Semi-supervised Learning with an Imperfect Supervisor

Massih-Reza Amini 1 Patrick Gallinari 1
1 MALIRE - Machine Learning and Information Retrieval
LIP6 - Laboratoire d'Informatique de Paris 6
Abstract : Real-life applications may involve huge data sets with misclassified or partially classified training data. Semi-supervised learning and learning in the presence of label noise have recently emerged as new paradigms in the machine learning community to cope with this kind of problems. This paper describes a new discriminant algorithm for semi-supervised learning. This algorithm optimizes the classification maximum likelihood (CML) of a set of labeled–unlabeled data, using a discriminant extension of the Classification Expectation Maximization algorithm. We further propose to extend this algorithm by modeling imperfections in the estimated class labels for unlabeled data. The parameters of this label-error model are learned together with the semi-supervised classifier parameters. We demonstrate the effectiveness of the approach using extensive experiments on different datasets.
Document type :
Journal articles
Complete list of metadatas
Contributor : Lip6 Publications <>
Submitted on : Thursday, July 2, 2015 - 11:46:31 AM
Last modification on : Thursday, March 21, 2019 - 1:03:37 PM

Links full text



Massih-Reza Amini, Patrick Gallinari. Semi-supervised Learning with an Imperfect Supervisor. Knowledge and Information Systems (KAIS), Springer, 2005, 8 (4), pp.385-413. ⟨10.1007/s10115-005-0219-4⟩. ⟨hal-01170739⟩



Record views