Personal Data

Name: Gaetano Zanghirati
Citizenship: Italian
Birth: Ferrara, August 23, 1966
Address: University of Ferrara, Department of Mathematics, second site,
      Scientific-Technological Campus, Building B
      via Saragat 1, 44121 Ferrara, Italy
Phone: +39 0532 974784
Fax: +39 0532 974787
E-mail: g.zanghirati@unife.it

Current position
Associate Professor of Numerical Analysis (MAT/08) at the Mathematics, Physics and Natural Sciences Faculty of the University of Ferrara.

Education

1991
Degree in Mathematics (cum laude) at the University of Ferrara, Italy
(thesis: "Congruences, remainder classes and probabilistic primality tests")

1991-1992
Lecturer at the Department of Mathematics, University of Ferrara.

1993-1996
Ph.D. student, Computational Mathematics and Mathematical Computer Science Ph.D. School, Department of Pure and Applied Mathematics, University of Padua (Italy).
Studentship winner (mark: 58/60; rank 2/23).

1997
Ph.D. in Computational Mathematics and Mathematical Computer Science (thesis defense at the University of Naples, Italy).

1996-1998
Two fellowships of the Ferrara Research Center on Mathematical Methods for SPECT and MRI medical images reconstruction.

1998-1999
Post-Doc position at the Department of Mathematics, University of Ferrara.


Summer Schools

1996
Mathematical Models and Inverse Problems, June 30 - July 16:
Mathematical models and Inverse problems,
Prof. I. Galligani, University of Bologna (Italy).
The Dirichelet to Neumann map and inverse problems,
Prof. G. Uhlmann, University of Washington in Seattle (USA).

1994
Numerical Analysis, July 3 - 23:
Numerical solution of differential equations,
Prof. A. Iserles, Cambridge University (UK).
Boundary value methods,
Prof. D. Trigiante, University of Florence (Italy).


Experience

1997-1998
Head of the parallel software development team for the Esprit - PST European Community
HPCN (High Performance Computing and Networking) Project called "PALMA" (Parallel Assets and Liabilities MAnagement)

1989-1991
High-school teacher in Mathematics and Physics


Research activity

Research interests

Numerical algorithms for large-scale nonlinear optimization problems. Optimization problems in Statistical Learning and Inverse Problems. Parallel algorithms for large-scale quadratic programs. Projection-type methods for large-scale quadratic programs. Parallel Interior-Point methods for large-scale quadratic problems. Parallel numerical algorithms for specially structured nonlinear systems.

Visiting

2007, April to July
University of Dundee (UK), Department of Mathematics and Computer Science.
Inviting: Prof. Roger Fletcher

2006, November
University of Groningen (The Netherlands), Neuro-Imaging Center.
Inviting: Prof. Christian Keysers

2006, November
University of Chicago (USA), Toyota Technological Institute.
Inviting: Dr. A. Caponnetto

2006, January
Agricultural University of China, Beijing (China), Department of Mathematics
Inviting: Prof. Y. Deng

2005, July
University of Dundee (UK), Department of Mathematics and Computer Science.
Inviting: Prof. Roger Fletcher

2005, May
University of Dundee (UK), Department of Mathematics and Computer Science.
Inviting: Prof. Roger Fletcher


Referred papers

  • R. Zanella, G. Zanghirati, R. Cavicchioli, L. Zanni, P. Boccacci, M. Bertero, G. Vicidomini,
    Towards real-time image deconvolution: application to confocal and STED microscopy,
    SCIENTIFIC REPORTS 3 (2013), pp. 1-8, with additional material (16 pages), ISSN: 2045-2322, doi: 10.1038/srep02523.

  • R. Cavicchioli, A. Prearo, R. Zanella, G. Zanghirati, L. Zanni,
    Optimization methods for digital image restoration on MPP multicore architectures,
    Quaderni di Matematica 27 (2012), 93-116. Special issue "Recent Advances in Nonlinear Optimization and Equilibrium Problems: A tribute to Marco D'Apuzzo", a cura di V. De Simone, D. di Serafino, G. Toraldo; Aracne Editrice.

  • R. Fletcher, G. Zanghirati,
    Binary separation and training support vector machines,
    Acta Numerica 19 (2010), Cambridge University Press, UK, pp. 121-158.

  • M.C. Pietrogrande, D. Bacco, N. Marchetti, M. Mercuriali, G. Zanghirati,
    2D autocovariance function for comprehensive analysis of two-way GC-MS data matrix: Application to environmental samples,
    Talanta (2011), pp. 1225-1232.

  • L. Sambo, J.B. Stephen, S. Bonettini, G. Zanghirati, F. Frontera,
    Improving the angular resolution of coded aperture instruments using a modified Lucy-Richardson algorithm for deconvolution,
    Proceedings of Science, vol. PoS(extremesky2009)099 (2009), SISSA, Trieste (Italy), pp. 1-6.

  • A. Giovannini, G. Zanghirati, M.A. Beaumont, L. Chikhi, G. Barbujani,
    A Novel Parallel Approach to the Likelihood-based Estimation of Admixture in Population Genetics,
    Bioinformatics 25(11)(2009), pp. 1440-1441.

  • M. Bottoni, S. Mantovani, G. Zanghirati,
    Numerical Simulations Of Thermal Confinement of Particles in Rarefied Gases with Molecular Dynamics,
    Computational Fluid Dynamics Journal 18(1) (2009), pp. 116-134.

