[Forum SIS] Seminari SOYER e POLSON al CNR IMATI - Milano - 3 Maggio
Antonio Pievatolo
antonio.pievatolo a mi.imati.cnr.it
Mer 2 Maggio 2018 11:55:11 CEST
Ricordo i seguenti seminari che si svolgeranno presso
CNR IMATI, via Alfonso Corti 12, Milano, Aula A
Giovedì 3 maggio 2018, ore 15:00 - 16:30
[notare il cambio di aula; gentilmente scrivere a simona at mi.imati.cnr.it
se si intende partecipare]
1) Refik Soyer, School of Business, George Washington University ore
15:00-15:45
BAYESIAN MODELING OF NON GAUSSIAN MULTIVARIATE TIME SERIES
Abstract: Modeling of multivariate non Gaussian time series of correlated
observations is considered. In so doing, we focus on time series from
multivariate counts and durations.
Dependence among series arises as a result of sharing a common dynamic
environment. We discuss characteristics of the resulting multivariate time
series models and develop Bayesian inference for them using particle
filtering and Markov chain Monte Carlo methods.
We illustrate application of the proposed approach using conditionally
multivariate Poisson and gamma time series.
Joint work with Tevfik Aktekin, University of New Hampshire and Nicholas
Polson, University of Chicago
2) Nicholas Polson, Booth School of Business, University of Chicago ore
15:45-16:30
DEEP LEARNING: A BAYESIAN PERSPECTIVE
Deep learning is a form of machine learning for nonlinear high dimensional
pattern matching and prediction. By taking a Bayesian probabilistic
perspective, we provide a number of insights into more efficient algorithms
for optimisation and hyper-parameter tuning. Traditional high-dimensional
data reduction techniques, such as principal component analysis (PCA),
partial least squares (PLS), reduced rank regression (RRR), projection
pursuit
regression (PPR) are all shown to be shallow learners. Their deep learning
counterparts exploit multiple deep layers of data reduction which provide
predictive performance gains. Stochastic gradient descent (SGD) training
optimisation and Dropout (DO) regularization provide estimation and variable
selection. Bayesian regularization is central to finding weights and
connections in networks to optimize the predictive bias-variance
trade-off. To illustrate our methodology, we provide an analysis of
international bookings on Airbnb. Finally, we conclude with directions
for future research.
Joint work with Vladimir Sokolov, Sistems Engineering and Operations
Research, George Mason University
Maggiori informazioni sulla lista
Sis