Wednesday 13/6/2018 at 2:30 p.m.
Prof. Stefano Favaro
Protection against disclosure is a legal and ethical obligation for statistical agencies releasing microdata files for public use. Given a cross classification of sample records by categorical key variables, any decision about release is supported by measures of disclosure risk, the most common being the number τ of sample uniques cells that are also population uniques. In this paper we depart from the dominant literature that infers τ by modeling association among key variables, and we consider modeling directly sample records. We develop a novel nonparametric Bayesian approach under the minimal assumption of a generalized Dirichlet prior for the random partition induced by the cross-classified sample records. This allows to derive an explicit, and simple, expression for the posterior distribution of τ, as well as a large sample Binomial approximation of it. Such a closed-form results, combined with an estimator for prior parameters designed in such a way to recognizes a primary role of small cells, make inference on τ exact, of easy implementation, computationally efficient and scalable to massive datasets. The proposed approach is tested on benchmark data from the U.S. 2000 census for the state of California, showing the same good performance of recent semiparametric Bayesian models for key variables.

Presso: Sala Pentagonale II Piano Via Bassini 12 20133 Milano

Martedì 5 Giugno 2018 presso la sala conferenze dell’IMATI-CNR di Pavia, verranno tenuti *due* seminari nell'ambito del Seminario di Matematica Applicata (IMATI-CNR e Dipartimento di Matematica, Pavia), http://matematica.unipv.it/it/seminari-matematica-applicata

** Alle ore 15.00 precise, il

Dr. Giorgio Saracco, Università degli Studi di Pavia

terrà un seminario dal titolo:

The Cheeger problem and an application to the (constant) Prescribed Mean Curvature problem

** Alle ore 16.00 precise, il

Dr. Stefano Almi, TUM Monaco

terrà un seminario dal titolo:

Energy release rate and stress intensity factors in planar elasticity in presence of smooth cracks

Giovedì 7 Giugno 2018, alle ore 15 precise, presso la sala conferenze dell’IMATI-CNR di Pavia, il

Dr. Robert Lasarzik, WIAS Berlino

terrà un seminario dal titolo:

Generalized solution concepts to the Ericksen-Leslie equations modeling liquid crystal flow

nell'ambito del Seminario di Matematica Applicata (IMATI-CNR e Dipartimento di Matematica, Pavia),

http://matematica.unipv.it/it/seminari-matematica-applicata

Al termine della conferenza sarà organizzato un piccolo rinfresco.

Giovedì, 31 Maggio 2018

Aula B, via Alfonso Corti 12, Milano, Italy

15:00-16:00

Non-homogeneous Inference Using Dynamic Bayesian Networks: Minimising the Impact of Dredging on Seagrass Ecosystems 

Il Dott. Wu è uno dei due relatori, insieme alla Prof.ssa Kerrie Mengersen (QUT), della scuola estiva ABS18 (Applied Bayesian Statistics) su

BAYESIAN STATISTICAL MODELLING AND ANALYSIS IN SPORT che si terrà la prossima settimana  a Como e per cui sono ancora

disponibili alcuni posti: http://www.mi.imati.cnr.it/conferences/abs18/index.html

 


It is challenging to predict the dynamic response of a complex system to stressors due to interdependencies and interactions between multiple system components under uncertainty. Potentially, the behaviour of the system itself can change over time as a result of exogeneous inputs and/or changes to the system state. For instance, an ecosystem that has already been subjected to stress may respond differently to further stresses, such as a reduced ability to resist and recover. Such changing dynamics are characteristic of non-homogeneous complex systems.

Dynamic Bayesian Networks (DBNs) provide an approach for predictive, whole-of-systems modelling of complex systems under uncertainty. Here, we discuss an approach to non-homogeneous inference with DBNs. The method enables dynamic updates of DBN parameters and inference of posterior marginal probabilities by propagating the effect of observations forwards in time. It also enables approximate inference for forwards-backwards inference. The approach is demonstrated on evaluating dredging impacts on seagrass ecosystems, with discussion on broader application to other domains.

Martedì 29 Maggio 2018, alle ore 15 precise, presso la sala conferenze dell’IMATI-CNR di Pavia, il

prof. Massimo Fornasier, TUM Monaco

terrà un seminario dal titolo:

Consistency of probability measure quantization by means of power repulsion-attraction potentials

nell'ambito del Seminario di Matematica Applicata (IMATI-CNR e Dipartimento di Matematica, Pavia),

http://matematica.unipv.it/it/seminari-matematica-applicata

Al termine della conferenza sarà organizzato un piccolo rinfresco.

Martedì 22 Maggio 2018, alle ore 15 precise, presso la sala conferenze dell’IMATI-CNR di Pavia, il

Dr. Antonio De Rosa, Courant Institute

terrà un seminario dal titolo:

Sharp regularity for weak notions of surfaces

nell'ambito del Seminario di Matematica Applicata (IMATI-CNR e Dipartimento di Matematica, Pavia),

http://matematica.unipv.it/it/seminari-matematica-applicata

Al termine della conferenza sarà organizzato un piccolo rinfresco.

Martedì 15 Maggio 2018, alle ore 15 precise, presso la sala conferenze dell’IMATI-CNR di Pavia, la

D.ssa Silvia Villa, Politecnico di Milano

terrà un seminario dal titolo:

Forward-backward algorithm for inverse problems and machine learning

nell'ambito del Seminario di Matematica Applicata (IMATI-CNR e Dipartimento di Matematica, Pavia),

Thursday, 3 May 2018, 15:00 - 16:30

Aula Expo, via Alfonso Corti 12, Milano, Italy

15:00-15:45
Bayesian Modeling of Non Gaussian Multivariate Time Series
(Refik Soyer, School of Business, George Washington University)

 

15:45-16:30
Deep Learning: a Bayesian Perspective
(Nicholas Polson, Booth School of Business, University of Chicago)