EVENTS

Affiliation: Centro de Investigación en Ciencias, Universidad Autónoma del Estado de Morelos (Cuernavaca, Morelos, México)

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Centro de Investigación en Ciencias, Universidad Autónoma del Estado de Morelos (Cuernavaca, Morelos, México)

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Cambridge University - UK 


Venue: office 36, 3rd floor of the Mathematics Department Building

Venue: Campus Universitário de Santiago of the University of Aveiro, Portugal

Affiliation: College of William and Mary, U.S.A.


Venue: office 27, 3rd floor of the Mathematics Department Building

Departamento de Matemática da Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa (FCT NOVA)

Affiliation: Faculty of Mathematics and Natural Science, South-West University "Neot Rilski", Bulgaria


Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Institute of Mathematics, Potsdam University, Germany


Venue: office 2, 3rd floor of the Mathematics Department Building

Departamento de Matemática da Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa (FCT NOVA)

The 11th European Conference on Mathematical and Theoretical Biology (ECMTB 2018) will be held in Lisbon, Portugal, from 23 to 27 July, 2018.

Departamento de Matemática da Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa (FCT NOVA)

Affiliation: University of Granada, Spain

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Department of Mathematical Sciences, Durham University, Durham

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Department of Philosophy, Université Paris 1 Panthéon-Sorbonne

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: University of Copenhagen

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: University of Tunis El Manar

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Universidade Federal Fluminense

Venue: office 5, 3rd floor of the Mathematics Department Building

Affiliation: Komi Science Centre Ural Division, Russian Academy of Sciences, Department of Mathematics, Syktyvkar, Russian Federation and Syktyvkar State University

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: University of St Andrews, UK

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Institute of Mathematics, Potsdam University, Germany

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Centro de Investigación en Ciencias, Universidad Autónoma del Estado de Morelos (Cuernavaca, Morelos, México)

Venue: office 36, 3rd floor of the Mathematics Department Building

Affiliation:

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Faculty of Mathematics and Natural Science, South-West University "Neot Rilski", Bulgaria

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Università Degli Studi Di Salerno

Venue: office 35, 3rd floor of the Mathematics Department Building

Affiliation: King's College London

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Centro de Investigación en Ciencias, Universidad Autónoma del Estado de Morelos (Cuernavaca, Morelos, México)

Venue: Office 2, 3rd floor of the Mathematics Department Building

Departamento de Matemática da Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa (FCT NOVA)

Sala 1.A - Edifício VII

Affiliation: Departamento de Matemática, Federal University of Minas Gerais (UFMG), Brazil

Venue: Office 5, 3rd floor of the Mathematics Department Building

Affiliation: College of William & Mary, USA

Venue: office 27, 3rd floor of the Mathematics Department Building

The purpose of this first DCV - DM Meeting was to present the research of both groups seeking to create new synergies in the area of biomathematics. The discussion was very dynamic with 25 researchers present. New meetings are already scheduled, in the form of more specific seminars. CMA as several researchers working on Mathematical Biology  

Affiliation: Faculty of Mathematics and Natural Science, South-West University "Neofit Rilski", Bulgaria

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Institute of Mathematics, University of Potsdam, Germany

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Department of Statistics and Operations Research, Institute of Mathematics,
University of Sevilha, Sevilha, Spain

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: School of Mathematics and Physical Sciences, University of Hull, United Kingdom

Venue: office 6, 3rd floor of the Mathematics Department Building

Affiliation: School of Mathematics and Statistics, University of St Andrews, United Kingdom

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: UNIMIB - Department of Mathematics and its Applications,
Università degli Studi di Milano-Bicocca, Milan, Italy

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Division of Mathematical Sciences, Nanyang Technological University, Singapore

Venue: office 2, 3rd floor of the Mathematics Department Building

Department of Mathematics of NOVA School of Science and Technology (FCT NOVA)

The conference will be held from Monday July 22nd through Friday July 26th, at Instituto Superior Técnico, University of Lisbon, Portugal.

