Machine Learning and Knowledge Discovery in Databases European Conference Ecml PKDD 2022 Grenoble France September 1923 2022 Proceedings Part I Massihreza Amini Download
Machine Learning and Knowledge Discovery in Databases European Conference Ecml PKDD 2022 Grenoble France September 1923 2022 Proceedings Part I Massihreza Amini Download
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2022-grenoble-
france-september-1923-2022-proceedings-part-v-massihreza-
amini-49420186
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2010-barcelona-
spain-september-2024-2010-proceedings-part-i-1st-edition-christos-
faloutsos-auth-2022770
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2008-antwerp-
belgium-september-1519-2008-proceedings-part-i-1st-edition-franoise-
fogelmansouli-auth-2039778
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2010-barcelona-
spain-september-2024-2010-proceedings-part-iii-1st-edition-joni-
pajarinen-2537120
Machine Learning And Knowledge Discovery In Databases European
Conference Ecml Pkdd 2012 Bristol Uk September 2428 2012 Proceedings
Part Ii 1st Edition Ruilin Liu
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2012-bristol-uk-
september-2428-2012-proceedings-part-ii-1st-edition-ruilin-liu-4142516
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2009-bled-
slovenia-september-711-2009-proceedings-part-ii-1st-edition-sangkyun-
lee-4142518
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2009-antwerp-
belgium-september-711-2009-proceedings-buntine-w-4142520
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2011-athens-
greece-september-59-2011-proceedings-part-ii-1st-edition-satoshi-
hara-4142522
https://ebookbell.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2010-barcelona-
spain-september-2024-2010-proceedings-part-i-1st-edition-christos-
faloutsos-auth-4142528
Massih-Reza Amini · Stéphane Canu ·
Asja Fischer · Tias Guns · Petra Kralj Novak ·
Grigorios Tsoumakas (Eds.)
Knowledge Discovery
in Databases
European Conference, ECML PKDD 2022
Grenoble, France, September 19–23, 2022
Proceedings, Part I
123
Lecture Notes in Computer Science
Series Editors
Randy Goebel, University of Alberta, Edmonton, Canada
Wolfgang Wahlster, DFKI, Berlin, Germany
Zhi-Hua Zhou, Nanjing University, Nanjing, China
The series Lecture Notes in Artificial Intelligence (LNAI) was established in 1988 as a
topical subseries of LNCS devoted to artificial intelligence.
The series publishes state-of-the-art research results at a high level. As with the LNCS
mother series, the mission of the series is to serve the international R & D community
by providing an invaluable service, mainly focused on the publication of conference and
workshop proceedings and postproceedings.
Massih-Reza Amini · Stéphane Canu ·
Asja Fischer · Tias Guns · Petra Kralj Novak ·
Grigorios Tsoumakas
Editors
© The Editor(s) (if applicable) and The Author(s), under exclusive license
to Springer Nature Switzerland AG 2023
Chapters 5, 7 and 26 are licensed under the terms of the Creative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/). For further details see license information in the chapters.
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The European Conference on Machine Learning and Principles and Practice of Knowl-
edge Discovery in Databases (ECML–PKDD 2022) in Grenoble, France, was once again
a place for in-person gathering and the exchange of ideas after two years of completely
virtual conferences due to the SARS-CoV-2 pandemic. This year the conference was
hosted for the first time in hybrid format, and we are honored and delighted to offer you
these proceedings as a result.
The annual ECML–PKDD conference serves as a global venue for the most recent
research in all fields of machine learning and knowledge discovery in databases, includ-
ing cutting-edge applications. It builds on a highly successful run of ECML–PKDD
conferences which has made it the premier European machine learning and data mining
conference.
This year, the conference drew over 1080 participants (762 in-person and 318 online)
from 37 countries, including 23 European nations. This wealth of interest considerably
exceeded our expectations, and we were both excited and under pressure to plan a
special event. Overall, the conference attracted a lot of interest from industry thanks to
sponsorship, participation, and the conference’s industrial day.
The main conference program consisted of presentations of 242 accepted papers and
four keynote talks (in order of appearance):
– Francis Bach (Inria), Information Theory with Kernel Methods
– Danai Koutra (University of Michigan), Mining & Learning [Compact] Representa-
tions for Structured Data
– Fosca Gianotti (Scuola Normale Superiore di Pisa), Explainable Machine Learning
for Trustworthy AI
– Yann Le Cun (Facebook AI Research), From Machine Learning to Autonomous
Intelligence
In addition, there were respectively twenty three in-person and three online work-
shops; five in-person and three online tutorials; two combined in-person and one com-
bined online workshop-tutorials, together with a PhD Forum, a discovery challenge and
demonstrations.
Papers presented during the three main conference days were organized in 4 tracks,
within 54 sessions:
– Research Track: articles on research or methodology from all branches of machine
learning, data mining, and knowledge discovery;
– Applied Data Science Track: articles on cutting-edge uses of machine learning, data
mining, and knowledge discovery to resolve practical use cases and close the gap
between current theory and practice;
– Journal Track: articles that were published in special issues of the journals Machine
Learning and Data Mining and Knowledge Discovery;
vi Preface
– Demo Track: short articles that propose a novel system that advances the state of the
art and include a demonstration video.
We received a record number of 1238 abstract submissions, and for the Research
and Applied Data Science Tracks, 932 papers made it through the review process (the
remaining papers were withdrawn, with the bulk being desk rejected). We accepted 189
(27.3%) Research papers and 53 (22.2%) Applied Data science articles. 47 papers from
the Journal Track and 17 demo papers were also included in the program. We were able
to put together an extraordinarily rich and engaging program because of the high quality
submissions.
