学报,2017,12(3):386.[doi:10.11992/tis.201705027]
WANG Yinglong,ZENG Qi,QIAN Wenbin,et al.Attribute reduction algorithm of the incomplete neighborhood decision system with variable precision[J].CAAI Transactions on Intelligent Systems,2017,12():386.[doi:10.11992/tis.201705027]
[8]徐健锋,何宇凡,汤涛,等.概率粗糙集三支决策在线快速计算算法研究[J].智能系统学报,2018,13(5):741.[doi:10.11992/tis.201706047]
XU Jianfeng,HE Yufan,TANG Tao,et al.Research on a fast online computing algorithm based on three-way decisions with probabilistic rough sets[J].CAAI Transactions on Intelligent Systems,2018,13():741.[doi:10.11992/tis.201706047]
[9]刘盾,李天瑞,梁德翠,等.三支决策的时空性[J].智能系统学报,2019,14(1):141.[doi:10.11992/tis.201804045]
LIU Dun,LI Tianrui,LIANG Decui,et al.Temporality and spatiality of three-way decisions[J].CAAI Transactions on Intelligent Systems,2019,14():141.[doi:10.11992/tis.201804045]
[10]黄琴,钱文彬,王映龙,等.代价敏感数据的多标记特征选择算法[J].智能系统学报,2019,14(5):929.[doi:10.11992/tis.201807027]
HUANG Qin,QIAN Wenbin,WANG Yinglong,et al.Multi-label feature selection algorithm for cost-sensitive data[J].CAAI Transactions on Intelligent Systems,2019,14():929.[doi:10.11992/tis.201807027]
备注/Memo
收稿日期:2022-08-11。
基金项目:中国博士后科学基金资助项目(2022M713491)；国家自然科学基金项目(61976158)
作者简介:张远健,博士,中国银联股份有限公司博士后,中国计算机学会会员,主要研究方向为多标记分类、粒计算、联邦学习,主持中国博士后面上基金1项。发表学术论文10余篇;赵天娜,博士研究生,中国人工智能学会会员,主要研究方向为标记分布学习、粒计算、不确定性。发表学术论文7篇;苗夺谦,教授,博士,国际粗糙集学会理事长、中国人工智能学会会士、中国计算机学会杰出会员,主要研究方向为粒计算、不确定性、大数据分析。荣获中国人工智能学会吴文俊人工智能自然科学二等奖1项;主持国家自然科学基金面上项目7项,出版教材和学术著作10余部。发表学术论文180余篇,ESI高被引8篇
通讯作者:张远健. E-mail：zhangyuanjian@unionpay.com
更新日期/Last Update:
1900-01-01