1st WM2 - Women in Mathematics Meeting The Women in Mathematics Meeting will take place at Faculdade de Ciências e Tecnologia of the Universidade Nova de Lisboa, Portugal, from 22 to 24 July.

Affiliation: Institute of Mathematics, Potsdam University, Germany

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Laboratoire G-SCOP, France

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Federal University of Minas Gerais, Brazil

Venue: office 5, 3rd floor of the Mathematics Department Building

Affiliation: Faculty of Mathematics and Natural Science, South-West University "Neofit Rilski", Bulgaria

Venue: office 18, 3rd floor of the Mathematics Department Building

Department of Mathematics of NOVA School of Science and Technology (FCT NOVA)

Affiliation: Departamento de Biologia e CESAM, University of Aveiro

Venue: office 18, 3rd floor of the Mathematics Department Building

Workshop*
Sessões nos dias 9/10 e 10/10, das 14h às 15h30, no laboratório 2.2 do edifício 7

Modelação Estatística em Biologia: Como lidar com a não-normalidade de dados contínuos, com dados discretos ou dados qualitativos?

Prof.ª Regina Bispo, Departamento de Matemática & Centro de Matemática e Aplicações, FCT Nova

Resumo: Neste workshop pretende-se abordar alguns aspetos iniciais da Modelação Estatística, em duas situações frequentemente encontradas na Biologia: (1) quando a variável resposta, cuja variação se pretende explicar, apesar de contínua, não tem distribuição Normal e (2) quando a variável resposta não é contínua, podendo nesse caso ser quer uma quantidade discreta (contagem) que uma qualidade binária. Serão apresentados alguns casos reais que servirão de ponto de partida para uma abordagem prática, utilizando o software R.

* O workshop destina-se a investigadores e alunos de final do 1º ciclo ou 2º ciclo das áreas de Matemática e Biologia

[Entrada livre mediante inscrição prévia obrigatória AQUI ]

Workshop*
Sessões nos dias 9/10 e 10/10, das 14h às 15h30, no laboratório 2.2 do edifício 7

Modelação Estatística em Biologia: Como lidar com a não-normalidade de dados contínuos, com dados discretos ou dados qualitativos?

Prof.ª Regina Bispo, Departamento de Matemática & Centro de Matemática e Aplicações, FCT Nova

Resumo: Neste workshop pretende-se abordar alguns aspetos iniciais da Modelação Estatística, em duas situações frequentemente encontradas na Biologia: (1) quando a variável resposta, cuja variação se pretende explicar, apesar de contínua, não tem distribuição Normal e (2) quando a variável resposta não é contínua, podendo nesse caso ser quer uma quantidade discreta (contagem) que uma qualidade binária. Serão apresentados alguns casos reais que servirão de ponto de partida para uma abordagem prática, utilizando o software R.

* O workshop destina-se a investigadores e alunos de final do 1º ciclo ou 2º ciclo das áreas de Matemática e Biologia

[Entrada livre mediante inscrição prévia obrigatória AQUI ]

Mathematical Biology Day @ FCT NOVA
10/10/2019 Palestra
Dia 10/10, das 15h30 às 16h30, na sala de seminários do edifício 7

Título: Mathematical modeling: a fundamental tool in biology
Prof. Ruy Ribeiro, Laboratório de Biomatemática, Faculdade de Medicina da Universidade de Lisboa and Theoretical Biology and Biophysics, Los Alamos National Laboratory

[Entrada livre mediante inscrição prévia obrigatória AQUI ]
(https://eventos.fct.unl.pt/mathematicalbiologydayatfctnova)

Affiliation: Department of Mathematics, Denver University, USA

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Alfréd Rényi Institute of Mathematics & Eötvös University, Budapest, Hungary

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Fluminense Federal University, Brazil