Research articles that were judged to be of exceptional quality and deserving of
special distinction were chosen by the awards committee:
– Machine Learning Best Paper Award: “Bounding the Family-Wise Error Rate in Local
Causal Discovery Using Rademacher Averages”, by Dario Simionato (University of
Padova) and Fabio Vandin (University of Padova)
– Data-Mining Best Paper Award: “Transforming PageRank into an Infinite-Depth
Graph Neural Network”, by Andreas Roth (TU Dortmund), and Thomas Liebig (TU
Dortmund)
– Test of Time Award for highest impact paper from ECML–PKDD 2012: “Fairness-
Aware Classifier with Prejudice Remover Regularizer”, by Toshihiro Kamishima
(National Institute of Advanced Industrial Science and Technology AIST), Shotaro
Akashi (National Institute of Advanced Industrial Science and Technology AIST),
Hideki Asoh (National Institute of Advanced Industrial Science and Technology
AIST), and Jun Sakuma (University of Tsukuba)
We sincerely thank the contributions of all participants, authors, PC members, area
chairs, session chairs, volunteers, and co-organizers who made ECML–PKDD 2022 a
huge success. We would especially like to thank Julie from the Grenoble World Trade
Center for all her help and Titouan from Insight-outside, who worked so hard to make
the online event possible. We also like to express our gratitude to Thierry for the design
of the conference logo representing the three mountain chains surrounding the Grenoble
city, as well as the sponsors and the ECML–PKDD Steering Committee.
General Chairs
Program Chairs
Local Chairs
Proceedings Chairs
Demonstration Chairs
Awards Chairs
Sponsorship Chairs
Web Chairs
Publicity Chair
Program Committees
Area Chairs
Sponsors
Contents – Part I
Anomaly Detection
Deep Learning Based Urban Anomaly Prediction from Spatiotemporal Data . . . 242
Bhumika and Debasis Das
Fast and Accurate Importance Weighting for Correcting Sample Bias . . . . . . . . . 659
Antoine de Mathelin, Francois Deheeger, Mathilde Mougeot,
and Nicolas Vayatis
1 Introduction
Truncated singular value decomposition (SVD) has broad applications in data
analysis and machine learning, such as dimension reduction, matrix comple-
tion, and information retrieval. However, for the large and high-dimensional
input data from social network analysis, natural language processing and rec-
ommender system, etc., computing truncated SVD often consumes tremendous
computational resource.
– We propose a technique to reduce the number of passes over the matrix in the
basic randomized SVD algorithm. It takes advantage of the row-major format
of the matrix and reads it row by row to build AΦ and AT AΦ with just one
pass over matrix. With this algorithm, the passes over the matrix in the basic
randomized SVD algorithm is reduced by half, with negligible loss of accuracy.
– Inspired by the shift technique in the power method [7], we propose to use
the shift skill in the power iteration called shifted power iteration to improve
the accuracy of results. A dynamic scheme of updating the shift value in
each power iteration is proposed to optimize the performance of the shifted
power iteration. This facilitates a pass-efficient randomized SVD algorithm,
i.e. PerSVD, which accurately computes truncated SVD of large matrix on a
limited-memory computer.
– Experiments on synthetic and real large data demonstrate that the proposed
techniques are all beneficial to improve the accuracy of result with same
number of passes over the matrix. With same 4 passes the over matrix, the
Pass-Efficient Randomized SVD with Boosted Accuracy 5
2 Preliminaries
Below we follow the Matlab conventions to specify indices of matrix and
functions.
A = UΣVT , (1)
where Uk and Vk are matrices with the first k columns of U and V respectively,
and the diagonal matrix Σk is the k × k upper-left submatrix of Σ. Notice that,
Ak is the best rank-k approximation of A in both spectral and Frobenius norm [5].
Fratelli d'Italia,
l'Italia s'è desta....
Fine.
OPERE DI LUCIO D'AMBRA (Renato Manganella)
Romanzi e novelle.
Teatro.
Critica.
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside
the United States, check the laws of your country in addition to
the terms of this agreement before downloading, copying,
displaying, performing, distributing or creating derivative works
based on this work or any other Project Gutenberg™ work. The
Foundation makes no representations concerning the copyright
status of any work in any country other than the United States.
1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if
you provide access to or distribute copies of a Project
Gutenberg™ work in a format other than “Plain Vanilla ASCII” or
other format used in the official version posted on the official
Project Gutenberg™ website (www.gutenberg.org), you must,
at no additional cost, fee or expense to the user, provide a copy,
a means of exporting a copy, or a means of obtaining a copy
upon request, of the work in its original “Plain Vanilla ASCII” or
other form. Any alternate format must include the full Project
Gutenberg™ License as specified in paragraph 1.E.1.
• You pay a royalty fee of 20% of the gross profits you derive
from the use of Project Gutenberg™ works calculated using
the method you already use to calculate your applicable
taxes. The fee is owed to the owner of the Project
Gutenberg™ trademark, but he has agreed to donate
royalties under this paragraph to the Project Gutenberg
Literary Archive Foundation. Royalty payments must be paid
within 60 days following each date on which you prepare (or
are legally required to prepare) your periodic tax returns.
Royalty payments should be clearly marked as such and sent
to the Project Gutenberg Literary Archive Foundation at the
address specified in Section 4, “Information about donations
to the Project Gutenberg Literary Archive Foundation.”
• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.
1.F.
Most people start at our website which has the main PG search
facility: www.gutenberg.org.
ebookbell.com