Privileged information for data clustering

Many machine learning algorithms assume that all input samples are independently and identically distributed from some common distribution on either the input space X, in the case of unsupervised learning, or the input and output space X Y in the case of supervised and semi-supervised learning. I...

Full description

Bibliographic Details
Main Authors: Feyereisl, Jan, Aickelin, Uwe
Format: Article
Published: Elsevier 2012
Online Access:https://eprints.nottingham.ac.uk/2026/
_version_ 1848790707677954048
author Feyereisl, Jan
Aickelin, Uwe
author_facet Feyereisl, Jan
Aickelin, Uwe
author_sort Feyereisl, Jan
building Nottingham Research Data Repository
collection Online Access
description Many machine learning algorithms assume that all input samples are independently and identically distributed from some common distribution on either the input space X, in the case of unsupervised learning, or the input and output space X Y in the case of supervised and semi-supervised learning. In the last number of years the relaxation of this assumption has been explored and the importance of incorporation of additional information within machine learning algorithms became more apparent. Traditionally such fusion of information was the domain of semi-supervised learning. More recently the inclusion of knowledge from separate hypothetical spaces has been proposed by Vapnik as part of the supervised setting. In this work we are interested in exploring Vapnik’s idea of ‘master-class’ learning and the associated learning using ‘privileged’ information, however within the unsupervised setting. Adoption of the advanced supervised learning paradigm for the unsupervised setting instigates investigation into the dierence between privileged and technical data. By means of our proposed aRi-MAX method stability of the K-Means algorithm is improved and identification of the best clustering solution is achieved on an artificial dataset. Subsequently an information theoretic dot product based algorithm called P-Dot is proposed. This method has the ability to utilize a wide variety of clustering techniques, individually or in combination, while fusing privileged and technical data for improved clustering. Application of the P-Dot method to the task of digit recognition confirms our findings in a real-world scenario.
first_indexed 2025-11-14T18:16:54Z
format Article
id nottingham-2026
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T18:16:54Z
publishDate 2012
publisher Elsevier
recordtype eprints
repository_type Digital Repository
spelling nottingham-20262020-05-04T20:22:28Z https://eprints.nottingham.ac.uk/2026/ Privileged information for data clustering Feyereisl, Jan Aickelin, Uwe Many machine learning algorithms assume that all input samples are independently and identically distributed from some common distribution on either the input space X, in the case of unsupervised learning, or the input and output space X Y in the case of supervised and semi-supervised learning. In the last number of years the relaxation of this assumption has been explored and the importance of incorporation of additional information within machine learning algorithms became more apparent. Traditionally such fusion of information was the domain of semi-supervised learning. More recently the inclusion of knowledge from separate hypothetical spaces has been proposed by Vapnik as part of the supervised setting. In this work we are interested in exploring Vapnik’s idea of ‘master-class’ learning and the associated learning using ‘privileged’ information, however within the unsupervised setting. Adoption of the advanced supervised learning paradigm for the unsupervised setting instigates investigation into the dierence between privileged and technical data. By means of our proposed aRi-MAX method stability of the K-Means algorithm is improved and identification of the best clustering solution is achieved on an artificial dataset. Subsequently an information theoretic dot product based algorithm called P-Dot is proposed. This method has the ability to utilize a wide variety of clustering techniques, individually or in combination, while fusing privileged and technical data for improved clustering. Application of the P-Dot method to the task of digit recognition confirms our findings in a real-world scenario. Elsevier 2012 Article NonPeerReviewed Feyereisl, Jan and Aickelin, Uwe (2012) Privileged information for data clustering. Information Sciences, 194 . pp. 4-23. ISSN 0020-0255 http://dx.doi.org/10.1016/j.ins.2011.04.025 doi:10.1016/j.ins.2011.04.025 doi:10.1016/j.ins.2011.04.025
spellingShingle Feyereisl, Jan
Aickelin, Uwe
Privileged information for data clustering
title Privileged information for data clustering
title_full Privileged information for data clustering
title_fullStr Privileged information for data clustering
title_full_unstemmed Privileged information for data clustering
title_short Privileged information for data clustering
title_sort privileged information for data clustering
url https://eprints.nottingham.ac.uk/2026/
https://eprints.nottingham.ac.uk/2026/
https://eprints.nottingham.ac.uk/2026/