Please use this identifier to cite or link to this item: http://scholarbank.nus.edu.sg/handle/10635/124176
Title: EFFICIENT EXTREME CLASSIFICATION WITH LABEL TAXONOMY BASED NEURAL NETWORKS
Authors: NICOLAS XAVIER MAURICE
Keywords: EXTREME,CLASSIFICATION,NEURAL,NETWORK,MACHINE,LEARNING
Issue Date: 4-Dec-2015
Citation: NICOLAS XAVIER MAURICE (2015-12-04). EFFICIENT EXTREME CLASSIFICATION WITH LABEL TAXONOMY BASED NEURAL NETWORKS. ScholarBank@NUS Repository.
Abstract: Almost every classi cation problems solved by computers count at most a few hundred of classes while Humans are able to discriminate between several thousands of categories. Also the past few years have been the witnesses of an exponential growth of the amount of data uploaded every day on the internet. In order to deal with this kind of very large datasets and create user friendly applications such as search engines, it is critical to first create machines that can e fficiently deal with a large number of categories. This paper presents an approach to Extreme Classification that relies on a Neural Network whose architecture as been drawn by following a hierarchy of labels available as prior information. The Neural Network is then trained in order to find powerful low dimensionnal embeddings fo the input space and perform the classification in low dimensional spaces.
URI: http://scholarbank.nus.edu.sg/handle/10635/124176
Appears in Collections:Master's Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
Efficient_Extreme_Classification_with_LTNN.pdf960.82 kBAdobe PDF

OPEN

NoneView/Download

Page view(s)

91
checked on Nov 16, 2018

Download(s)

82
checked on Nov 16, 2018

Google ScholarTM

Check


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