  • G. Frassoldati, L. Zanni, G. Zanghirati,
    New adaptive stepsize selections in gradient methods,
    Journal of Industrial and Management Optimization 4(2) (2008), pp. 299-312.
    Also available at Optimization Online 2007.

  • G. Zanghirati, R. Fletcher,
    Parallel gradient methods for some classes of large-scale nonlinear programming methods,
    Science and Supercomputing in Europe (2008), (P. Alberigo, G. Erbacci, F. Garofalo, S. Monfardini, Eds.), pp. 842-846.

  • M. Bottoni, S. Mantovani, G. Zanghirati,
    Review of Ongoing Developments for Molecular Dynamics Simulation of Thermophoretic Problems,
    Proceedings of the 20th International Conference on Systems Research, Informatics and Cybernetics (July 24-30, 2008, Baden-Baden, Germany),
    IIAS Transactions on Systems Research and Cybernetics 8(1) (2008), pp. 33-42.

  • M. Prato, L. Zanni, G. Zanghirati,
    On Recent Machine Learning Algorithms for Brain Activity Interpretation,
    Proceedings of the 23rd Annual Review of Progress in Applied Computational Electromagnetics, Verona (Italy), March 19-23, 2007,
    Applied Computational Electromagnetics Society Journal 22 (2007), pp. 1939-1946.

  • L. Zanni, T. Serafini, G. Zanghirati,
    Parallel Software for Training Large-Scale Support Vector Machines on Multiprocessor Systems,
    Journal of Machine Learning Research 7 (2006), pp. 1467-1492.

  • T. Serafini, L. Zanni, G. Zanghirati,
    Some improvements to a parallel decomposition technique for training support vector machines,
    Lecture Notes in Computer Science 3666 (2005), pp. 9-17.

  • T. Serafini, G. Zanghirati, L. Zanni,
    Gradient projection methods for quadratic programs and applications in training support vector machines,
    Optimization Methods and Software 20 (2005), pp. 353-378.
    Also available at Optimization Online 2003.

  • T. Serafini, G. Zanghirati, L. Zanni,
    Parallel Decomposition Approaches for Training Support Vector Machines,
    in "Parallel Computing: Software Technology, Algorithms, Architectures and Applications", G.R. Joubert, W.E. Nagel, F.J. Peters and W.V. Walter, Eds.,
    Advances in Parallel Computing 13, Amsterdam, The Netherlands, 2004, pp. 259-266.

  • T. Serafini, G. Zanghirati, L. Zanni,
    Large quadratic programs in training Gaussian support vector machines,
    Rendiconti di Matematica e delle sue Applicazioni, University "La Sapienza", Rome, Series VII, Vol. 23 (2003), pp. 257-275.

  • G. Zanghirati, L. Zanni,
    A parallel solver for large quadratic programs in training support vector machines,
    Parallel Computing 29(4) (2003), pp. 535-551.

  • E. Loli Piccolomini, F. Zama, G. Zanghirati, A.R. Formiconi,
    Regularization methods in dynamic MRI,
    Applied Mathematics and Computation 132 (2002), pp. 325-339.

  • C. Durazzi, V. Ruggiero, G. Zanghirati,
    Solving a special class of discrete optimal control problems via a parallel interior-point method,
    in "Equilibrium Problems and Variational Models" (P. Daniele, F. Giannessi, A. Maugeri, Eds.),
    Kluwer Academic Publishers, Dordrect, The Netherlands, 2003, pp. 141-161.

  • C. Durazzi, V. Ruggiero, G. Zanghirati,
    A Parallel Interior-Point Method for Linear and Quadratic Programs with Special Structure,
    Journal of Optimization Theory and Applications 110(2) (2001), pp. 289-313.

  • A.R. Formiconi, E. Loli Piccolomini, S. Martini, F. Zama, G. Zanghirati,
    Numerical methods and software for functional magnetic resonance images reconstruction,
    Annals of the University of Ferrara 46(suppl.) (2000), pp. 87-102.

  • G. Zanghirati, F. Cocco, G. Paruolo, F. Taddei,
    A Cray T3E implementation of a parallel stochastic dynamic assets and liabilities management model,
    Parallel Computing 26(2000), pp. 539-567.

  • G. Zanghirati,
    Global convergence of nonmonotone strategies in parallel methods for block-bordered nonlinear systems,
    Applied Mathematics and Computation 107 (2000), pp. 137-168.

  • G. Zanghirati, F. Cocco, F. Taddei, G. Paruolo,
    Cray T3E performance of a parallel code for a stochastic dynamic ALM model,
    Lecture Notes in Computer Science 1685, Euro-Par '99 Parallel Processing, Springer, 1999, pp. 1176-1186.

  • G. Zanghirati,
    Some theoretical properties of Feng-Schnabel algorithm for block-bordered nonlinear systems,
    Optimization Methods and Software 10(6) (1999), pp. 783-801.

  • G. Zanghirati,
    Parallel computational experience and dynamic scaling for a class of nonlinear systems,
    in "Numerical Methods in Optimization" (A. Maugeri, E. Galligani, Eds),
    Rendiconti del Circolo Matematico di Palermo 58 (1999), serie II (suppl.),
    pp. 229-246.

  • P. Codecà, L. Biasini, G. Zanghirati,
    Implementazione di tests probabilistici di primalità,
    Annals of the University of Ferrara, Sec. VII, Mathematical Sciences, Vol. XXXVIII (1992), pp. 117-132.


