Please use this identifier to cite or link to this item:
Title: Learning neural models for end-to-end clustering
Authors : Meier, Benjamin Bruno
Elezi, Ismail
Amirian, Mohammadreza
Dürr, Oliver
Stadelmann, Thilo
Proceedings: Proceedings of the 8th IAPR TC3 Workshop on Artificial Neural Networks in Pattern Recognition (ANNPR)
Conference details: 8th IAPR TC3 Workshop on Artificial Neural Networks in Pattern Recognition (ANNPR), Siena, 19-21 September 2018
Publisher / Ed. Institution : IAPR
Issue Date: 2018
License (according to publishing contract) : Not specified
Type of review: Peer review (publication)
Language : English
Subjects : Perceptual grouping; Learning to cluster; Speech & image clustering
Subject (DDC) : 004: Computer science
Abstract: We propose a novel end-to-end neural network architecture that, once trained, directly outputs a probabilistic clustering of a batch of input examples in one pass. It estimates a distribution over the number of clusters k, and for each 1 <= k <= k_max, a distribution over the individual cluster assignment for each data point. The network is trained in advance in a supervised fashion on separate data to learn grouping by any perceptual similarity criterion based on pairwise labels (same/different group). It can then be applied to different data containing different groups. We demonstrate promising performance on high-dimensional data like images (COIL-100) and speech (TIMIT). We call this “learning to cluster” and show its conceptual difference to deep metric learning, semi-supervise clustering and other related approaches while having the advantage of performing learnable clustering fully end-to-end.
Departement: School of Engineering
Organisational Unit: Institute of Applied Information Technology (InIT)
Institute of Data Analysis and Process Design (IDP)
Publication type: Conference paper
DOI : 10.21256/zhaw-3850
Appears in Collections:Publikationen School of Engineering

Files in This Item:
File Description SizeFormat 
ANNPR_2018a.pdf3.51 MBAdobe PDFThumbnail

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.