Clustering is one of the most important unsupervised learning problems and it deals with finding a structure in a collection of unlabeled data; however, different clustering algorithms applied to the same data-set produce different solutions. In many applications the problem of multiple solutions becomes crucial and providing a limited group of good clusterings is often more desirable than a single solution. In this work we propose the Least Square Consensus clustering that allows a user to extrapolate a small number of different clustering solutions from an initial (large) set of solutions obtained by applying any clustering algorithm to a given data-set. Two different implementations are presented. In both cases, each consensus is accomplished with a measure of quality defined in terms of Least Square error and a graphical visualization is provided in order to make immediately interpretable the result. Numerical experiments are carried out on both synthetic and real data-sets. © 2010 Springer-Verlag.

Preface - MASCOT series of International Workshops has been organized every year, since 2001, under the auspices of the International Association for Mathematics and Computers in Simulation (IMACS) and the International Society of Grid Generation (ISGG). The Istituto per le Applicazioni del Calcolo} (IAC) of the CNR (Consiglio Nazionale delle Ricerche) has been, from the beginning, co-sponsor and main organizer of these events. MASCOT09, the 9th Meeting on Applied Scientific Computing and Tools: Grid Generation, Approximation and Visualization, was held in Rome, Italy, October 28-30, 2009, and hosted its animated group of young and senior participants at the CNR Head Office, in Piazzale A. Moro 7. Included as MASCOT09 co-sponsors were SIMAI- Societa' Italiana per la Matematica Applicata e Industriale and the Italian GNCS- Gruppo Nazionale per il Calcolo Scientifico. IAC, within the CNR RSTL (Ricerca Spontanea a Tema Libero) research project Numerical Grid Generation for Applied Scientific Computing, and MASCOT09 promoted the organization of the Guest Special Session " Together for Research: from the University of L'Aquila Onwards" by opening five positions for supported participation of young lecturers. This Special Session was designed to stress the disaster caused on the area of L'Aquila by the earthquake on April 6, 2009, at 3:32 am. In particular to pay attention to its effects on research surviving at the University of L'Aquila, and highlight the difficulties of the student community. As session focus Prof. Elisabetta Santi, from that University, gave a touching talk documenting the historical importance and the present large ruins of her town. Then five young fellows, from international submission, illustrated their scientific results. Even this workshop edition included a large spectrum of topics of the applied scientific computing, dealing with numerical methods and tools for a variety of aspects of modeling, approximation and simulation of crucially important application fields. Workshop presentations showed advanced studies in approximation and in numerical grid generation, developments of efficient solvers for partial differential equations and effective software tools, even for scientific visualization. This Book of Proceedings collects papers illustrating scientific backgrounds, research approaches and results presented at the Workshop, along with a few from the Special Session. Senior and junior expertise matured in several countries, both in west and east Europe, and overseas appears, coming from high qualified universities and research centers, in particular Italian ones, even from the University of L'Aquila. We would like to thank all who gave scientific, technical and organizational support to ensure the success of this edition, MASCOT09. We thank all the participants, senior and young colleagues, who were interested and planned to participate and all actively and warmly made the meeting successful Francesca Pistella Rosa Maria Spitaleri MASCOT09 Chairs- Scientific and Organizing Committees

About 50 % of the glaciated area outside the large ice sheets is located in the Arctic, and they contribute about 30 % to the runoff. IPCC predicts that the largest contribution to global sea level rise will stem from glaciers and ice caps. However, the uncertainties are large, up to 50%. Present estimates of mass balance of Svalbard glaciers are scarce and vary from close to balance to significantly negative. The dynamic response of the glaciers varies on the different glacier types: 1) Ice caps 2) tidewater glaciers and 3) glaciers ending on land. We are interested in estimating how changes in climate can affect the future mass balance of Svalbard glaciers and, consequently, the contribution to sea-level rise. This will be accomplished through modelling of ice flow, including calving fluxes, supported by field data. The proposed modelling work also includes Regional Climate modelling providing surface mass balance estimates for the whole of Svalbard, surface mass balance modelling for targeted glaciers and sensitivity analysis by different approaches using degree-day and energy balance models as well as couplings to atmosphere, hydrology and dynamics. A warmer climate may change both surface processes (snow accumulation, internal refreezing, superimposed ice and ablation) and dynamics. Predictions of future mass balance and dynamic response require boundary information about the thermal structure of the ice, the present and past surface mass balance, meteorological data/atmospheric field studies (AWS), surface and bed topography and current flow. These points will be the main focus for the field and remote sensing investigations. Remote sensing data is the only way to get enough spatial data, but must be validated by field data. We propose to address the above questions in a set of complementary field, remote sensing and modelling programs.

We present a comparison of three methods for the solution of the magnetoencephalography inverse problem. The methods are: a linearly constrained minimum variance beamformer, an algorithm implementing multiple signal classification with recursively applied projection and a particle filter for Bayesian tracking. Synthetic data with neurophysiological significance are analyzed by the three methods to recover position, orientation and amplitude of the active sources. Finally, a real data set evoked by a simple auditory stimulus is considered.

IL TALENTO MATEMATICO DI MAURO PICONE Nella puntata n. 33 in onda mercoledì 28 aprile 2010 Per "STORIA DELL'INFORMATICA ITALIANA" "L'artigliere e l'istituto di calcolo" L' Istituto per le Applicazioni del Calcolo, nato a Napoli nel 1927 e in seguito, per volontà di Marconi, trasferito a Roma, si prefiggeva l'obbiettivo di riunire un insieme di macchine di calcolo e di persone, capaci di utilizzarle al meglio, anche per svilupparne di nuove, al servizio della ricerca scientifica. Un'idea apparentemente semplice, ma in realtà assolutamente innovativa. Alla sua nascita infatti, l'Istituto di Calcolo italiano è il primo istituto di calcolo al mondo! Dietro quest'idea, una delle figure più significative della storia del calcolo e dell'informatica italiana: Mauro Picone.

We present a multiphase mathematical model for tumor growth which incorporates the remodeling of the extracellular matrix and describes the formation of fibrotic tissue by tumor cells. We also detail a full qualitative analysis of the spatially homogeneous problem, and study the equilibria of the system in order to characterize the conditions under which fibrosis may occur. © 2010 Elsevier Ltd.