Theoretical foundations of multi-task and lifelong learning (Record no. 358238)

000 -LEADER
fixed length control field 02140ntm a22002897a 4500
003 - CONTROL NUMBER IDENTIFIER
control field AT-ISTA
005 - DATE AND TIME OF LATEST TRANSACTION
control field 20190813141841.0
008 - FIXED-LENGTH DATA ELEMENTS--GENERAL INFORMATION
fixed length control field 170609s2016 au ||||| m||| 00| 0 eng d
040 ## - CATALOGING SOURCE
Transcribing agency IST
100 ## - MAIN ENTRY--PERSONAL NAME
Personal name Pentina, Anastasia
9 (RLIN) 3355
245 ## - TITLE STATEMENT
Title Theoretical foundations of multi-task and lifelong learning
260 ## - PUBLICATION, DISTRIBUTION, ETC. (IMPRINT)
Name of publisher, distributor, etc. IST Austria
Date of publication, distribution, etc. 2016
500 ## - GENERAL NOTE
General note Thesis
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note Acknowledgments
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note Abstract
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note 1 Introduction
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note 2 Background
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note 3 Multi-task learning
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note 4 Lifelong learning
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note 5 Future directions
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note A Proofs of theorems in chapter 3
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note B Proofs of theorems in chapter 4
505 ## - FORMATTED CONTENTS NOTE
Formatted contents note C Supplementary lemmas
520 ## - SUMMARY, ETC.
Summary, etc. Traditionally machine learning has been focusing on the problem of solving a single task in isolation. While being quite well understood, this approach disregards an important aspect of human learning: when facing a new problem, humans are able to exploit knowledge acquired from previously learned tasks. Intuitively, access to several problems simultaneously or sequentially could also be advantageous for a machine learning system, especially if these tasks are closely related. Indeed, results of many empirical studies have provided justification for this intuition. However, theoretical justifications of this idea are rather limited. The focus of this thesis is to expand the understanding of potential benefits of information transfer between several related learning problems. We provide theoretical analysis for three scenarios of multi-task learning - multiple kernel learning, sequential learning and active task selection. We also provide a PAC-Bayesian perspective on lifelong learning and investigate how the task generation process influences the generalization guarantees in this scenario. In addition, we show how some of the obtained theoretical results can be used to derive principled multi-task and lifelong learning algorithms and illustrate their performance on various synthetic and real-world datasets.
856 ## - ELECTRONIC LOCATION AND ACCESS
Uniform Resource Identifier <a href="https://doi.org/10.15479/AT:ISTA:TH_776">https://doi.org/10.15479/AT:ISTA:TH_776</a>
942 ## - ADDED ENTRY ELEMENTS (KOHA)
Source of classification or shelving scheme
Holdings
Withdrawn status Lost status Source of classification or shelving scheme Damaged status Not for loan Permanent Location Current Location Date acquired Barcode Date last seen Price effective from Koha item type
  Not Lost       Library Library 2017-06-09 AT-ISTA#001354 2018-11-06 2017-06-09 Book

Powered by Koha