exciting week[s]

Posted in Mountains, pictures, Running, Statistics with tags , , , , , , , , , , , , , , on June 27, 2017 by xi'an

The past week was quite exciting, despite the heat wave that hit Paris and kept me from sleeping and running! First, I made a two-day visit to Jean-Michel Marin in Montpellier, where we discussed the potential Peer Community In Computational Statistics (PCI Comput Stats) with the people behind PCI Evol Biol at INRA, Hopefully taking shape in the coming months! And went one evening through a few vineyards in Saint Christol with Jean-Michel and Arnaud. Including a long chat with the owner of Domaine Coste Moynier. [Whose domain includes the above parcel with views of Pic Saint-Loup.] And last but not least! some work planning about approximate MCMC.

On top of this, we submitted our paper on ABC with Wasserstein distances [to be arXived in an extended version in the coming weeks], our revised paper on ABC consistency thanks to highly constructive and comments from the editorial board, which induced a much improved version in my opinion, and we received a very positive return from JCGS for our paper on weak priors for mixtures! Next week should be exciting as well, with BNP 11 taking place in downtown Paris, at École Normale!!!

fiducial on a string

Posted in Books, pictures, Statistics, Travel, University life with tags , , , , , , on June 26, 2017 by xi'an

A very short note in arXiv today by Gunnar Taraldsen and Bo Henry Lindqvist (NTU, Norway). With the above title. I find the note close to unreadable, I must say, as the notations are not all or well- defined. The problem starts from Teddy Seidenfeld [whom I met in Harvard around Dutch book arguments] arguing about the lack of unicity of fiducial distributions in a relatively simple setting. Actually the note is also inspired from Bayes, Fiducial and Frequentist, and comments from Teddy, a talk I apparently missed by taking a flight back home too early!

What I find surprising in this note is that the “fiducial on a string” is a conditional distribution on the parameter space restricted to a  curve, derived from the original fiducial distribution by a conditioning argument. Except that since the conditioning is on a set of measure zero, this conditional is not only not-unique, but it is completely undefined and arbitrary, since changing it does not modify the properties of the joint distribution.

Trump lies

Posted in pictures with tags , , , , , on June 25, 2017 by xi'an

“Trump achieved something remarkable: He said something untrue, in public, every day for the first 40 days of his presidency.” David Leonhardt

Today, I spotted in the NYT this interactive catalogue of lies uttered publicly by Trump since he became President.

the end of travel?

Posted in Books, Linux, Travel, University life with tags , , , , , , , on June 25, 2017 by xi'an

First came the bad news that travelling [by plane] with any electronic device larger than a small cell phone would be prohibited unless the device was checked in. I first thought this was a sort of weird protectionism for American airlines, since only Middle-East carriers seem to be impacted at the moment, but there are serious hints this could soon be extended to all flights to the USA. And to the UK. And likely to about every flight in a near future. That is fairly annoying and more, not only because it hugely reduces the ability to work on a plane [unless carrying paper printouts of every paper one wanted to read or review during a long flight, until this is as well considered as a fire hazard!], but mostly because quite likely these electronic devices in checked-in bags will be damaged or stolen. So the only convenient [least inconvenient] solution may be to stop travelling with laptops and operate on remote or cheap disposable machines from whenever one is. Since I presume even the smallest PCs like Raspberry Pi’s will be banned at some point. The only good news are for computer companies. And book sellers maybe.On top of this, I learned that the UK has [again!] copied the US in requiring visitors and even its own citizens to open devices to Border or Customs officers. And to disclose social media identities and associated passwords. A British Human Rights activist was recently arrested at London Heathrow for refusing to do so. And I read today that divulging those media IDs are now optional on the visa waiver program and possibly soon to be compulsory. Which sounds insane as a way to fight terrorism as it is obvious to set parallel accounts. And yet another travel nuisance.

At some point in the past, I had mused that we would soon be forced to travel with no personal item, maybe not even our own clothes, but a uniform provided by the airline. With luggages on a drone plane following by a safe margin. The invasion of privacy now contemplated and soon implemented by states that do not any longer seem concerned with Human Right goes way beyond this fantasy scenario. If travelling between countries means a massive reduction of one’s rights and dignity [which is already quite reduced under the current conditions], travelling may soon become a rare occurrence…

a trip back in time [and in Rouen]

Posted in Kids, pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , on June 24, 2017 by xi'an

On Monday, I took part in a celebration of the remarkable career of a former colleague of mine in Rouen, Gérard Grancher, who is retiring after a life-long position as CNRS engineer in the department of maths of the University of Rouen, a job title that tells very little about the numerous facets of his interactions with mathematics, from his handling of all informatics aspects in the laboratory to his support of all colleagues there, including fresh PhD students like me in 1985!, to his direction of the CNRS lab in 2006 and 2007 at a time of deep division and mistrust, to his numerous collaborations on statistical projects with local actors, to his Norman federalism in bringing the maths departments of Caen and Rouen into a regional federation, to an unceasing activism to promote maths in colleges and high schools and science fairs all around Normandy, to his contributions to professional training in statistics for CNRS agents, and much, much more… Which explains why the science auditorium of the University of Rouen was packed with mathematicians and high schools maths teachers and friends! (The poster of the day was made by Gérard’s accomplices in vulgarisation, Élise Janvresse and Thierry Delarue, based on a sample of points randomly drawn from Gérard’s picture, maybe using a determinantal process, and the construction of a travelling salesman path over those points.)

