A puzzling trend in French engineer school students is to contest the role of companies in their training or even the very training towards an engineering job, as illustrated by the above Le Monde article of last weekend. A former Polytechnique student has even published a book entitled Désertons, calling for deserting the job market. (Paradoxically on sale on Amazon!) These schools are indeed a French peculiarity in that they were initially created to train civil servants for the French State and later expanded to larger cohorts free to join public or private companies. They are radically distinct (and independent) from universities, in their entry mechanism (with a specific competition exam, requiring its own specific preparation public programme), in their training, and in their funding (three times higher per student and mostly issued from other ministries than the Ministry of Education, with minute fees if any). The State support extend to paying a salary to Polytechnique students during their studies, which should on principle be reimbursed when the students do not join a civil service or a public company although it is unclear if anyone has yet been asked to since 2000. (Not due to the 2000 bug!) I thus fail to get the logic of this contestation when the professional purpose of these schools is clear to everyone, when there exist alternatives in the (public) university system, and when the French State is supporting much more intensively this branch of the education system…
Archive for Telecom ParisTech
le beurre et l’argent du beurre (have your cake and eat it too)[nicht den Fünfer und das Weggli]
Posted in Books, Kids, University life with tags École Polytechnique, beurre, bug, cake, Centrale Supélec, classes préparatoires, Engineering, ENSAE, enterprises, French education ministry, French politics, grandes écoles, have your cake and eat it too, Le Monde, Mines de Paris, sponsorship, Telecom ParisTech on October 21, 2025 by xi'anCongrats, Dr. Marival!
Posted in Books, Statistics, University life with tags coauthors, importance Markov chains, importance MCMC, Institut Polytechnique de Paris, jury, Palaiseau, parallelisation, Paris-Saclay campus, PhD thesis, projection, sequential Monte Carlo, SMC, speed of convergence, Telecom ParisTech, thesis defence, variational approximations on March 26, 2025 by xi'an
adaptive equi-energy sampling
Posted in Statistics, University life with tags MCMC, Metropolis-Hastings algorithm. equi-energy sampler, Telecom ParisTech on July 14, 2014 by xi'anToday, I took part in the thesis defence of Amandine Shreck at Telecom-ParisTech. I had commented a while ago on the Langevin algorithm for discontinuous targets she developed with co-authors from that school towards variable selection. The thesis also contains material on the equi-energy sampler that is worth mentioning. The algorithm relates to the Wang-Landau algorithm last discussed here for the seminars of Pierre and Luke in Paris, last month. The algorithm aims at facilitating the moves around the target density by favouring moves from one energy level to the next. As explained to me by Pierre once again after his seminar, the division of the space according to the target values is a way to avoid creating artificial partitions over the sampling space. A sort of Lebesgue version of Monte Carlo integration. The energy bands
require the choice of a sequence of bounds on the density, values that are hardly available prior to the simulation of the target. The paper corresponding to this part of the thesis (and published in our special issue of TOMACS last year) thus considers the extension when the bounds are defined on the go, in a adaptive way. This could be achieved based on earlier simulations, using some quantiles of the observed values of the target but this is a costly solution which requires to keep an ordered sample of the density values. (Is it that costly?!) Thus the authors prefer to determine the energy levels in a cheaper adaptive manner. Namely, through a Robbins-Monro/stochastic approximation type update of the bounds,
My questions related with this part of the thesis were about the actual gain if any in computing time versus efficiency, the limitations in terms of curse of dimension and storage, the connections with the Wang-Landau algorithm and pseudo-marginal approximations, and the (degree of) likelihood of an universal and automatised adaptive equi-energy sampler.