Other papers

  • M.C. Pietrogrande, M. Mercuriali, N. Marchetti, L. Pasti, D. Bacco, G. Zanghirati, F. Dondi,
    A chemometric approach based on the AutoCoVariance Function for handling complex signals from environmental monitoring,
    TIES, International Environmetrics Society, University of Bologna, Book of Abstracts 2009, pp. 83-93.

  • T. Serafini, L. Zanni, G. Zanghirati,
    Parallel training of Large-Scale Kernel Machines,
    in "Science and Supercomputing at CINECA", Report 2005 (2006), Bologna, Italy, pp. 415-419.

  • T. Serafini, L. Zanni, G. Zanghirati,
    Training Support Vector Machines on Parallel Architectures,
    in "Science and Supercomputing at CINECA", Report 2003 (2004), Bologna, Italy, pp. 391-394.

  • C. Durazzi, V. Ruggiero, G. Zanghirati,
    Some Applications via a Parallel Interior-Point Method,
    in "Science and Supercomputing at CINECA", Report 2001 (2002), Bologna, Italy, pp. 418-424.

  • E. Loli Piccolomini, F. Zama, G. Zanghirati, A. R. Formiconi, S. Martini,
    MRItool: a MATLAB tool for functional Magnetic Resonance Images reconstruction,
    Monograph n. 2, COFIN '97 MURST Project "Numerical Analysis: Methods and Mathematical Software", Monograf, Bologna, Italy, 2000, 40pp..

  • G. Zanghirati,
    Uso della decomposizione ai valori singolari nella compressione di immagini digitali,
    Quaderni di Matematica Applicata 11 (1993), University of Ferrara, Ferrara, Italy, pp. 1-11.

  • G. Zanghirati,
    Metodi paralleli per problemi di programmazione quadratica con vincoli di uguaglianza,
    Quaderni di Matematica Applicata 12 (1995), University of Ferrara, Ferrara, Italy, pp. 27-43.

  • G. Zanghirati,
    Convergenza quadratica locale di metodi paralleli quasi-Newton per sistemi di equazioni non lineari bordati a blocchi,
    Atti dell'Accademia delle Scienze dell'Istituto di Bologna, Classe di Scienze Fisiche, anno 283, serie V, n. 7 (1996), pp. 41-66.

  • G. Zanghirati,
    Metodi paralleli per sistemi di equazioni non lineari bordati a blocchi,
    Ph.D. Thesis (in Italian), Department of Pure and Applied Mathematics, University of Padua (1997), Padua, Italy.

  • G. Zanghirati,
    Metodi paralleli per sistemi di equazioni non lineari bordati a blocchi,
    Bollettino U.M.I. 8(suppl.) (1998), pp. 205-208.

  • G. Zanghirati, V. Ruggiero,
    A Cray T3D numerical experience on block bordered nonlinear systems,
    Science and Supercomputing at CINECA, Report '97 (1998), pp. 570-586.

  • A. Guizzardi, P. Paruolo, G. Zanghirati,
    Stochastic simulation of models for expected asset returns based on neural networks: a parallel experience,
    Science and Supercomputing at CINECA, Report '97 (1998), pp. 225-227.


Books

Monographs

  • E. Loli Piccolomini, F. Zama, G. Zanghirati, A. R. Formiconi, S. Martini,
    MRItool: a MATLAB tool for functional Magnetic Resonance Images reconstruction,
    Monograph n. 2, MURST COFIN '97 Project "Numerical Analysis: Methods and Mathematical Software", Ferrara, Italy, 2000.


Technical Reports

  • R. Cavicchioli, A. Prearo, R. Zanella, G. Zanghirati, L. Zanni,
    Iterative optimization methods for efficient image restoration on multicore architectures, Technical Report 373, Department of Mathematics, University of Ferrara, Italy (2011).

  • G. Frassoldati, L. Zanni, G. Zanghirati,
    New Adaptive Stepsize Selections in Gradient Methods,
    Technical Report 77, Department of Pure and Applied Mathematics "G. Vitali", University of Modena and Reggio Emilia, Italy (2007).

  • E. Galligani, V. Ruggiero, G. Zanghirati,
    Splitting methods for nonlinear diffusion filtering,
    Technical Report 73, Department of Pure and Applied Mathematics "G. Vitali", University of Modena and Reggio Emilia, Italy (2005).

  • T. Serafini, G. Zanghirati, L. Zanni,
    Gradient Projection Methods for Quadratic Programs and Applications in Training Support Vector Machines,
    Technical Report 48, Department of Pure and Applied Mathematics "G. Vitali", University of Modena and Reggio Emilia, Italy (2005).

  • G. Zanghirati, L. Zanni,
    Variable Projection Methods for Large Quadratic Programs in Training Support Vector Machines,
    Technical Report 339, Department of Mathematics, University of Ferrara, Italy (2003).

  • G. Zanghirati, L. Zanni,
    Large Quadratic Programs in Training Gaussian Support Vector Machines,
    Technical Report 320, Department of Mathematics, University of Ferrara, Italy (2002).

  • A.R. Formiconi, E. Piccolomini, F. Zama, G. Zanghirati,
    Regularization of constrained dynamic magnetic resonance imaging.
    Technical Report 270, Department of Mathematics, University of Ferrara, Italy (1998).



