Notiziario Scientifico
a cura del Dipartimento di Matematica G. Castelnuovo, Sapienza Università di Roma
Settimana dal 01-06-2020 al 07-06-2020
Mercoledì 03 giugno 2020
Ore 14:00, streaming all'indirizzo https://meet.google.com/vsz-nxkm-hav, Dipartimento di Matematica, Sapienza Università di Roma
seminario di Algebra e Geometria
Cristiano Spotti (Aarhus University)
On log Kähler-Einstein metrics
In this talk I will discuss examples and some geometric properties of KE metrics with cone angle singularities along possibly singular (in general worse than normal crossing) divisors.
Giovedì 04 giugno 2020
Ore 17:00, On line seminar, Web site for information: https://www.dinamici.org/dai-seminar/
DinAmicI: Another Internet Seminar (DAI Seminar)
Sandro Vaienti (Centre de Physique Théorique, Marseille (France))
Thermodynamic formalism for random weighted covering systems
We develop a quenched thermodynamic formalism for random dynamical systems generated by countably branched, piecewise-monotone mappings of the interval that satisfy a random covering condition. Joint with J. Atnip, G. Froyland and C. Gonzalez-Tokman Note: The zoom link to the seminar will be posted on https://www.dinamici.org/dai-seminar/ and on https://mathseminars.org/seminar/DinAmicI. Moreover, it will be streamed live on youtube via the DinAmicI channel: https://www.youtube.com/channel/UCyNNg155G3iLS7l-qZjboyg
Per informazioni, rivolgersi a: sorrentino@mat.uniroma2.it
Venerdì 05 giugno 2020
Ore 14:30, Ore 14:30, Videoconference talk - join the conference via this link, Dipartimento di Matematica - Università di Roma "Tor Vergata"
ARTS - Algebra and Representation Theory Seminar
Pawel DLOTKO (Swansea University)
TDA for medical data analysis, how can we help in the current pandemic?
Topological data analysis is a source of stable and explainable methods to analyze data. Those features are of the key importance in medical applications. In this talk I will review concepts of conventional and ball mapper. I will highlight how those tools have been used to analyze medical data, starting from work performed in Ayasdi, ending up in my work related to the current pandemic. I will finish by describing my work with Oxford Covid19 database (OxCDB).
N.B.: this talk is part of the activity of the MIUR Excellence Department Project CUP E83C18000100006
Per informazioni, rivolgersi a: gavarini@mat.uniroma2.it
Venerdì 05 giugno 2020
Ore 15:00, Modalità telematica, Dipartimento di Matematica e Fisica, Università degli Studi RomaTre
Seminario di Logica e Informatica Teorica
Michele Pagani (Universite Paris VII - Denis-Diderot)
Automatic differentiation in PCF
Backpropagation is a classic automatic differentiation algorithm computing the gradient of functions specified by a certain class of simple, first-order programs, called computational graphs. It is a fundamental tool in several fields, most notably machine learning, where it is the key for efficiently training (deep) neural networks. Recent years have witnessed the quick growth of a research field called differentiable programming, the aim of which is to express computational graphs more synthetically and modularly by resorting to actual programming languages endowed with control flow operators and higher-order combinators, such as map and fold. We extend the backpropagation algorithm to a paradigmatic example of such a programming language: we define a compositional program transformation from PCF (a Turing complete simply-typed lambda-calculus) to itself augmented with a notion of linear negation, and prove that this computes almost everywhere the gradient of the source program with the same efficiency as first-order backpropagation. The transformation is completely effect-free and thus provides a purely logical understanding of the dynamics of backpropagation.
Per partecipare al seminario, richiedere il link all'indirizzo email vitomichele.abrusci@uniroma3.it o cliccare sul seguente link Teams Meeting: link
Le comunicazioni relative a seminari da includere in questo notiziario devono pervenire esclusivamente
mediante apposita form da compilare online, entro le ore 24 del giovedì precedente la settimana
interessata. Le comunicazioni pervenute in ritardo saranno ignorate.
Per informazioni, rivolgersi all'indirizzo di posta elettronica
seminari@mat.uniroma1.it.
Coloro che desiderano ricevere questo notiziario via e-mail sono pregati
di comunicare il proprio indirizzo di posta elettronica a
seminari@mat.uniroma1.it.