Seminari
Prossimi Seminari
- A mathematical-physics approach to machine learning
Pierluigi Contucci, Dipartimento di Matematica Università di Bologna
giovedì 30 gennaio 2020 alle ore 14:00, Aula Saleri VI piano - Modeling and simulation of thermo-poroelastic processes in fractured geothermal reservoirs
Eirik Keilegavlen, Department of Mathematics, University of Bergen, Norway
giovedì 20 febbraio 2020 alle ore 11:30, Aula Saleri - VI piano
Seminari Passati
- Going deep into shallowness
Alessandro Verri, MaLGa - Università di Genova
giovedì 23 gennaio 2020 alle ore 14:00, Aula Consiglio VII pianoABSTRACTIn the first part of my talk I quickly review the work we have been doing at UniGe in the last decades on machine learning. Ranging from theoretical to applied work I’ll highlight strengths and weaknesses of the regularization approach to learning. In the second part I argue, through examples and some surprisingly basic considerations, that while in certain application domains the widespread enthusiasm for deep learning is well justified, in others a more careful and critical approach might be the key to build truly intelligent systems.
Contatto: luca.formaggia@polimi.it
- Nonlinear Peridynamic Models
Giuseppe Maria Coclite, Politecnico di Bari
mercoledì 22 gennaio 2020 alle ore 15:15, Sala Consiglio 7° pianoABSTRACTSome materials may naturally form discontinuities such as cracks as a result of scale effects and long range interactions. Peridynamic models such behavior introducing a new nonlocal framework for the basic equations of continuum mechanics. In this lecture we consider a nonlinear peridynamic model and discuss its well-posedness in suitable fractional Sobolev spaces.
Those results were obtained in collaboration with S. Dipierro (Perth), F. Maddalena (Bari) and E. Valdinoci (Perth). - Stochastic Optimization with Multiple Time Scales
Martin Glanzer, University of Vienna
martedì 21 gennaio 2020 alle ore 11:00 precise, Aula Seminari Terzo pianoABSTRACTReal-world multistage stochastic optimization problems are often characterized by the fact that the decision maker may take actions only at specific points in time, even if relevant data can be observed much more frequently. In such a case there are not only multiple decision stages present but also several observation periods between consecutive decisions, where profits/costs occur contingent on the stochastic evolution of some uncertainty factors. We present a tailor-made modeling framework for such problems, which allows for a computationally efficient solution. We first establish new results related to the approximation of (Markovian) stochastic processes by scenario lattices. In a second step, we incorporate the multiscale feature by leveraging the theory of stochastic bridge processes. The ingredients to our proposed modeling framework are elaborated explicitly for various popular examples, including both diffusion and jump models. In particular, we present new results related to the simulation of compound Poisson bridges. Finally, we discuss a valuation problem of a thermal power plant, where implementing our multiscale modeling framework turned out to be particularly convenient. If time permits, we incorporate model ambiguity into the power plant valuation problem and show some numerical results. - "Piu` che l'doppiar de li scacchi s'immilla" (Dante, Pd XXVIII, 93)
Riccardo Rosso, Università di Pavia
mercoledì 15 gennaio 2020 alle ore 15:00, Sala Consiglio - piano 7° - edificio 14ABSTRACTIn questo seminario verranno proposti alcuni problemi di indole matematica, ispirati dal gioco degli scacchi: dallo studio del percorso di un cavallo al numero massimo di regine che si possono disporre su una scacchiera senza che si possano minacciare. Si discuterà anche il metodo proposto nel XIX secolo da due matematici francesi, Delannoy e Lucas, per risolvere problemi di probabilità servendosi della scacchiera. Si darà risalto alla traduzione dei problemi in linguaggio matematico, in vista di un'eventuale fruizione didattica.
- Long-time asymptotics for evolutionary crystal dislocations models
Matteo Cozzi, University of Bath
martedì 17 dicembre 2019 alle ore 15:30, Aula seminari 3° pianoABSTRACTIn this talk, I will discuss a recent result concerning the long-time behavior of solutions to evolutionary Peierls-Nabarro type equations, related to crystal dislocations.
I will present the construction of solutions that, at large times, behave like a superposi- tion of an arbitrary finite number of fundamental dislocations, equally oriented and centered near points that evolve according to a repulsive dynamical system.
This result has been obtained in collaboration with J. D ?avila and M. del Pino (University of Bath). - Translating cardiac models into the clinic
Steven Niederer, Biomedical Engineering, King’s College London
giovedì 12 dicembre 2019 alle ore 14:00, Aula Consiglio VII Piano - Edificio 14, Dipartimento di Matematica POLITECNICO DI MILANOABSTRACTThe ability to measure the heart, its shape, its structure and its function across multiple spatial and temporal scales continues to grow. Interpreting this data remains challenging. Computational biophysical models of the heart allow us to quantitatively link and interpret these large disparate data sets within the context of known cardiac physiology and invariable physical constraints. Within these models, we can infer unobservable states, propose and test new hypothesis and predict how systems will respond to challenges increasing our ability to interrogate and understand biological systems. We are increasingly applying this approach to modelling human hearts to investigate clinical applications. In this presentation, I will give an overview on our modelling work simulating anthracycline-induced heart failure, how we are using models of individual patients to study cardiac resynchronisation therapy and how we are using simulations to characterise the anatomy and pathophysiology of atrial fibrillation patients. Finally, I will present some of our preliminary results on simulating the four-chamber heart to begin simulating the interactions between atrial and ventricular function.
Contact: alfio.quarteroni@polimi.it
This seminar is organized within the ERC-2016-ADG Research project iHEART - An Integrated Heart Model for the simulation of the cardiac function, that has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 740132) - The mysteries of L-values
Sarah Zerbes, University College London
martedì 10 dicembre 2019 alle ore 14:00, Sala di Rappresentanza, Dipartimento di Matematica, Via C. Saldini 50ABSTRACTL-functions are one of the central objects of study in
number theory. There are many beautiful theorems and many more open
conjectures linking their values to all kinds of arithmetic problems.
I will talk about the mysteries surrounding these L-values and
describe some of the progress that has recently been made towards
understanding them. - Explainability, intepretability and sensitivity analysis
Emanuele Borgonovo, Department of Decision Sciences, BIDSA, Bocconi University, Milano
venerdì 6 dicembre 2019 alle ore 14:30, Aula Saleri - VI pianoABSTRACT
A growing research activity is developing for increasing interpretability of machine findings. When complex architectures are used, analysts are, in fact, exposed to the black-box effect. This seminar will review several methods used both in the machine learning and in the simulation community to make the black box more transparent. We shall discuss tools such as partial dependence functions, layerwise relevance propagation, as well as present several local and global sensitivity analysis methods, also proposing new tools and new findings on popular tools.
Contatto: piercesare.secchi@polimi.it