Software


  • SGP-dec - Scaled Gradient Projection technique for image deconvolution v. 1.0 (2011)
    by R. Cavicchioli, R. Zanella, G. Zanghirati, L. Zanni
    Available at www.unife.it/prisma/software
    Description. A Matlab package for the deconvolution of 2D and 3D images corrupted by Poisson noise. Following a maximum likelihood approach, SGP-dec computes a deconvolved image by early stopping the scaled gradient projection (SGP) algorithm for the solution of the optimization problem coming from the minimization of the generalized Kullback-Leibler divergence between the blurred image and the observed image.
    This code is developed with the support of the PRIN2008 Italian national research project "PRISMA - Optimization Methods and Software for Inverse Problems", grant n. 2008T5KA4L.

  • PGPDT - Parallel Gradient-Projection-based Decomposition Technique
    by T. Serafini, L. Zanni, G. Zanghirati
    Available at dm.unife.it/gpdt
    Description. PGPDT is the parallel version of the C++ software GPDT. This version is designed for distributed memory multiprocessor architectures and it shares with the scalar version the same problem decomposition and gradient-projection methods.

  • GPDT - Gradient-Projection-based Decomposition Technique
    by T. Serafini, L. Zanni, G. Zanghirati
    Available at dm.unife.it/gpdt
    Description. GPDT is a C++ software designed to train large-scale Support Vector Machines (SVMs) for binary classification on scalar environments. It uses a popular problem decomposition technique to split the SVM quadratic programming (QP) problem into a sequence of smaller QP subproblems, each one being solved by a suitable gradient-projection method (GPM). The currently implemented GPMs are the Generalized Variable Projection Method (GVPM) and the Dai-Fletcher method (DFGPM).
    The software is available also in the Machine Learning repository mloss.org and it has been selected for inclusion in other widely used software packages such as the Matlab toolbox SHOGUN.

  • ParLEA - Parallel Likelihood-based Estimation of Admixture. Version 0.2 (2008)
    by A. Giovannini, G. Zanghirati, M. Beaumont, L. Chikki, G. Barbujani
    Available at dm.unife.it/parlea
    Description. Estimating genetic population admixture is a difficult task that is shown to have meaningful applications in evolutionary and conservation Biology. Unfortunately state-of-the-art probabilistic approaches are computationally demanding. The ability of effectively exploiting the computational power of modern multiprocessor systems can thus contribute a beneficial impact to Monte-Carlo-based simulation of admixture modeling. Parallel LEA is a novel parallel approach for Likelihood-based Estimation of Admixture in population genetics, MPI and C++ based.

  • JEpidemicS - a Java tool for epidemic spread simulation through distance-dependent probabilities. Version 0.9beta (2010)
    by E. Ganzaroli, A. Corli, G. Zanghirati.
    Description. The software package is designed mainly for the simulation of the African Suine Fever (ASF) epidemic spreading in wide geographical areas and medium-to-long distances. It uses a graph-based multilayer structure and follows the object-oriented programming (OOP) paradigm, to allow for maximum flexibility, extendibility, easy of maintenance and connectivity to other available tools. The software is implemented in the Java programming language to ensure for maximum portability on a number of different hardware/software platforms, without the need for any source code adaptation. The whole project is developed by using only open-source libraries and public-domain databases of Earth maps. It allows the user to manually set the data or to load field data from external files. The simulations are based on a deeply studied model and allow the user to save the results in external files for statistical purposes. The package is different from other available simulation tools in the underlying model and in the features provided.
    This software is part of a joint project of the underlying study and the developed code have been carried on in strict cooperation with Dr. V. Guberti. The project was a joint effort of the "Mathematics for Technology" Center of the University of Ferrara and the Food and Agricolture Organization (FAO) of the United Nations (UN).

  • EpidemicS - Epidemic Simulator. Version 1.0 (2009).
    by E. Ganzaroli, A. Corli, G. Zanghirati.
    Description. This is a MATLAB code based on cellular automata for the simulation of the disease spreading in wild animals, such as the suine fever in wild boars. The project has been developed in strict cooperation with Dr. V. Guberti and Dr. M. Fenati and was a joint effort of the "Mathematics for Technology" Center of the University of Ferrara and of the Italian National Institute for Wild Animals (INFS-CNR).

  • ParTHERCONF - Parallel THERmal CONFinement. Version 0.8 (2008)
    by M. Bottoni, S. Mantovani, G. Berdondini, G. Zanghirati.
    Description. Parallel code that describes the displacement due to thermophoresis of tiny particles in a gas-filled domain subjected to a temperature gradient. The behavior of the gas molecules and their interaction with material surfaces and with the particles floating in the gas are modeled with the statistical methods of molecular dynamics, based on the "Direct Simulation Monte Carlo" (DSMC). In case of polyatomic molecules sharing of the energy between translational and rotational modes is simulated with the Larsen-Borgnakke model. Interaction of polyatomic molecules with material surfaces is computed on the basis of the Cercignani-Lampis-Lord model. The computational code has extensive post-processing capabilities for visualization of computational results.
    The software is coded in Fortran 77 with MPI communication routines. The serial version is called THERCONF-2D. The code is not available for distribution: please contact Dr. Eng. Maurizio Bottoni for additional information.