Venue: office 5, 3rd floor of the Mathematics Department Building

Affiliation: Departamento de Matemática Aplicada, State University of Campinas, São Paulo, Brazil

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Department of Mathematical and Statistical Methods, Poznań University of Life Sciences, Poland

Venue: CMA room, 2nd floor of the Mathematics Department Building

Affiliation: Moscow State University, Russia 

Venue: Office 18, 2nd floor of the Mathematics Department Building

Affiliation: University of Zielona Góra, Poland

Venue: CMA room, 2nd floor of the Mathematics Department Building

Affiliation: University of St Andrews, United Kingdom

Venue: Office 17, 3rd floor of the Mathematics Department Building

Faculdade de Ciências e Tecnologia – Universidade Nova de Lisboa (FCT NOVA), Seminar Room – Departamento de Matemática – Building VII

Affiliation: King's College London, London, United Kingdom

Venue: Office 2, 3rd floor of the Mathematics Department Building

Dates: 15/12/2019 to 21/12/2019

Departamento de Matemática da Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa (FCT NOVA)

Affiliation: School of Mathematics, University of Manchester, United Kingdom

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Faculty of Mathematics and Natural Science, South-West University "Neofit Rilski", Bulgaria

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Institute of Mathematics, Potsdam University, Germany

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Department of Mathematics, Faculty of Science, Khon Kaen University, Thailand

Venue: office 2, 3rd floor of the Mathematics Department Building

Affiliation: Laboratoire LMNO, Université de Caen Normandie, France

Venue: office 18, 3rd floor of the Mathematics Department Building

Affiliation: Instituto de Matemática e Estatística da Universidade de São Paulo

Venue: office 35, 3rd floor of the Mathematics Department Building

Affiliation: School of Mathematics and Statistics, University of St Andrews

Venue: office 2, 3rd floor of the Mathematics Department Building

09:45 - Opening

António Malheiro (CMA)

Luís Caires (NOVA LINCS)

Marcos Raydan (CMA)

Pedro Barahona (NOVA LINCS)

10:00 - 10:45

Statistics and Machine Learning in Biomedical and Environmental Sciences, Marta Lopes (CMA / NOVA LINCS)

10:45 - 11:15

How to test different block diagonal structures in one or several covariance matrices, Filipe Marques (CMA)  

11:15 - 11:30

SI-MORENA’s Data Driven Projects, Carlos Damásio and João Moura Pires (NOVA LINCS)

11:30 - 12:00

Classification with symbolic data, Paula Amaral (CMA)

12:00 - 12:30

Pot-pourri of Systems for Big Data and Data Science, Nuno Preguiça (NOVA LINCS)

12:30 - 14:00

Lunch

14:00 - 14:30

Semantic Similarity and Textual Entailment in the Portuguese language, Rui Rodrigues (CMA)

14:30 - 15:00

Title TBA, João Magalhães (NOVA LINCS)

15:00 - 15:30

Feature selection for marine species origin prediction, Regina Bispo (CMA)

15:30 - 16:00

Deep Models in Data Science: some problems and applications, Ludwig Krippahl (NOVA LINCS)

16:00 - 16:30

General discussion and closing

Addicional information here.

Addicional information here.

Título: Monitorização da transmissibilidade da COVID-19 em Portugal

Orador: Liliana Antunes, INSA

Título: Estimação de parâmetros epidemiológicos da COVID-19: internamentos e óbitos

Orador: João Pereira, INSA, UTAD (bolseiro do Projeto COVID-19 in-CTRL)

Título: Modelação matemática da dinâmica de transmissão de COVID-19 em Portugal

Orador: Constantino Caetano, INSA

More information is available here.

Addicional information here.