This was a great day with mostly vulgarisation talks (including one about Rasmus’ socks..!) and reminiscences about Gérard’s carreer at Rouen. As I had left the university in 2000 to move to Paris-Dauphine, this was a moving day as well, as I met with old friends I had not seen for ages, including our common PhD advisor, Jean-Pierre Raoult.

This trip back in time was also an opportunity to (re-)visit the beautifully preserved medieval centre of Rouen, with its wooden houses, Norman-style, the numerous churches, including Monet‘s cathedral, the Justice Hall… Last time I strolled those streets, George Casella was visiting!

Le Monde puzzle [#1013]

Posted in Books, Kids with tags , , , , , on June 23, 2017 by xi'an

A purely arithmetic Le Monde mathematical puzzle:

An operation þ applies to all pairs of natural integers with the properties

0 þ (a+1) = (0 þ a)+1, (a+1) þ (b+1)=(a þ b)+1, 271 þ 287 = 77777, 2018 þ 39 = 2018×39

Find the smallest integer d>287 such that there exists c<d leading to c þ d = c x d, the smallest integer f>2017 such that 2017 þ f = 2017×40. Is there any know integer f such that f þ 2017 = 40×2017?

The major appeal in this puzzle (where no R programming seems to help!) is that the “data” does not completely defines the operation  þ ! Indeed, when a<b, it is straightforward to deduce that a þ b = (0 þ 0)+b, hence solving the first two questions by deriving (0 þ 0)=270×287 [with d=2×287 and f=2017×40-270×287], but the opposed quantity b þ a is not defined, apart from (2018-39) þ 0. This however brings a resolution since

(2018-39) þ 0 = 2017×39 and (2018-39+2017) þ 2017 = 2017×39+2017 = 2017×40

leading to f=2018-39+2017=3996.

non-Bayesian decision riddle

Posted in Statistics with tags , , , on June 22, 2017 by xi'an

As a continuation of the Bayesian resolution of last week riddle, I looked at a numeric resolution of the four secretaries problem, while in the train back from Rouen (and trying to block the chatter of my neighbours, a nuisance I find myself more and more sensitive to!). The target function is defined as

gainz=function(b,c,T=1e4,type="raw"){
  x=matrix(runif(4*T),ncol=4)
  maz=t(apply(x,1,cummax))
  zam=t(apply(x[,4:1],1,cummax))
  if (type=="raw"){return(mean(
   ((x[,2]>b*x[,1])*x[,2]+
    (x[,2]<b*x[,1])*((x[,3]>c*maz[,2])*x[,3]+
        (x[,3]<c*maz[,2])*x[,4]))/maz[,4]))} 
  if (type=="global"){return(mean( 
    ((x[,2]>b*x[,1])*(x[,2]==maz[,4])+
     (x[,2]<b*x[,1])*((x[,3]>c*maz[,2])*(x[,3]==maz[,4])+
         (x[,3]<c*maz[,2])*(x[,4]==maz[,4])))))} 
  if (type=="remain"){return(mean( 
    ((x[,2]>b*x[,1])*(x[,2]==zam[,3])+
     (x[,2]<b*x[,1])*((x[,3]>c*maz[,2])*(x[,3]==zam[,2])+
          (x[,3]<c*maz[,2])*(x[,4]==zam[,2])))))}}

where the data is generated from a U(0,1) distribution as the loss functions are made scaled free by deciding to always sacrifice the first draw, x¹. This function is to be optimised in (b,c) and hence I used a plain vanilla simulated annealing R code:

avemale=function(T=3e4,type){
  b=c=.5
  maxtar=targe=gainz(b,c,T=1e4,type)
  temp=0.1
  for (t in 1:T){
    bp=b+runif(1,-temp,temp)
    cp=c+runif(1,-temp,temp)
    parge=(bp>0)*(cp>0)*gainz(bp,cp,T=1e4,type)
    if (parge>maxtar){
      b=bs=bp;c=cs=cp;maxtar=targe=parge}else{
    if (runif(1)<exp((parge-targe)/temp)){
      b=bp;c=cp;targe=parge}}
    temp=.9999*temp}
  return(list(bs=bs,cs=cs,max=maxtar))}

with outcomes

  • b=1, c=.5, and optimum 0.8 for the raw type
  • b=c=1 and optimum 0.45 for the global type
  • b undefined, c=2/3 and optimum 0.75 for the remain type