  • MRItool - A Matlab tool for functional Magnetic Resonance Image reconstruction. Version 1.0 (2000)
    by E. Loli Piccolomini, M. Romagnoli, F. Zama, G. Zanghirati, A.R. Formiconi, S. Martini
    Available at dm.unife.it/mritool
    Description. This package performs the reconstruction of sequences of Functional Magnetic Resonance images with the Generalized Series approach. It has been developed in the context of the Italian MURST project "Numerical Analysis: Methods and Mathematical Software".

  • PALMA - Parallel Assets and Liabilities MAnagement
    by G. Zanghirati and Prometeia Calcolo s.r.l. (1998)
    Description. This is a C++ parallel software developed jointly with SMART s.r.l. and Prometeia Calcolo s.r.l. as a EU Esprit project in the High-Performance Computing and Networking (HPCN) action. The package is aimed to simulate and analyze the assets and liabilities dynamics for a financial firm or a bank, considering a variety of possible parameter setting and a large number of independent temporal scenarios.


Recent Conference Presentations

  • First Order Methods for Image Deconvolution in Microscopy
    • , accepted at SIOPT2014 - SIAM Conference on Optimization 2014, May 19-22, 2014, San Diego (CA, USA) (with R. Zanella, G. Vicidomini, L. Zanni).

  • Optimization methods for real-time image deconvolution on GPU architectures, NUMTA2013 - International Conference on Numerical Computations: Theory and Algorithms, June 17-23, 2013, Falerna (CZ) (with R. Zanella, L. Zanni, P. Boccacci, M. Bertero, G. Vicidomini).

  • A joint inversion algorithm for image reconstruction (with A. Giovannini),
    SC2011 – Scientific Computing 2011, International Conference, Cagliari (Italy), October 19–23, 2011.

  • Constrained optimization methods for image reconstruction on multicore systems (with R. Zanella, R. Cavicchioli, L. Zanni),
    AFG11 – Austrian-French-German Conference on Optimization 2011, Toulouse (France), September 19–23, 2011.

  • An HPC joint inversion code for image reconstruction (with A. Giovannini),
    AFG11 – Austrian-French-German Conference on Optimization 2011, Toulouse (France), September 19–23, 2011.

  • A novel approach to the joint inversion of loosely connected data (with A. Giovannini, G. Vignoli),
    AIP2009 - International Conference on Applied Inverse Problems, Wien (Austria), July 20-24, 2009.

  • A new look to the primal problem in SVM training (with R. Fletcher),
    EURO2009 - 23rd International Conference on Operational Research, Bonn (Germany), July 5-8, 2009.

  • Improving the angular resolution of coded aperture instruments using a modified Lucy-Richardson algorithm for deconvolution (with L. Sambo, J.B. Stephen, S. Bonettini, F. Frontera),
    Intn. Conf. "The Extreme Sky: Sampling the Universe above 10 keV", poster session, Otranto (Lecce, Italy), October 13-17, 2009.

  • A fully flexible, agent-based model for epidemic spreading in wild animals (with E. Ganzaroli, A. Corli),
    wANPE08 - Intn. Workshop on Analysis and Numerics of Population dynamics and Epidemics models, Udine (Italy), December 15-17, 2008.

  • Modelling external noise in epidemics models of bacteriophage infection (with M. Carletti, G. Notarangelo),
    wANPE08 - Intn. Workshop on Analysis and Numerics of Population dynamics and Epidemics models, Udine (Italy), December 15-17, 2008.

  • Review of ongoing developments for molecular dynamics simulations of thermophoretic problems (with M. Bottoni, S. Mantovani),
    20th Meeting of the IIAS (International Institute for Advanced Studies in Systems Research and Cybernetics), Baden-Baden (Germany), July 24-30, 2008.

  • Separating clusters of points and training SVMs
    (with R. Fletcher),
    International workshop on Optimization Techniques for Inverse Problems, Modena (Italy), April 28-29, 2008.

  • Separating clusters of points (with R. Fletcher),
    plenary opening talk, NA07 - 22nd Biennial Conference on Numerical Analysis, Dundee (UK), June 26-29, 2007.

  • On Adaptive Step-Size Selections in Gradient Methods (with G. Frassoldati, L. Zanni),
    invited talk, ICIAM07 - 6th International Conference on Industrial and Applied Mathematics, Zurich (Switzerland), July 16-20, 2007.

  • On Recent Machine Learning Algorithms for Brain Activity Interpretation (with M. Prato, L. Zanni),
    23rd International Review of Progress in Applied Computational Electromagnetics, Verona (Italy), March 2--5, 2007.

  • Effective large-scale optimization techniques for support vector machines (with T. Serafini, L. Zanni),
    invited talk, 8th Workshop of the ECRIM Working Group on Matrix Computations and Statistics, Salerno (Italy), September 2-3, 2006.

  • Some properties of gradient-based methods with application to machine learning (with T. Serafini, L. Zanni),
    invited talk, EuroXXI - 21st European Conference on Operational Research, Reykjavik (Iceland), July 2-5, 2006.

  • Large-scale Support Vector Machines: Decomposition and Cascade Approaches (with T. Serafini, L. Zanni),
    invited talk, EuroXXI - 21st European Conference on Operational Research, Reykjavik (Iceland), July 2-5, 2006.

  • On gradient projection-based and other techniques for SVMs training
    invited seminar, China Agricultural University, Beijing (China), January 13, 2006.

