UNIMIB SIAM Student Chapter - Matematica in azione

"Matematica in azione" is a cycle of seminars aimed at gathering young researchers (Master’s and PhD students) working in the field of Applied Maths in an accessible and informal environment.

This is the webex call link

Here you can see our old seminars

Learning Dynamical Models from Data

(16/10/2026, 15:30, U3-01)

Learning dynamical models from data means turning observations into systems we can predict, estimate, and control. From Gauss recovering the orbit of Ceres to modern data-driven control and AI, linear system identification shows how a few elegant tools of linear algebra, such as Hankel matrices, orthogonal projections, and the singular value decomposition, can reveal the hidden dynamics behind the data.

Learning Dynamical Models from Data: A Short Journey Through Linear System Identification

Prof. Vittorio de Iuliis

University of San Raffaele, Roma

Abstract

Estimating models from data is an old problem: in 1801 Gauss recovered the orbit of Ceres, laying the foundation for the method of least squares. Moving beyond static curve fitting, this seminar explores the identification of dynamical systems, where models must serve prediction, estimation, and control. We will see how central questions in linear system identification translate directly into linear algebra, involving equivalence classes under change of basis, ranks of Hankel matrices, orthogonal projections, and the singular value decomposition. Finally, the talk connects these classical foundations to the rapidly growing field of data-driven control and modern AI.

Poster 🖼️

Optimal Control Algorithms for Epidemiological Time-Delay Models

(15/09/2026)

Prof. Rami Katz

Tel Aviv University

Abstract

Optimal control is a widely used approach for controller synthesis in numerous disciplines, including epidemiology (intervention policies), biological engineering and synthetic biology (design of bio-molecular circuits), aerospace engineering (trajectory optimization) and robotics and automation (motion planning). In such applications, dynamical models that account for delays in the evolution of the system (i.e. time-delay systems - TDSs) are ubiquitous; for example, an epidemiological model can take into account the incubation period of a disease. Compared to regular ODEs, optimal control of TDSs imposes significant additional challenges in synthesizing controllers and designing reliable algorithms with theoretical guarantees to identify optimal controllers. In this talk, I will first present the general problem of optimal control of TDSs and discuss some of its analogies with the problem of minimizing functions. Then, I will present a recently developed sequential algorithm for optimal control of TDSs, and discuss its guarantees. Finally, I will show numerical simulations that demonstrate the efficacy of the algorithm in solving problems emanating from mathematical epidemiology.

Poster 🖼️

Bayesian Inference for Biodiversity

(09/06/2026)

Statistical inference on biodiversity acts as a powerful mathematical lens that reveals the hidden patterns of life within our most complex ecosystems. At its core, it leverages Bayesian nonparametric methods to map the hierarchical structure of nature, allowing us to estimate the fundamental diversity of our planet—from the smallest fragments to the vast tree flora of the Amazon.

A Bayesian theory for estimation of biodiversity

Prof. Tommaso Rigon

University of Milano-Bicocca

Abstract

Statistical inference on biodiversity has a rich history going back to RA Fisher. An influential ecological theory suggests the existence of a fundamental biodiversity number, denoted alpha, which coincides with the precision parameter of a Dirichlet process. Motivated by this theory, we develop Bayesian nonparametric methods for statistical inference on biodiversity, building on the literature on Gibbs-type priors. We argue that sigma-diversity is the most natural extension of the fundamental biodiversity number and discuss strategies for its estimation. Furthermore, we develop novel theory and methods starting with an Aldous-Pitman process, which serves as the building block for any Gibbs-type prior with a square-root growth rate. We propose a modeling framework that accommodates the hierarchical structure of Linnean taxonomy, offering a more refined approach to quantifying biodiversity. The analysis of a large and comprehensive dataset on Amazon tree flora provides a motivating application.

Poster 🖼️

Sustainability around the Earth

(20/05/2026)

Dynamical systems theory provides the mathematical lens to navigate the complex environment of Earth’s orbit, turning natural perturbations into a tool for environmental preservation. At its core, it reveals the existence of "de-orbiting corridors"—natural gravitational highways that allow us to safely dispose of space debris and ensure the long-term sustainability of our cosmic infrastructure.

How dynamical systems theory can help in the design of passive mitigation strategies for the space debris problem

Elisa Maria Alessi

IMATI - CNR

Abstract

Most of the artificial bodies orbiting the Earth are non-functional spacecraft, rocket bodies and fragments. As operational satellites are key assets for life on Earth, the international community is facing the space debris problem by means of different measures, from observations to regulations, from spacecraft design to advanced mathematics. The ultimate goal is to make sustainable the space exploitation. At the moment, one of the most effective mitigation measures is the so-called end-of-life disposal, i.e., the satellite, once the operations are concluded, cannot be left in its orbit, but must be disposed towards an orbit that will not cause any risk to the operational satellites. Natural gravitational and non-gravitational perturbations can be considered to reduce the cost of this maneuver. In this talk, I will introduce the space debris problem and I will focus on the coupled effect of Earth's oblateness and solar radiation pressure. The perturbation in the long term can be written as a series of resonant terms, that produce a variation in the orbit of the satellite. By isolating one term at a time, we can obtain an integrable model. The corresponding phase space can be used to define suitable initial conditions and area-to-mass rati to de-orbit the spacecraft towards the Earth's atmosphere, by increasing naturally the orbital eccentricity. Numerical simulations based on a full dynamical model have confirmed the existence of these trajectories, that have bene called "de-orbiting corridors".

Poster 🖼️

The nature of mathematical models

(13/05/2026)

Mathematical modeling acts as a formal bridge between conceptual theory and empirical observation, mapping complex realities—from physics to biomedicine—into a precise geometric language. At its core, it reveals that identifying a model is not just a calculation, but a way to represent the abstract correspondence between our ideas and the physical world through the geometry of estimation.

Prof. Andrea De Gaetano

Director of CNR-IRIB; Distinguished Professor at Óbuda University; Adjunct Professor at Mahidol University

Abstract

Mathematical modeling has become pervasive in applications, not only in physics or economy, but also in biomedicine and other "soft'' sciences. To the conceptual formulation of a model there often follows its identification by statistical parameter estimation, given available observations. While the nature of the modeling process as well as its relationship with the attending statistical computations could both appear obvious to the practitioner, it may be useful to formalize them in a precise way. Insight into the process of (linear and nonlinear) model parameter estimation can be obtained from the description of the geometry of estimation in case space. The objective then is to also describe the geometry of modeling in the abstract, and to show how the correspondence between the conceptual context and the computational context can be formally represented.

Poster 🖼️

Conservation laws (01/04/2026)

Conservation laws are mathematical equations that describe how a physical quantity changes over time. At their core, they state a simple but powerful principle: the total amount of a substance in a region changes only by the amount that flows in or out through its boundaries. Today, we explore two fascinating and diverse applications of these laws:

General Non Local Balance Laws: from Clustering to Cryptography

Andrea Salvadori

University of L'Aquila, University of Tours

ARZ-type model of vehicular traffic with local constrained flow

Davide Meroni

University of Milano-Bicocca

Abstract Meroni 📑

Abstract Salvadori 📑

Poster 🖼️