Tomorrow I am off to Venezia for three days, attending the ESOBE 2016 workshop, where ESOBE stands for European Seminar on Bayesian Econometrics. This year it is indeed taking place in Venezia, Università Ca’ Foscari, in this beautiful building on the Gran Canale, and I have been invited to give a talk. Excited to get back to this unique place, hoping the high water will not be too high to prevent getting around (at random as usual).
Archive for Italia
Today, I went to Milano for 13 hours to give a seminar at l’Università Bocconi. Where I thus gave a talk on Testing via mixtures (using the same slides as at ISBA last Spring). It was the first time I was in Milano (and thus at Bocconi) for more than a transfer to MCMski or to Pavia and it was great to walk through the city. And of course to meet and share with many friends there. While I glimpsed the end of the sunrise on the Italian Alps (near Monte Rosa?!), I was too late on my way back for the sunset.
I remember fondly the early Valencia meetings where we did not have to pick between sessions. Then one year there were two sessions and soon more. And we now have to pick among equally tantalising sessions. [Complaint of the super wealthy, I do realise.] After a morning trip to San’Antioco and the southern coast of Sardinia, I started my ISBA 2016 with an not [that Bayesian] high dimension session with Michael Jordan (who gave a talk related to his MCMski lecture), Isa Verdinelli and Larry Wasserman.
Larry gave a [non-Bayesian, what else?!] talk on the problem of data splitting versus double use of the same data. Or rather using a model index estimated from a given dataset to estimate the properties of the mean of the same data. As in model selection. While splitting the data avoids all sorts of problem, not splitting the data but using a different loss function could avoid the issue. (And the infinite regress that if we keep conducting inference, we may have to split further and further the data.) Namely, if we were looking only at quantities that do not vary across models. So it is surprising that prediction get affected by this.
In a second session around Bayesian tests and model choice, Sarah Filippi presented the Bayesian non-parametric test she devised with Chris Holmes, using Polya trees. And mentioned our testing-by-mixture approach as a valuable alternative! Veronika Rockova talked about her new approach to efficient variable selection by spike-and-slab priors, through a mix of particle MCMC and EM, plus some variational Bayes motivations. (She also mentioned extensions by repulsive sampling through the pinball sampler, of which her recent AISTATS paper reminded me.)
Later in the evening, I figured out that the poster sessions that make the ISBA/Valencia meetings so unique are alas out of reach for me as the level of noise and my reduced hearing capacities (!) make impossible any prolonged discussion on any serious notion. No poster session for ‘Og’s men!, then, even though I can hang out at the fringe and chat with friends!
Just to remind participants to AISTATS 2016 and to ISBA 2016 that the deadlines for early registration are January 31 and February 10, getting close. Since both fees are quite high, it certainly makes sense to take advantage of those deadlines (and to make all travel reservations). (While I did try to see the fees of AISTATS 2016 set to a lower value, half of the fees are paying for coffee breaks and the banquet and the welcome party and were not negotiable. As my suggestion of cancelling the banquet was not accepted either! At least, the offer of accommodations in Cadiz is reasonable, from the list of hotels on the website, to a large collection of airbnb listings [minus the one I just reserved!]. And both Spain and Italy set an heavy 20% tax on conferences… Warning: the AISTATS 2016 do not cover the shuttle bus transfer from Sevilla, the major airport in the vicinity.)