Website

Website

More information on the webpage

More information available here

More information available here

Venue: office 2, 3rd floor of the Mathematics Department Building

More information on the webpage

Venue: office 2, 3rd floor of the Mathematics Department Building

More information here

More information here

Venue: office 2, 3rd floor of the Mathematics Department Building

Venue: office 2, 3rd floor of the Mathematics Department Building

More information is available here

More information is available here

Venue: office 2, 3rd floor of the Mathematics Department Building

More information is available here

Abstract: In this expository talk, which will be accessible to everyone, I will introduce an alternative to real analysis called p-adic analysis and describe some of its main (dis)advantages.

More information is available here

More information is available here

Abstract: Compressed sensing is a new paradigm in signal processing which states that for certain matrices sparse representations can be obtained by a simple $\ell_1$−minimization. In this seminar we explore this paradigm for higher-dimensional signals. In particular three cases are being studied: signals taking values in a bicomplex algebra, quaternionic signals, and complex signals which are representable by a nonlinear Fourier basis, a so-called Takenaka-Malmquist system.

We formulate the optimal balance sheet management problem as a linear program and study it using a duality approach. In addition to helping determine the optimal balance sheet, the dual problem also provides us the interest rate risk and credit risk pricing. We deploy our methodology to determine premia on credit risk and interest rate risk for commercial banks, which allows us to manage the risk allocation for a bank given a risk budget. Moreover, our approach will be of interest to regulators, who can use it to assess the price of credit and interest rate risk at each point in the economic cycle. Finally, we apply this methodology to real data and show how it can be used in a real-world setting, using diversification constraints and a greedy algorithm that results in the optimal asset-liability allocation.

This research is independent from the activity at Montepio Bank and does not reflect the views of this institution.

Pedro Júdice is a director at Montepio Bank in Lisbon and an integrated researcher at ISCTE Business Research Unit. He started his banking career in 2003 as a quantitative research associate at JPMorgan Chase in London, and in 2006 he joined Montepio Bank as a director. Since then, he had several different roles at this bank, in market risk, non-core asset sales, strategic planning, asset-liability management, and financial sector research. Pedro has a Ph.D. from the Courant Institute of Mathematical Sciences (New York University, 2003) and was a postdoctoral research associate at the University of Chicago Booth School of Business (2011-2012). He was an invited assistant professor at ISCTE Business School in Lisbon (2010-2012), a Visiting Professor at Western Michigan University (Feb 2022-Apr 2022), and served as a regional co-director for the Global Association of Risk Professionals (2008-2012). His academic research has focused on practical banking problems, and he has published in the Journal of Banking and Finance and Expert Systems with Applications, among other journals.

OR Seminar: June 01 (Wednesday)

Title: Robust Optimization in Radiotherapy Treatment Planning: models and algorithms

Time:  3:00 pm to 4:00 pm,  in room (soon we will disclose) Building VII.   

Author:  Joana Matos Dias has a BSc in Computers Engineering (University of Coimbra, 1996), a MSc in Operations Research (University of Lisbon, 2000), a MSc in Quantitative Finance (University of London, 2011), a PhD and Habilitation in Management Science (University of Coimbra, 2006, 2017). She is Associate Professor at the Faculty of Economics of the University of Coimbra, where she has been responsible for several curricular units like informatics, logistics, operations research, modeling in management, simulation. She is researcher at Inesc-Coimbra. Her main research interest is decision making models and algorithms in general, and operations research applied to health problems, combinatorial optimization, multiobjective optimization, in particular.

Personal homepage: https://www.uc.pt/feuc/joanamatosdias
Scopus ID: 36238953500
ORCID: orcid.org/0000-0003-2517-7905

