Dealing With Multipositive Unlabeled Learning Combining Metric Learning and Deep Clustering

被引:1
作者
Racanati, Amedeo [1 ]
Esposito, Roberto [1 ]
Ienco, Dino [2 ]
机构
[1] Univ Turin, Comp Sci Dept, I-10149 Turin, Italy
[2] Univ Montpellier, UMR TETIS, INRAE, F-34090 Montpellier, France
关键词
Training; Measurement; Task analysis; Supervised learning; Standards; Reliability; Prototypes; Multi-positive unlabelled learning; weakly supervised learning; tabular data; metric learning; deep clustering; CLASSIFICATION;
D O I
10.1109/ACCESS.2022.3174590
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Standard supervised classification methods make the assumption that the training data is fully annotated thus requiring an a-priory labelling process which is both costly and time-consuming. To relax this requirement, many different flavors of weakly supervised learning have been proposed. Among weakly supervised learning strategies, Positive Unlabelled learning (PUL) is gaining attention from the research community due to the wide spectrum of applications it can fit. However, the majority of research studies related to PUL only consider binary classification tasks while real-world applications commonly involve multiple categories. To deal with this limitation, Multi-Positive Unlabelled learning (MPUL) has been recently introduced to learn from examples labelled with multiple positive labels and a single unknown negative label. Up to today, only a limited number of research works were proposed to cope with this more general setting. In this paper, we propose a new MPUL framework based on deep learning strategies. Our framework, named ProtoMPUL (Prototype based Multi-Positive and Unlabelled Learning), combines metric learning and clustering strategies to model the set of positive classes as well as to characterize the unknown negative one. Experimental evaluations on real-world benchmarks considering recent MPUL competitors demonstrates that the proposed framework achieves state-of-the-art performances, thus supporting the validity of the proposed approach.
引用
收藏
页码:51839 / 51849
页数:11
相关论文
共 36 条
  • [1] Agarap AF, 2018, ARXIV
  • [2] Aggarwal CC, 2014, CH CRC DATA MIN KNOW, P457
  • [3] HPC4AI, an AI-on-demand federated platform endeavour
    Aldinucci, Marco
    Rabellino, Sergio
    Pironti, Marco
    Spiga, Filippo
    Viviani, Paolo
    Drocco, Maurizio
    Guerzoni, Marco
    Boella, Guido
    Mellia, Marco
    Margara, Paolo
    Drago, Idilio
    Marturano, Roberto
    Marchetto, Guido
    Piccolo, Elio
    Bagnasco, Stefano
    Lusso, Stefano
    Vallero, Sara
    Attardi, Giuseppe
    Barchiesi, Alex
    Colla, Alberto
    Galeazzi, Fulvio
    [J]. 2018 ACM INTERNATIONAL CONFERENCE ON COMPUTING FRONTIERS, 2018, : 279 - 286
  • [4] Ensembles of density estimators for positive-unlabeled learning
    Basile, T. M. A.
    Di Mauro, N.
    Esposito, F.
    Ferilli, S.
    Vergari, A.
    [J]. JOURNAL OF INTELLIGENT INFORMATION SYSTEMS, 2019, 53 (02) : 199 - 217
  • [5] Learning from positive and unlabeled data: a survey
    Bekker, Jessa
    Davis, Jesse
    [J]. MACHINE LEARNING, 2020, 109 (04) : 719 - 760
  • [6] Novel Cytonuclear Combinations Modify Arabidopsis thaliana Seed Physiology and Vigor
    Boussardon, Clement
    Martin-Magniette, Marie-Laure
    Godin, Beatrice
    Benamar, Abdelilah
    Vittrant, Benjamin
    Citerne, Sylvie
    Mary-Huard, Tristan
    Macherel, David
    Rajjou, Loic
    Budar, Francoise
    [J]. FRONTIERS IN PLANT SCIENCE, 2019, 10
  • [7] Multiple instance learning: A survey of problem characteristics and applications
    Carbonneau, Marc-Andre
    Cheplygina, Veronika
    Granger, Eric
    Gagnon, Ghyslain
    [J]. PATTERN RECOGNITION, 2018, 77 : 329 - 353
  • [8] CLAESEN M, 2015, ARXIV150407389
  • [9] SUPPORT-VECTOR NETWORKS
    CORTES, C
    VAPNIK, V
    [J]. MACHINE LEARNING, 1995, 20 (03) : 273 - 297
  • [10] Demsar J, 2006, J MACH LEARN RES, V7, P1