a journal of the plague, sword, and famine year

Read two successive books about seeking lost sisters, The Last House on Needless Street and Second Sister, after finishing the third book in a row involving a dead god, aptly named Three parts dead. This third one was rather enjoyable, thanks to the world construction, except for a blah ending. The first one, by Catriona Ward, is perplexing, complex and frankly a bit stretched in its gradual exposition of a multiple personality (disorder) patient. The “horror” side never really set for me, which is fine as it never does. Furthermore, this is the very first book I ever read where I saw a few words (correctly) written in Breton, as well as a thread with the Breton myth of ar Ankou, the local Death personification. Kudos for that! The second one, a physical book that I picked rather instinctively / hurriedly in a Barnes & Noble in Philadelphia is a thriller set in Hong Kong. Despite a bit too much of infodump on internet (in)security and hacking, and some caricaturesque sides, incl. the final coup de théâtre!, I enjoyed it as a page-turner. (But I now wonder if I am not getting prejudiced against Kindle books..!) Except for the anti-protest paragraph. Also read a nice BD, Les Animaux Dénaturés, borrowed from Andrew, which is an adaptation the 1952 book by Vercors, that I saw eons ago as a theatre play. The interrogation on what constitutes humanity (vs. simianity) is the driving force of the story, but it is somewhat marred by the killing of a newborn child that seems to negate the whole fight of the main characters.

Thanks to a short (train) visit to Coventry, I stayed overnight in the center of the city and enjoyed a fabulous dinner with friends at Jinseon Korean BBQ Restaurant, recently reviewed by Jay Rayner in The Guardian. Marinated thin slices of beef, pork, and lamb almost immediately cooked on the white hot (ring) coals, along rice and plenty of kimchi and hot sauce. And a sip of soju. Not an everyday fare, for sure, but quite delightful (and even more as my single true meal over two days!)

Watched a fraction of Swedish Black Crab, with Naomi Rapace playing the central character, but despite potential connections with the current survival war of Ukraine against the Russian terror, I quickly lost interest in the very shallow plot and in the accumulation of unrealistic scenes and heavily programmed eliminations of the characters (sorry for the spoiler!). For one thing, expert skaters skating 100km should not take days to cover the distance. For another, a military commando operating in the far North should wear appropriate clothes, not a sweater and a loose scarf!  Luckily enough, I have had no screen nearby [me] to distract me on my round trip flight to NYC from reviewing Biometrika submissions. (The flight back to Paris amazingly took less than 6 hours, thanks to extremely strong tail winds.)

Winter workshop, Gainesville (day 2)

On day #2, besides my talk on “empirical Bayes” (ABCel) computation (mostly recycled from Varanasi, photos included), Christophe Andrieu gave a talk on exact approximations, using unbiased estimators of the likelihood and characterising estimators garanteeing geometric convergence (bounded weights, essentially, which is a condition popping out again and again in the Monte Carlo literature). Then Art Owen (father of empirical likelihood among other things!) spoke about QMC for MCMC, a topic that always intringued me.

Indeed, while I see the point of using QMC for specific integration problems, I am more uncertain about its relevance for statistics as a simulation device. Having points uniformly distributed over the unit hypercube in a much more efficient way than a random sample is not helping much when only a tiny region of the unit hypercube, namely the one where the likelihood concentrates, matters. (In other words, we are rarely interested in the uniform distribution over the unit hypercube: we instead want to simulate from a highly irregular and definitely concentrated distribution.) I have the same reservation about the applicability of stratified sampling: the strata have to be constructed in relation with the target distribution. The method Art advocates using a CUD (completely uniformly distributed) sequence as the underlying (deterministic) pseudo-unifom sequence. Highly interesting and I want to read the paper in greater details, but the fact that most simulation steps use a random number of uniforms seems detrimental to the performances of the method in general.

After a lunch break at a terrific BBQ place, with a stop at Lake Alice to watch the alligator(s) I had missed during my morning run, I was able this time to attend till the end Xiao-Li Meng’s talk, where he presented new improvements on bridge sampling based on location-scale (or warping) transforms of the original two-samples to make them share mean and variance. Hani Doss concluded the meeting with a talk on the computation of Bayes factors when using (non-parametric) Dirichlet mixture priors, whose resolution does not require simulations for each value of the scale parameter of the Dirichlet prior, thanks to a Radon-Nykodim derivative representation. (Which nicely connected with Art’s talk in that the latter mentioned therein that most simulation methods are actually based on Riemann integration rather than Lebesgue integration. Hani’s representation is not, with nested sampling being another example.)

We ended up the day with a(nother) barbecue outside, under the stars, in the peace and quiet of a local wood, with wine and laughs, just like George would have concluded the workshop. This was a fitting ending to a meeting dedicated to his memory…