-------------------------------------

Abstract: Radiotherapy is one of the possible treatments for cancer, and it is indeed delivered to more than half of all cancer patients. Proton therapy has specific features that make it highly appropriate for the treatment of cancer as compared with photon radiotherapy. Intensity Modulated Proton Therapy (IMPT), the most advanced form of proton therapy, is delivered using electromagnetically scanned narrow beams (“beamlets”) of protons. The process of treatment planning consists of deciding the number and directions of beams and the optimization of intensities and energies of beamlets comprising each of the beams so that a homogeneous dose distribution is obtained in the target volume and the normal tissues are maximally spared. Proton beams have unique depth-dose characteristics, since protons slow down as they penetrate matter, their rate of energy transfer increases with depth, coming to an abrupt stop just beyond where energy deposition is maximum producing the so-called Bragg peak. By positioning several Bragg peaks inside the target volume, excellent tumor conformity is obtained, and adjacent organs at risk are spared. However, proton therapy is much more vulnerable to uncertainties due to this very steep Bragg peak. These uncertainties should be explicitly taken into account when planning the treatment, otherwise it is not possible to guarantee that the treatment will achieve its objectives, guaranteeing proper tumor coverage and sparing of the organs at risk. And this is where Operational Research can make the difference!

In this presentation the crucial role of Operational Research in radiotherapy treatment planning, contributing to cancer treatment, will be highlighted, and developments considering the application of robust optimization to deal with the underlying uncertainties will be described. Fuzzy inference systems, quadratic programming, pattern search methods and Monte Carlo Simulation all come together, contributing to the calculation of high quality plans in an almost automated way.

Author:   Miguel Alves Pereira is a post-doctoral researcher at the Institute for Systems and Computer Engineering, Technology and Science (INESC TEC). He earned his doctoral degree in Engineering and Management at the Instituto Superior Técnico of the Universidade de Lisboa in 2021, where he also earned his integrated master’s degree in Biomedical Engineering in 2018. His main interests concern operational research and management science (with an emphasis on the development of efficiency measurement and performance assessment models), and their applications to sustainable development and the health and water and sanitation sectors. He is the author of 13 scientific publications in several top international peer-reviewed journals, as well as a referee in 20 of them.

-------------------------------------

Abstract: 

Convergence in productivity examines if entities in an industry get closer to the best practices or if the gap between the frontiers of the best and worst performers decreases over time. In a multi-input multi-output setting, the assessment of sigma- and beta-convergence can be measured with the use of non-parametric frontier techniques, such as data envelopment analysis. We propose an innovative approach to estimate convergence in the context of performance assessments resting on composite indicators, accounting for desirable and undesirable indicators. This methodology rests on ‘Benefit-of-the-Doubt’ models, specified with a directional distance function. It is applied to the Member States of the World Health Organization (WHO) in order to study their convergence in terms of the United Nations’ Sustainable Development Goal (SDG) ‘Good health and well-being’. We collected data for all years since the proposal of the SDGs, covering the period between 2016 and 2020. The results show that all WHO regions are beta-divergent, especially because of the generalised decline of the Worst Practice Frontier (WPF), alongside an improvement at a lower rate of the Best Practice Frontier (BPF). The regional analysis also revealed sigma-convergence in the Region of the Americas and the Eastern Mediterranean Region; the South-East Asia and African Regions exhibited sigma-divergence; the Western Pacific and European Regions remained stable in terms of the performance spread regarding the BPF. At the worldwide level, we also observed an increase of the gap between the BPF and the WPF, although the performance spread around the worldwide BPF remained relatively stable.

Venue: office 2, 3rd floor of the Mathematics Department Building

OR Seminar: June 28 (Tuesday)

Title: Designing Master Surgery Schedules with Downstream Unit Integration via Stochastic Programming

Time: 2:30 pm - 3:30 pm, Location:  Building VII,  Seminar room (2nd floor) . 

Author:  Daniel Rebelo dos Santos (Professor Auxiliar - Departamento de Engenharia e Gestão- Instituto Superior Técnico- Universidade de Lisboa)

------------------------------------

Abstract: Surgical activity has a substantial impact in all areas of hospitals. Additionally, social concerns arise related to equity and speed of access. Therefore, operating room management is paramount in modern society. This work studies the master surgery scheduling problem which is the problem of assigning surgical specialties to operating room blocks, which represent a shift of an operating room. For a master surgery schedule to be applicable in practice, multiple considerations must be taken into account. The particular focus of this work is in the integration of downstream units (i.e., wards or the ICU). Although, in a tactical planning scenario, operational bed requirements are unknown, these may be estimated based on historical data. We propose a novel stochastic programming model that captures the uncertainty in the bed requirements, with a recourse function that penalizes the overutilization of beds. A solution approach based on Benders decomposition is developed and results for generated instances mimicking real-life data are presented.

