Access count of this item: 262

Files in This Item:
File Description SizeFormat 
yjohk00596.pdfAbstract_要旨140.53 kBAdobe PDFView/Open
djohk00596.pdfDissertation_全文14.19 MBAdobe PDFView/Open
Title: Geometry-Aware Learning Algorithms for Histogram Data Using Adaptive Metric Embeddings and Kernel Functions
Other Titles: 距離の適応埋込みとカーネル関数を用いたヒストグラムデータからの幾何認識学習アルゴリズム
Authors: Le, Thanh Tam
Keywords: Metric learning
Kernel method
Aitchison geometry
the probability simplex
Bag of features
Riemannian manifold
Issue Date: 25-Jan-2016
Publisher: 京都大学 (Kyoto University)
Conferring University: Kyoto University (京都大学)
Degree Level: 新制・課程博士
Degree Discipline: 博士(情報学)
Degree Report no.: 甲第19417号
Degree no.: 情博第596号
Degree Call no.: 新制||情||104(附属図書館)
Degree Serial no.: 32442
Degree Affiliation: 京都大学大学院情報学研究科知能情報学専攻
Examination Committee members: (主査)教授 山本 章博, 教授 黒橋 禎夫, 教授 鹿島 久嗣, 准教授 Cuturi Marco
Provisions of the Ruling of Degree: 学位規則第4条第1項該当
Rights: The content of Chapter 2 was published in Journal of Machine Learning Workshop and Conference Proceedings Vol. 29: Asian Conference on Machine Learning (ACML), pp. 293-308, 2013 and Machine Learning Journal (MLJ), pp. 169-187, 2014. Springer and the Machine Learning Journal, volume 99, 2015, pp. 169-187, Adaptive Euclidean Maps for Histograms: Generalized Aitchison Embeddings, Tam Le and Marco Cuturi, original copyright notice is given to the publication in which the material was originally published, by adding; with kind permission from Springer Science and Business Media. The content of Chapter 3 was published in Journal of Machine Learning Research Workshop and Conference Proceedings Vol. 37: Proceedings of the 32nd International Conference on Machine Learning (ICML), pp. 2002-2011, 2015. The content of Chapter 4 was published in Journal of the Institute of Image Electronics Engineers of Japan (IIEEJ), Vol. 42, No. 2, pp. 214-221, 2013.
DOI: 10.14989/doctor.k19417
Appears in Collections:Doctoral Dissertation (Informatics)

Show full item record

Export to RefWorks

Export Format: 

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