  • Splitting methods for nonlinear diffusion filtering (with E.~Galligani, V.~Ruggiero),
    3rd IASC World Conference on "Computational Statistics and Data Analysis", Limassol (Cyprus), October 28-31, 2005.

  • Gradient Projection-Type Quadratic Solvers in Parallel Decomposition Techniques for Support Vector Machines (with T. Serafini, L. Zanni),
    3rd IASC World Conference on "Computational Statistics and Data Analysis", Limassol (Cyprus), October 28-31, 2005.

  • Some Improvements to a Parallel Decomposition Technique for Training Support Vector Machines (with T. Serafini, L. Zanni),
    EURO PVM-MPI 2005, 12th European Parallel Virtual Machine and Message Passing Interface Conference, Sorrento (Italy), September 18--21, 2005.

  • Numerical Topics on SVMs Classification (with T. Serafini, L. Zanni),
    workshop "ASTAA Project Meeting 2005", Genova (Italy), May 9-10, 2005.

  • Decomposition techniques and gradient projection methods in training Support Vector Machines (with T. Serafini, L. Zanni),
    "SIAM Conference on Optimization 2005", Stocholm (Sveden), May 15--19, 2005.

  • On gradient projection-based decomposition techniques for training SVMs on parallel architectures (with T. Serafini, L. Zanni),
    "PASCAL Workshop", Thurnau (Germany), March 16-18, 2005.


Before 2005

  • Recent improvements to gradient projection-based decomposition techniques for Support Vector Machines (with T. Serafini, L. Zanni),
    International Congress "MML04 - Mathematical Methods for Learning 2004", Villa Geno (Como, Italy), June 21-24, 2004.

  • Recent Advances in Parallel Approaches to Large-Scale Support Vector Machines (with T. Serafini, L. Zanni),
    International Congress ``Scicomp9'', Bologna (Italy), March 23-26, 2004.

  • Nonstandard Training techniques for SVMs (with T. Serafini, L. Zanni),
    workshop "ASTAA Project Meeting 2004", Sestri Levante (Genova, Italy), February 6-7, 2004.

  • Accelerazione della convergenza in metodi del gradiente proiettato per problemi di programmazione quadratica (with T. Serafini, L. Zanni),
    XVII Convegno Nazionale dell'Unione Matematica Italiana, Milano (Italy), September 8-13, 2003.

  • Steplength selections in gradient projection methods for large-scale quadratic programs (with T. Serafini, L. Zanni),
    AIRO Conference 2003, Venezia (Italy), September 2-5, 2003.

  • Parallel Decomposition Approaches for Training Support Vector Machines (with T. Serafini, L. Zanni),
    International Conference "ParCo2003 - Parallel Computing 2003", Dresda (Germany), September 2-5, 2003.

  • Parallel Training of Support Vector Machines (with T. Serafini, L. Zanni),
    INdAM International Workshop "OPT2003 – Numerical Methods for Local and Global Optimization", Cortona (italy), July 14-20, 2003.

  • Adaptive Steplength Selections in Gradient Projection Methods for Quadratic Programs (with T. Serafini, L. Zanni),
    NA03 – 20th Biennial Conference on Numerical Analysis, Dundee (UK), June 24-27, 2003.

  • Variable Projection Decomposition Techniques for Large-Scale Support Vector Machines (with T. Serafini, L. Zanni),
    National Congress "CNAN2002 - Analisi Numerica: Stato dell'Arte", Rende (Cosenza, Italy), September 2002.

  • Parallel Solution of Large Quadratic Programs in Training Support Vector Machines (with L. Zanni),
    NATO-ASI Conference on $quot;Learning Theory and Practice", Loeven (Belgium), July 8-19, 2002.

  • Decomposition techniques in training support vector machines: inner QP solvers and parallel approaches (with L. Zanni),
    International Workshop on "Mathematical Diagnostics", Erice (Trapani, Italy), June 17-26, 2002.

  • Decomposition Techniques for Large Quadratic Programs in Training Support Vector Machines (with L. Zanni),
    invited talk at the International Conference APMOD 2002, Milano (Italy), June 17-21, 2002.

  • A Parallel Solver for Large Quadratic Programs in Training Support Vector Machines (with L. Zanni),
    invited talk at the 2002 SIMAI National Conference, Chia (Cagliari, Italy), May 26-31, 2002.

  • Variable Projection Methods for Quadratic Programs in Training Support Vector Machines with Gaussian Kernels (with L. Zanni),
    International Conference "SIAM Meeting on Optimization 2002", Toronto (Canada), May 20-22, 2002.

  • Programmazione quadratica di grandi dimensioni in problemi di apprendimento da esempi (with L. Zanni),
    National Workshop "GNCS2002, Scientific Computing National Group Annual Conference", Ferrara (Italy), February 12-13 2002.

  • Tecniche di decomposizione per l'Ottimizzazione nell'addestramento di SVMs (with L. Zanni),
    National Workshop "Alcune applicazioni e questioni aperte in ottimizzazione non lineare", Modena (Italy), January 23-24, 2002.

  • Large Quadratic Programs in Training Support Vector Machines (with L. Zanni),
    International Workshop on "Numercial Methods for Evolutionary Problems", Peschici (Foggia), September 17-21, 2001.

  • Training Support Vector Machines: a Parallel Approach (with L. Zanni),
    International Conference "Parallel Computing 2001", Napoli (Italy), September 4-7, 2001.

  • Numerical solution of large quadratic programs in training support vector machines with Gaussian kernels (with L. Zanni),
    20th IFIP TC7 Conference on "System Modelling and Optimization", Trier (Germany), July 23-27, 2001.

  • Parallel solution of quadratic programs in training support vector machines (with A. Verri, L. Zanni),
    33rd International Workshop on "High Performance Algorithms and Software for Nonlinear Optimization", Erice (Trapani, Italy), June 30 - July 8, 2001.


Organization of Conferences and Schools

Contributions to research projects

  • Principal Investigator in international projects

    • Study and development of parallel gradient methods for some classes of large-scale nonlinear programming problems, EU FP VI project "HPC-Europa. Pan-European Research Infrastructure on High-Performance Computing", www.hpc-europa.org), contract n. R113-CT-2003-506079, 2007.


  • Principal Investigator of national and local research projects

    • LaSPIM - Large-scale parallel computing for inverse problems in imaging, ISCRA class C project at the CINECA Supercomputing Center, n. HPHP10C312GC, April 2012 - March 2013.

    • ParJoInv - Parallel Joint Inversion, ISCRA class C project at the CINECA Supercomputing Center, n. HP10CQOVYA, November 2010 - September 2011.

    • NOCSiMA - Numerical Optimization, Scientific Computing and Simulation for Multidisciplinary Applications, local research project FAR2011, University of Ferrara, 2011 - 2014.

    • MultiNOPaC - Multidisciplinary issues of Numerical Optimization and Parallel Computing, local research project FAR2011, University of Ferrara, December 2010 - December 2013.

    • MultiNOPaC - Multidisciplinary issues of Numerical Optimization and Parallel Computing, local research project FAR2011, University of Ferrara, November 2009 - November 2012.


  • Italian Ministry-funded national projects

  • National Group for Scientific Computing (GNCS) projects

    • Numerical Methods for the Analysis of Satellite Images and Data (Metodi Numerici per l'Analisi di Immagini e Dati da Satellite) (2005).
      Principal Investigator: Prof. M. Piana, University of Verona.
    • Inverse Problems in Astronomy: Models, Algorithms, Applications (Problemi Inversi in Astronomia: Modelli, Algoritmi, Applicazioni) (2004).
      Principal Investigator: Prof. M. Bertero, University of Genova.
    • Mathematical Software in Parallel and/or Distributed Computing Environments (Software Matematico in Ambienti di Calcolo Parallelo e/o Distribuito) (2003).
      Principal Investigator: Prof. L. Brugnano, University of Florence.
    • Analysis and Development of Parallel Computational Kernels for Systems Solution (Analisi e Sviluppo di Nuclei Computazionali Paralleli per la Soluzione di Sistemi) (2002).
      Principal Investigator: Prof. E. Galligani, University of Modena.
    • Iterative Methods for Nonlinear Equations Systems and Finite-Dimensional Optimization Problems (Metodi Iterativi per Sistemi di Equazioni Non Lineari e Problemi di Ottimizzazione di Dimensione Finita) (2000/01).
      Principal Investigator: Prof. V. Ruggiero, University of Ferrara.

  • National CNR projects

    • Large-Scale Systems and Parallel Computing (Sistemi di Grandi Dimensioni e Calcolo Parallelo) (1998/99)
      Principal Investigator: Prof. V. Ruggiero, University of Ferrara.


Collaborations with international journals

Ph.D. Students and international cooperations

  • Tatiana Alessanda Bubba
    Project (2013): Inverse problems regularization with shearlets: theory and applications
    Visiting at the Department of Mathematics, University of Houston (USA).
    Inviting: Prof. Demetrio Labate.
    Period: March 29 - July 27, 2014.

  • Matteo Boccafoli
    Ph.D. thesis (2013): Assistenza sanitaria a domicilio: problemi multi-obiettivo d'instradamento di veicoli e fidelizzazione paziente-infermiera. (Home health care: multi-objective problems of both vehicle routing and patient-nurse retention)
    Project: Multi-objective vehicles-routing problems for the optimization of both load balancing and customer loyalty
    Co-supervisor: Prof. Maddalena Nonato, University of Ferrara, Italy.
    Visiting at the CIRRELT - Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation (Montreal, Canada)
    Inviting: Prof. Gilbert Laporte. Period: September 2011 - March 2012

  • Ambra Giovannini, Ph.D.
    Ph.D. thesis (2011): Parallel Large-Scale Edge-Preserving Joint Inversion with PETSc and TAO
    Awarded as the best Ph.D. thesis in Computer Science 2011, University of Ferrara.
    Visiting at the Argonne National Laboratory, Mathematics and Computer Science Division, Chicago (USA)
    Inviting: Prof. Hong Zhang, Dept. of Computer Science, Illinois Institute of Technology, Chicago (USA). Period: September - December 2009.
    Studentship: Erasmus project, 2009; INdAM-GNCS project, 2009.
    Current position: research staff, High-Performance Computing Department, CINECA Supercomputing Center (Bologna, Italy)

  • Girolama Notarangelo, Ph.D.
    Ph.D. thesis (2011): Stochastics Modelling and Numerics for the Immune Human Response
    Co-supervisor: Margherita Carletti, University of Urbino, Italy.
    Visiting at the Heriot-Watt University, Department of Mathematics, Edinburgh (UK).
    Inviting: Prof. Evelyn Buckwar. Period: September - December 2009.
    Studentship: University of Ferrara, 2009; INdAM-GNCS project, 2010.
    Current position: post-doc fellow, Istitut für Stochastik, Johannes Kepler Universität (Linz, Austria)

  • Elisa Stevanato, Ph.D.
    Ph.D. thesis (2010): The Bi-Objective Travelling Salesman Problem with Profits and its Connections to Computer Networks
    Co-supervisor: prof. Carlo Filippi, University of Brescia, Italy.
    Awarded as the best Ph.D. thesis in Computer Science 2010, University of Ferrara.
    Visiting at the University of "La Laguna", La Laguna, Tenerife (Spain), Department of Statistics, Operational Research and Computer Science
    Inviting: Prof. Juan-José Salazar-González. Periods: April - June 2009, October 2009.
    Current position: research fellow, Department of Economy, University of Brescia (Italy).


