Binary Stochastic Representations for Large Multi-class Classification

Handle URI:
http://hdl.handle.net/10754/626677
Title:
Binary Stochastic Representations for Large Multi-class Classification
Authors:
Gerald, Thomas; Baskiotis, Nicolas; Denoyer, Ludovic
Abstract:
Classification with a large number of classes is a key problem in machine learning and corresponds to many real-world applications like tagging of images or textual documents in social networks. If one-vs-all methods usually reach top performance in this context, these approaches suffer of a high inference complexity, linear w.r.t. the number of categories. Different models based on the notion of binary codes have been proposed to overcome this limitation, achieving in a sublinear inference complexity. But they a priori need to decide which binary code to associate to which category before learning using more or less complex heuristics. We propose a new end-to-end model which aims at simultaneously learning to associate binary codes with categories, but also learning to map inputs to binary codes. This approach called Deep Stochastic Neural Codes (DSNC) keeps the sublinear inference complexity but do not need any a priori tuning. Experimental results on different datasets show the effectiveness of the approach w.r.t. baseline methods.
Citation:
Gerald T, Baskiotis N, Denoyer L (2017) Binary Stochastic Representations for Large Multi-class Classification. Lecture Notes in Computer Science: 155–165. Available: http://dx.doi.org/10.1007/978-3-319-70087-8_17.
Publisher:
Springer International Publishing
Journal:
Lecture Notes in Computer Science
KAUST Grant Number:
OSR-2015-CRG4-2639
Issue Date:
23-Oct-2017
DOI:
10.1007/978-3-319-70087-8_17
Type:
Book Chapter
ISSN:
0302-9743; 1611-3349
Sponsors:
This publication is based upon work supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research (OSR) under Award No. OSR-2015-CRG4-2639.
Appears in Collections:
Publications Acknowledging KAUST Support

Full metadata record

DC FieldValue Language
dc.contributor.authorGerald, Thomasen
dc.contributor.authorBaskiotis, Nicolasen
dc.contributor.authorDenoyer, Ludovicen
dc.date.accessioned2018-01-04T07:51:38Z-
dc.date.available2018-01-04T07:51:38Z-
dc.date.issued2017-10-23en
dc.identifier.citationGerald T, Baskiotis N, Denoyer L (2017) Binary Stochastic Representations for Large Multi-class Classification. Lecture Notes in Computer Science: 155–165. Available: http://dx.doi.org/10.1007/978-3-319-70087-8_17.en
dc.identifier.issn0302-9743en
dc.identifier.issn1611-3349en
dc.identifier.doi10.1007/978-3-319-70087-8_17en
dc.identifier.urihttp://hdl.handle.net/10754/626677-
dc.description.abstractClassification with a large number of classes is a key problem in machine learning and corresponds to many real-world applications like tagging of images or textual documents in social networks. If one-vs-all methods usually reach top performance in this context, these approaches suffer of a high inference complexity, linear w.r.t. the number of categories. Different models based on the notion of binary codes have been proposed to overcome this limitation, achieving in a sublinear inference complexity. But they a priori need to decide which binary code to associate to which category before learning using more or less complex heuristics. We propose a new end-to-end model which aims at simultaneously learning to associate binary codes with categories, but also learning to map inputs to binary codes. This approach called Deep Stochastic Neural Codes (DSNC) keeps the sublinear inference complexity but do not need any a priori tuning. Experimental results on different datasets show the effectiveness of the approach w.r.t. baseline methods.en
dc.description.sponsorshipThis publication is based upon work supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research (OSR) under Award No. OSR-2015-CRG4-2639.en
dc.publisherSpringer International Publishingen
dc.subjectDeep learningen
dc.subjectMulti-class classificationen
dc.subjectBinary latent representationen
dc.titleBinary Stochastic Representations for Large Multi-class Classificationen
dc.typeBook Chapteren
dc.identifier.journalLecture Notes in Computer Scienceen
dc.contributor.institutionSorbonne Universités, UPMC Univ Paris 06, UMR 7606, LIP6, Paris, Franceen
kaust.grant.numberOSR-2015-CRG4-2639en
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.