--

This seminar is funded by national funds through the FCT - Fundação para a Ciência e a Tecnologia, I.P., under the scope of the projects UIDB/00297/2020 and UIDP/00297/2020 (NovaMath - Center for Mathematics and Applications).
 
------------------------------
NOVA Math - Center for Mathematics and Applications
NOVA School of Science and Technology | FCT NOVA

More information is available here

Neovascularization is the hallmark of neovascular age-related macular degeneration (AMD). Type 1 is a neovascular subtype that originates from the ingrowth of vessels from the choriocapillaris (CC) into the sub-RPE (retinal pigment epithelium) space (which is a region above CC). Some evidence suggested that type 1 macular neovascularization (MNV) might be able to provide nutritional support to the RPE and outer retina, by assuming the function of diseased choriocapillaris, which are areas with impaired blood flow (flow voids). 

Gray-scale images from CC were obtained through OCTA (Optical Coherence Tomography Angiography) and an image processing algorithm was applied to obtain the segmentation of the CC flow voids. Besides, the neovascularization segmentation outlines at the baseline and at the follow-up scans were also provided. 

Our main goal was to evaluate the correlation between CC texture features, flow voids (FV) size and density with MNV radial growth in eyes with AMD over time. Thirty case series were analyzed in our study. Image and statistical analysis were performed with a code designed in MATLAB version R2020a. Our algorithm selected a region extending 0 to 600 μm from the margin of the MNV at the baseline image. Though the division of this ring-shaped region around MNV in radial sectors (10/15/20/30º central angle sectors), we obtained per sector: five Haralick features (contrast, correlation, energy, entropy, and homogeneity), FV average size and FV density and average neovascular growth. The algorithm used conversion to polar coordinates (using as reference point: the centroid of the MNV at baseline scan) to facilitate the analyses of the sectors (ring-shaped to rectangular-shaped) in order to obtain these measurements. Kendall’s rank correlation coefficient (τ) was calculated to correlate each of these metrics with respective neovascular growth for each case series. A final global τ average was also calculated comprising all case series for all CC features studied. In our analysis, we have found weak/moderate positive correlations between some of the CC metrics and the neovascular growth. 

Additionally, all values of neovascular growth and the corresponding values of CC features of all case series were compiled together. What is the best way to analyze this global data set is one of our questions. Ultimately, our goal is (if possible) to develop a model to predict the most probable areas of MNV growth considering all (or some) of the variables studied.

(Joint work with Professor Miguel Seabra, Doctor Sandra Tenreiro and Doctor Telmo Pereira)

Venue: office 18, 3rd floor of the Mathematics Department Building

Venue: office 2, 3rd floor of the Mathematics Department Building

Venue: office 2, 3rd floor of the Mathematics Department Building

More information is available here

Venue: office 2, 3rd floor of the Mathematics Department Building

We consider discrimination between mammary cancer and mastopathy tissues, which plays a crucial role in clinical practice. Based on my recent projects, I will discuss results on how stochastic processes can model both tissue modalities, mammary cancer and mastopathies. I will use exponential families with geometric-topological canonical parameters, like Euler characteristic. Application of such discrimination techniques has been applied to real biopsy images from patients in German clinics with reasonable improvement of false positive rates. Such an approach can be considered as stochastic and statistical modelling for image processing which can provide good benchmarks for false-positive rates in most typical cancer assessments.