International commitments

  • External examiner for the final dissertation of the Ph.D. thesis at the School of Mathematics, The University of Edinburgh (UK).
    Candidate: Dr. Kristian Woodsend. Thesis title: Using Interior Point Methods in Large-Scale Support Vector Machine Training.
    Supervisor: Prof. Jacek Gondzio.
    Date: October 2, 2009.

Teaching Activity


Interests:
Numerical Analysis; numerical methods for Nonlinear Optimization; Parallel Computing; mathematical methods in Computer Science; Object-Oriented Programming.

Courses taught

  • Department of Mathematics and Computer Science, University of Ferrara:

    2013/14
    Numerical Optimization Methods, M.Sc. degree in Mathematics
    2012/13
    Programming Languages and Laboratory, B.Sc. degree in Computer Science.
    2012/13 -2013/14
    Numerical Approximation Methods, M.Sc. degree in Mathematics

  • Faculty of Mathematics, Physics and Natural Sciences, University of Ferrara:

    2011/12
    Programming Languages and Laboratory, B.Sc. degree in Computer Science.
    2009/10 - 2011/2012
    Numerical Approximation Methods, M.Sc. degree in Mathematics
    2006/07 - 2010/2011
    Methods for Math. Programming, Planning and Control, M.Sc. degree in Computer Science and M.Sc. degree in Mathematics
    2003/04
    Numerical Optimization, M.Sc. degree in Computer Science and M.Sc. degree in Mathematics
    2001/02 - 2006/07
    Programming Languages II: Object-Oriented Programming, B.Sc. degree in Computer Science
    2000/01
    Computer Programming, part I and part II, Computer Science diploma
    1999/2000
    Computational Mathematics, part I: Optimization Methods, Computer Science diploma
    1994/1995
    Graph theory for Discrete Mathematics, Computer Science diploma

  • Faculty of Engineering, University of Ferrara:

    2003/04 - 2009/10
    Numerical Analysis, M.Sc. degree in Mechanics Eng., Materials Eng, Civil Eng. and Environmental Eng.
    1995/96 - 1998/99
    Applied Mathematics, full M.Sc. degree in Electronic Eng.
    1991/92
    Calculus, full M.Sc. degree in Electronic Eng.

  • faculty of Pharmacy, University of Ferrara:

    1999/2000 - 2003/04
    All lectures on
    Mathematics, Statistics, Probability and basics of Computer Science     for all the full M.Sc. degree in Pharmacy, Chemistry and Pharmaceutical Technology (CTF), Pharmaceutical Biotechnologies, Science and Technology of Cosmetic Products (STP).
    2000/01 - 2003/04
    On-line course of Mathematics, Statistics and Probability for the full M.Sc. degree in Pharmacy.


Ph.D. lectures:

Jun. 2013
Support Vector Machines for binary classi cation: a special algorithm, course "Knowledge Discovery in Life Sciences", IUSS, University of Ferrara, for the Ph.D. Schools in Mathematics, Computer Science, Physics, Chemistry and Engineering.

Feb. 2009
An introduction to kernel methods and their applications in pattern recognition, course "Methodologies of Pattern Recognition & Imaging", for the Ph.D. Schools in Mathematics, Computer Science, Physics, Chemistry and Engineering.

Feb. 2007
An Overview of Parallel Software for High-Performance Scientific Computing for the Ph.D. School in High Mechanics and Automotive Design & Technology, University of Modena and Reggio Emilia


Other teaching-related activities:

2010/11 to date
Referent teacher for the B.Sc. degree in Computer Science, University of Ferrara.

2002/03 - 2009/10
Referent teacher for the M.Sc. degree in Computer Science, University of Ferrara.

2004/05 - 2005/06
President of the Evaluation Committee of the M.Sc. degree in Computer Science, University of Ferrara.

2001/02 - 2004/05
Member of the Student's Carrier Evaluation Committee of the M.Sc. degree in Computer Science, University of Ferrara.

2005/06 to date
Member of the Library Committee of the Department of Mathematics, University of Ferrara.

2005/06 to date
Member of the Board for the Ph.D. School in Mathematics and Computer Science, Department of Mathematics, University of Ferrara.

2005/2006 - 2008/2009
Responsible for the tutoring project "Laboratorio di MATLAB per l'Analisi Numerica" for the M.Sc. degrees in Mechanics Eng., Materials Eng, Civil Eng. and Environmental Eng.


Other activities:

1999 to date
Responsible of the Computer Laboratories of the Computer Science degrees of the University of Ferrara.

2000 - 2005
Responsible of the Computer Laboratories of Department of Mathematics, University of Ferrara.