NumPI May 2015

numpi@lists.dm.unipi.it

1 participants
4 discussions

Fwd: [Dipartimento.di] avviso ciclo di seminari - 8, 10, 11 giugno
by Federico Poloni 04 Jun '15

04 Jun '15

In the framework of the research project "Mathematical models and computational methods for complex networks" by the University of Pisa, Paola Boito (Faculté des Sciences et Techniques Université de Limoges) will give three seminars on "Complex networks, functions of matrices and quadrature rules". The abstract is provided below. The seminars will take place in June at the Department of Computer Science of the University of Pisa: -- June 8th, 15:00-16:00, sala seminari est (room 351); -- June 10th, 10:00-11:00, sala seminari est (room 351); -- June 11th, 15:00-16:00, sala seminari est (room 351). Abstract. One cornerstone of our modern society is the use of different kinds of networks. Our cities are connected by a network of streets and railways, telecommunication networks including their wireless, mobile components, the internet, and the World Wide Web build the most important infrastructure for communication and information worldwide. Designing and managing such networks pose challenging algorithmic and numerical problems. These three talks are concerned with the development, analysis and application of numerical linear algebra tools for the quantitative study of important properties of large-scale complex networks, such as centrality, betweenness and communicability measures. Specifically, we design efficient algorithms for the fast approximation of the entries of matrix functions such as the exponential and the resolvent of adjacency matrices and graph Laplacians, as well as their traces. While these matrices are quite sparse, in the case of complex networks they behave very differently from the matrices associated with regular lattices (grids), such as those arising from discretizations of partial differential equations. Hence, there is a need to develop new methods, both computational and analytical, to deal with these very large-scale problems. Potential applications include the analysis of social networks, the study of biological and neurological networks, applications in physics and operations research, and so forth.

1 1

Optimization & Numerical Analysis Seminar: N. Vannieuwenhoven
by Federico Poloni 03 Jun '15

03 Jun '15

Speaker: Nick Vannieuwenhoven (KU Leuven) Title: Towards a perturbation theory for the tensor rank decomposition Date: Thursday, 4 June 2015. Time: 11:00 Place: Sala Seminari Ovest, dipartimento di informatica Abstract: The tensor rank decomposition is a decomposition of a tensor into a linear combination of rank-1 tensors. One of the key advantages of higher-order tensors is that this decomposition is generally unique, which allows for an interpretation of the individual rank-1 terms. Because of this property, the tensor rank decomposition has found application in several domains. For instance, they can be used as a clustering algorithm in an unsupervised learning setting assuming a certain statistical model wherein each of the rank-1 tensors in the decomposition will correspond with a cluster. In the applications where tensor rank decompositions arise, the tensor is usually known only up to some small perturbation error. This raises the following interesting problem: Suppose that the tensor is perturbed by a small quantity, by how much can the individual rank-1 terms change relative to the magnitude of the perturbation? Can it be unbounded? In this presentation, I will explore the definition of a possible condition number that answers these questions up to first order. == Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 1

Reminder: Optimization & Numerical Analysis Seminars: F. Romani
by Federico Poloni 13 May '15

13 May '15

Date: Thursday May 14, 2015 Time: 11:00 Place: Sala Seminari Ovest, dipartimento di informatica Speaker: Francesco Romani, dipartimento di informatica, U Pisa. Title: *Krylov subspace methods for the solution of linear systems* Abstract: Krylov subspace methods are iterative methods for the solution of linear systems which operate by multiplying the system matrix (sometimes also its transpose) by a vector. They are recommended when the system matrix is large and sparse or with a strong structure. The most common Krylov subspace methods are presented focusing the attention mainly on the practical and empirical aspects. == The seminar will be in Italian (but with slides in English). Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis Seminars: F. Romani
by Federico Poloni 07 May '15

07 May '15

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

NumPI May 2015