Pure  machine learning techniques can benefit from these obtained results and thus contribute to addressing "inherent data ambiguity" in medical imaging. The inter-patient variability of fractal dimension for mammary cancer is high and therefore multi-fractality can serve as  a better concept. This is a feasible and parsimonious solution for a more delicate problem of multi-objective aggregation of information, relevant in modeling clinical assessor intervariability. Such a problem is well visible if one considers several fractal and texture characteristics of biopsy images. This helps to refine a fractal cancer hypothesis, well applicable for several standard cancers, e.g. mammary or prostate cancer. In this talk I will also introduce a very different type of cancer growth which is not related to fractal cancer hypothesis. Nephroblastoma (given by Wilms tumor) is the typical tumor of the kidneys appearing in childhood, which does not satisfy fractal cancer hypothesis. We will  illustrate by recent pre/post clinical study the effect on invasive treatments to Euclidean volumes of such tumors.

Abstract: We present here different situations in which the filtering of high or low modes is used either for stabilizing semi-implicit numerical schemes when solving nonlinear parabolic equations, or for building adapted damping operators in the case of dispersive equation. We consider numerical filtering provided by mutigrid-like techniques as well as the filtering resulting from operator with monotone symbols. Our approach applies to several discretization techniques and we focus on finite elements and finite differences. Numerical illustrations are given on Cahn-Hilliard, Korteweig-de Vries and Kuramoto-Sivashinsky equations.

More information is available here

More information is available here

More information is available here

More information is available here

Abstract: This talk will discuss some applications of Bayesian matrix/tensor factorization and mixed effects models on biological and health data. Bayesian matrix factorization is applied to the expanded disability status scale (EDSS) data of multiple sclerosis patients. Mixed effects models are applied to model the effect of pharmaceutical pollution on amphibians’ tadpoles development. The models are fitted in Python for Bayesian matrix factorization and SAS for mixed models. Several models are tested under different situations to discover the best modelling alternatives while giving details on analysis strategy and software implementation. Since the datasets used were relatively small, it was possible to explore different choices for the models within reasonable computation time and study how they impact the predictive performances. The results show that these methods can handle missing data properly and arrive at accurate predictions, especially when working with sparse and noisy real-world datasets.

Venue: office 18, 3rd floor of the Mathematics Department Building

Venue: office 2, 3rd floor of the Mathematics Department Building

Abstract: "In this paper we consider constrained optimization problems where both the objective and constraint functions are of the black-box type. Furthermore, we assume that the nonlinear inequality constraints are non-relaxable, i.e. their values and that of the objective function cannot be computed outside of the feasible region. This situation happens frequently in practice especially in the black-box setting where function values are typically computed by means of complex simulation programs which may fail to execute if the considered point is outside of the feasible region. For such problems, we propose a new derivative-free optimization method which is based on the use of a merit function that handles inequality constraints by means of a log-barrier approach and equality constraints by means of a quadratic penalty approach. We prove convergence of the proposed method to KKT stationary points of the problem under quite mild assumptions. Furthermore, we also carry out a preliminary numerical experience on standard test problems and comparison with state-of-the-art solvers which shows efficiency of the proposed method."

Abstract: "In this paper we consider constrained optimization problems where both the objective and constraint functions are of the black-box type. Furthermore, we assume that the nonlinear inequality constraints are non-relaxable, i.e. their values and that of the objective function cannot be computed outside of the feasible region. This situation happens frequently in practice especially in the black-box setting where function values are typically computed by means of complex simulation programs which may fail to execute if the considered point is outside of the feasible region. For such problems, we propose a new derivative-free optimization method which is based on the use of a merit function that handles inequality constraints by means of a log-barrier approach and equality constraints by means of a quadratic penalty approach. We prove convergence of the proposed method to KKT stationary points of the problem under quite mild assumptions. Furthermore, we also carry out a preliminary numerical experience on standard test problems and comparison with state-of-the-art solvers which shows efficiency of the proposed method."