Vivek Roy, Aixian Tan and James Flegal arXived a new paper, Estimating standard errors for importance sampling estimators with multiple Markov chains, where they obtain a central limit theorem and hence standard error estimates when using several MCMC chains to simulate from a mixture distribution as an importance sampling function. Just before I boarded my plane from Amsterdam to Calgary, which gave me the opportunity to read it completely (along with half a dozen other papers, since it is a long flight!) I first thought it was connecting to our AMIS algorithm (on which convergence Vivek spent a few frustrating weeks when he visited me at the end of his PhD), because of the mixture structure. This is actually altogether different, in that a mixture is made of unnormalised complex enough densities, to act as an importance sampler, and that, due to this complexity, the components can only be simulated via separate MCMC algorithms. Behind this characterisation lurks the challenging problem of estimating multiple normalising constants. The paper adopts the resolution by reverse logistic regression advocated in Charlie Geyer’s famous 1994 unpublished technical report. Beside the technical difficulties in establishing a CLT in this convoluted setup, the notion of mixing importance sampling and different Markov chains is quite appealing, especially in the domain of “tall” data and of splitting the likelihood in several or even many bits, since the mixture contains most of the information provided by the true posterior and can be corrected by an importance sampling step. In this very setting, I also think more adaptive schemes could be found to determine (estimate?!) the optimal weights of the mixture components.
Archive for the Mountains Category
By a great coincidence, I happened to be in Banff the same weekend as Salman Rushdie was giving a talk at the Banff Centre on his latest book! And got the news early enough to book a seat. The amphitheatre was unsurprisingly full and Salman Rushdie was interviewed by Eleanor Wachtel, first about the book and what led to its creation, especially the influence of his parents, and then second about his life and career, with an obvious focus on Khomeini’s fatwa. (The whole interview is podcasted on CBC.) Rushdie was witty and funny, even about the darkest moments, and discussed how in his youth no one would have imagined that religion would become such a central issue, defining and reducing people rather than being a part of them that would need no discussion. And how his family was de facto atheist, if not in words. The interview spent too little time on Rusdhie’s stand on freedom of expression, although he briefly spoke about the growing threats to this freedom, including those made in the name of religious freedom. (As we were reminded yesterday by the Warwick student union decision to bar Maryam Namazie from speaking on campus.) The experience was quite a treat, adding to the many bonuses of spending this weekend in Banff. Although I must admit I was fighting jetlag that late at night and hence must have dozed at points… (As an aside, I was rather surprised to see no security or police around the Banff Centre theatre, but of course this does not mean there was none. And this is Banff, not New York City or London.)
Well, the clue in the title should be obvious enough: A censorship board in New Zealand, the Film and Literature Board of Review, has just banned Ted Dawes’ “Into the River” from being sold or distributed or even exhibited. Following complaints orchestrated by Family First, a conservative organisation… The ban is actually temporary, until a committee reaches a decision about the classification of the book. The most surprising aspect of this story—besides the existence of a censoring institution in a democratic country in 2015!— is that the ban applies to everyone, including adult readers, and that all libraries had to take the book out of their shelves. Which is also fairly ridiculous in the era of e-books…
Yesterday, I flew from Amsterdam to Calgary to attend the Canadian Statistical Sciences Institute Leadership Retreat (15w2214) at the Banff International Research Station (BIRS). The point of this meeting is to brainstorm towards building a research policy and strategy for the newly created CANSSI. To which scientific advisory committee I joined last semester. I just hope my brain will remain functional enough to contribute to the discussion, if not to storm, despite the eight hour time lag! (The drive from Calgary to Banff was beautiful, with flashy yellow bursting from the green landscape: Fall is coming.)
I came upon the news by mere chance: the first Salomon Glencoe skyline mountain race took place last month, on Aug. 22, in one of my favourite mountaineering spots, the Valley of Glencoe. I have hiked and climbed in this valley six or seven times, and “bagged” seven of the nine local Munroes, mostly in Winter conditions. The race includes all of them in a 53km route with a 4,256m total ascent (and descent!), with a scramble of Buachaille Etive Mor via the classic Curved Ridge route and the west-to-east traverse of the Aonach Eagach. Absolutely awesome!
The winners completed the route in 7:36 for Joe Symmons and 7:44 for Emelie Forsberg, with the last runner finishing in 14 hours. I really wish I could enter this race but the organisers screen quite thoroughly entrances based on past experience, insisting on previous mountain races, and so this must remain a dream..!