Archive for Amsterdam

3-D bridge now opened!

Posted in pictures, Statistics, Travel with tags , , , , , , , on July 31, 2021 by xi'an

The 3D printed bridge I had mentioned in a previous blog was installed and officially opened over an Amsterdam canal, last month. Besides the company printing this steel bridge, over six months, researchers from the Turing Institute, including my friend Mark Girolami, developed a virtual version of the bridge using embedded sensors to monitor the behaviour and evolution of the real bridge over time and usage. Next time I am in the city, I’ll make sure to run bridge sampling!

the liberation [book review]

Posted in Books with tags , , , , , , , , , , on July 9, 2021 by xi'an

The third volume of Ian Tregillis’ The Alchemy Wars arrived in the mail, and I could not resist bing-read it, induced by a heat wave that made anything close to serious work near impossible in the late afternoons… The characters are essentially the same, with two central (human) female characters whose trajectories once again converge to the critical point. Plus, two female robots playing a contrapunt. And the biblical Daniel, reborn from slavery into a free willed, tolerant and pacific being.

“The Clockmakers had been playing a losing game of catch-up (…) They were too soft, too coddled, too accustomed to standing atop the pile. They weren’t well suited to life as underdogs. They were not French.”

The core of the action takes place in Amsterdam, occupied by liberated robots, prone to pogroms as well as re-enslaving other robots. The weakness in the plot is that there is no strong reason these robots do not completely take over the formerly ruling Guild of Alchemists, and lengthy plot-resolving discussions between fighting characters always irk me no end, but the conclusion still feels proper, with the author not at all reluctant to hack at bits and pieces of his character to raise the body-count. À ls George Martin! And the mild philosophical musing about the reversal of dominant-dominated positions in this society are overall enjoyable if not particularly deep. Overall, a striking trilogy.

a journal of the plague year [soon to turn one…]

Posted in Books, Kids, Mountains, pictures, Travel with tags , , , , , , , , , , , , , , , , , , , , , , , on March 6, 2021 by xi'an

Read in a few Sunday hours Living proof by John Harvey, a 1995 novel that I had found in the book exchange section of our library. A very easy read but rather enjoyable with several stories within stories and books within books. The resolution of the main murder mystery was disappointing but I enjoyed the inclusion of real artists like Ian Rankin and Mark Timlin. With a pastiche of P.D. James. And plenty of jazz references. Plus two characters meeting while studying at Warwick. And some shared glimpses of Nottingham like the statue of Robin Hood and the troglodyte Ye Olde Trip to Jerusalem. where I was once  invited for a pint… The story takes place during and around Nottingham’s Shots in the Dark festival. I hence grabbed another volume in the library in prevision of another lazy afternoon!

Baked rather decent chapattis for a take-home Bengali dinner but ruined the pan and started the fire alarm! Also tried to bake tortillas but mixed up the proportions of flour and water, ending up with a type of galette or injera instead (which worked as a container for the fajitas!).

Eventually watched the last two episodes of the Queen’s Gambit, but found them somewhat disappointing, between the main characters’ attitude that did not feel in tune with the 1960’s, the French femme fatale who cannot pronounce Jardins du Luxembourg, and the somewhat rosy tale of two orphans achieving financial freedom and professional success before their majority. Also watched the Korean Space Sweepers after an exhausting day, with a very shallow plot and a complete disregard for physics.

Read the duology of Six of Crows and Crooked Kingdom, by Leigh Bardugo, set in the same universe as the Grisha novels. Which I had read five years ago and somewhat forgotten than these novels were written as young adult books, with a resulting shallow plot, so full of sudden changes of fortune that any worry for the (caricaturesque) characters vanishes (till the one point when one should have) in a definitive suspension of suspense. (The Guardian reviewed Six of Crows in the Children’s book review section, which feels rather inappropriate given the degree of abuse the teens in the novels are submitted to, with two girls surviving sexual enslavement in the local brothels.) Just like the Grisha novels were set in a postcard version of Russia, these novels are taking place in a similarly thin (pannenkoek) version of Amsterdam (with waffles as the only culinary delicacy!). I do realise these series have a huge fan base, to the point of leading to an incoming Netflix series. But I found the more elaborate Ninth House much more enjoyable… (In tune with this series of reviews, the second book includes a plague episode with a modicum of realism, at least in its early stages.)

more EU, another 50 years! [jatp]

Posted in pictures, Travel with tags , , , , , , , on November 6, 2019 by xi'an

Gibbs clashes with importance sampling

Posted in pictures, Statistics with tags , , , , , on April 11, 2019 by xi'an

In an X validated question, an interesting proposal was made: at each (component-wise) step of a Gibbs sampler, replace simulation from the exact full conditional with simulation from an alternate density and weight the resulting simulation with a term made of a product of (a) the previous weight (b) the ratio of the true conditional over the substitute for the new value and (c) the inverse ratio for the earlier value of the same component. Which does not work for several reasons:

  1. the reweighting is doomed by its very propagation in that it keeps multiplying ratios of expectation one, which means an almost sure chance of degenerating;
  2. the weights are computed for a previous value that has not been generated from the same proposal and is anyway already properly weighted;
  3. due to the change in dimension produced by Gibbs, the actual target is the full conditional, which involves an intractable normalising constant;
  4. there is no guarantee for the weights to have finite variance, esp. when the proposal has thinner tails than the target.

as can be readily checked by a quick simulation experiment. The funny thing is that a proper importance weight can be constructed when envisioning  the sequence of Gibbs steps as a Metropolis proposal (in the dimension of the target). Sad enough, the person asking the question seems to have lost interest in the issue, a rather common occurrence on X validated!

%d bloggers like this: