[ABE-L] Seminário PGEST (UFPE) - Feature selection in biclassifying SVM by means of strict sparsity control

Seg Mar 17 11:33:11 -03 2025

We would like to extend this invitation to our academic community for the
next seminar hosted by our graduate program, PGEST/UFPE. The seminar
"Feature selection in biclassifying SVM by means of strict sparsity
control" will be held next Wednesday at 3 pm (GMT-3) by Professor Immanuel
M. Bomze from the Department of Statistics and Operations Research at the
University of Vienna. This is a joint work with F.d'Onofrio, L.Palagi and
B.Peng.

*Abstract*: In some ML communities, researchers claim that obtaining local
solutions of optimality criteria is often sufficient to provide a
meaningful and accurate data model in real-world analytics. However, this
is simply incorrect and sometimes dangerously misleading, particularly when
it comes to highly structured problems involving non-convexity such as
discrete decisions (binary variables). This talk will advocate the
necessity of research efforts in the quest for global solutions and strong
rigorous bounds for quality guarantees, showcased on one of the nowadays
most popular domains -- cardinality-constrained models. These models try to
achieve fairness, transparency and explainability in AI applications,
ranging from Math.Finance/Economics to social and life sciences.

Instead of using surrogates for the benefit of tractability, we propose to
incorporate the true zero-norm into the base model and treat this either by
MILP relaxations or else by lifting to tractable conic optimization models.
Both in practice and in theory, these have proved to achieve much stronger
bounds than the usual LP-based ones, and therefore they may, more reliably
and based upon exact arguments, assess the quality of proposals coming from
other techniques in a more precise way. With some effort invested in the
theory, the resulting models are still scalable and would guarantee
computational performance closer to reality and/or optimality.

*About Prof. Bomze*: Since 2004 until retirement 2023, Prof. Dr. Immanuel
M. Bomze held a chair (full professor) of Applied Mathematics and
Statistics at the University of Vienna and is now active as a Professorial
Research Fellow at various institutions of this university, as well as a
Distinguished Research Fellow at the Corvinus Institute of Advanced Studies
in Budapest.

Bomze’s research interests are in the areas of nonlinear optimization,
qualitative theory of dynamical systems, game theory, mathematical
modelling and statistics, where he has edited one and published four books,
as well as over 130 peer-reviewed articles in scientific journals and
monographs. The list of his co-authors comprises over ninety scientists
from more than a dozen countries in four continents. In 2014 he was elected
Fellow of EurOpt, the Continuous Optimization Working Group of EURO, the
Association of European Operational Research Societies
https://www.euro-online.org/websites/continuous-optimization/.

As a member of program and/or organizing committees, he co-organized
various scientific events and he is an Associate Editor for five
international journals. For several science foundations and councils (based
in Canada, the Czech Republic, Germany, Great Britain, Hong Kong, Israel,
Italy, the Netherlands, Norway, Portugal, Singapore, Spain, USA), and for
almost 50 scientific journals he acted as a reporting referee. 2011–2017 he
served as an Editor (Co-EiC) of the European Journal of Operational
Research, and since 2021 he serves as sole EiC of the EURO Journal of
Computational Optimization. As elected president of EURO, Bomze served
2018-2020 and, as Past President, in the EURO Executive Committee until end
of 2021.

[image: Seminário Immanuel Bomze (2025.1).png]

-- 
Thyago Nepomuceno

Assistant Professor - Departament of Statistics - CCEN - Universidade
Federal de Pernambuco
Deputy Coordinator of Masters and PhD. Programs in Statistics -
Universidade Federal de Pernambuco
Bolsista de Produtividade em Pesquisa PQ2 - Conselho Nacional de Des.
Científico e Tecnológico (CNPq)
Research Fellow - Department of Computer, Control, and Management
Engineering - Sapienza University of Rome
Editor in chief of Socioeconomic Analytics
<https://periodicos.ufpe.br/revistas/index.php/SECAN/index>
Contatos: thyago.nepomuceno em ufpe.br; nepomuceno em diag.uniroma1.it
...........................................
Recent publications: MCDA Distance Functions (DEA)
<https://doi.org/10.1016/j.omega.2024.103044>, Panel Data Envelopment
Analysis (DEA) <https://doi.org/10.1108/IJSE-11-2022-0767>, Maritime
Transportation Efficiency (DEA)
<http://dx.doi.org/10.1080/03088839.2023.2252431>, Knowledge Benchmarking
(DEA) <https://onlinelibrary.wiley.com/doi/abs/10.1111/exsy.12967>, Webpage
Performance (DEA) <http://dx.doi.org/10.3390/data8080134>, Systematic
Reviews as a Metaknowledge (Bibliometrics)
<https://onlinelibrary.wiley.com/doi/10.1111/itor.13309>, Information
Security (Bibliometrics) <http://dx.doi.org/10.3390/fi15120393>, State of
the Art on DEA (Bibliometrics)
<https://doi.org/10.1504/IJOR.2023.129960>, Review
of Hospitals' Efficiency (Bibliometrics)
<https://www.mdpi.com/2227-9032/10/7/1316>, Innovation Hubs (MCDA)
<http://dx.doi.org/10.3390/info14030169>, Relational Governance (ITO)
<http://dx.doi.org/10.1108/ARLA-10-2022-0182>, Bedouin Syndrome
(Econometrics) <https://www.mdpi.com/1660-4601/19/15/9711>, COVID-19
Mobility Impact (Econometrics)
<http://www.aimspress.com/article/doi/10.3934/mbe.2022332>.
...........................................
<http://lattes.cnpq.br/4815452948320665>
<https://orcid.org/0000-0001-8327-6472>
<https://www.researchgate.net/profile/Thyago-Nepomuceno-2>
<https://www.linkedin.com/in/thyago-nepomuceno-78a08a42/>
<https://scholar.google.com.br/citations?user=xvNP0ioAAAAJ&hl=pt-BR&oi=ao>
..........................................
<https://www.ufpe.br/caa> <https://www.ufpe.br/ccen>
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20250317/5ba1a74e/attachment-0001.htm>
-------------- Próxima Parte ----------
Um anexo não-texto foi limpo...
Nome: Seminário Immanuel Bomze (2025.1).png
Tipo: image/png
Tamanho: 2833039 bytes
Descrição: não disponível
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20250317/5ba1a74e/attachment-0001.png>