Archive for flight

Ainsi sommes nous…

Posted in Travel with tags , , , , , , , , , on January 4, 2022 by xi'an

Yesterday in Cayenne, while boarding the Air France plane to Paris, we came across Christiane Taubira who was also taking this flight, and who wished us a happy new year. While my wish for 2022 is that she would be the presidential candidate for a united Left, I greatly fear this is unlikely to happen… Here are some paragraphs from a call to unity she wrote in Le Monde last week:

C’est nous qui pouvons traduire en politiques publiques des revendications dispersées et légitimes, portées par les féministes et les ouvriers, les professeurs et les employés, les gens des villes et ceux des cités, les communes rurales qui tiennent bon et celles qui se dépeuplent, toutes causes portées et illustrées autant par l’inventivité constatée dans les ronds-points, que par les images familières des syndicaux le poing levé, des banderoles et pancartes imaginatives des personnels soignants d’hôpitaux en burn-out, par l’impatience bruyante des pêcheurs exaspérés, l’arc de couleurs des cortèges LGBT, les impromptus des artistes et créateurs en tous genres, les fantaisies joyeuses et sérieuses des lycéens, la ronde et les performances de ces adolescents qui marchent pour la Terre, la biodiversité, le climat, la vie !

Ainsi sommes-nous, égaux et solidaires, libres, fiers et volontaires, divers dans nos apparences, nos enracinements, nos lignées, nos parcours, nos bagages culturels, liés par un destin collectif qui transcende nos péripéties personnelles, conscients que se trouve là notre insondable puits d’énergie, de vitalité et de créativité.

Et nous savons ce qui nous fait tenir ensemble. La République. Notre chose publique. Notre bien commun suprême. La République laïque qui nous protège de toute emprise et promet à chacune et chacun d’entre nous les conditions pour accomplir sa destinée choisie. C’est la haute mission de l’école que de pourvoir en savoirs et préparer en citoyenneté ; de susciter et nourrir l’esprit critique, celui-là même qui nous autorise à tout interroger, en premier lieu les dogmes et commandements, les assignations et les interdits, à pouvoir récuser au nom de la raison et des connaissances.

La République laïque protège autant notre liberté de conscience que notre liberté de culte, n’encombre notre citoyenneté d’aucune obligation de croyance ou non-croyance, mieux : par la loi pénale, elle nous protège individuellement de toute pression. Libres de devenir, libres, quoi que contredisent les circonstances réelles d’évitement et d’empêchement, expérimentées au quotidien par des millions d’entre nous, et qui font obstacle à cette promesse. Elle reste une promesse formelle et ferme, toujours à accomplir. Je sais les difficultés, les injustices et parfois la suspicion ou la morgue. C’est pourtant par-dessus qu’il faut forcer un chemin républicain.

three decades back

Posted in Kids, Travel with tags , , , , , , , , , on October 25, 2021 by xi'an

Yesterday, I (we) found myself (ourselves) back in time, precisely 34 years ago, as we drove our daughter to Orly airport for her flight to Cayenne, French Guiana. And the start of her internship. Indeed, this is also the airport from which I left for Purdue University in 1987 and where my parents drove me then, more for sentimental reasons than out of necessity, as I had much less luggage than Rachel! This old airport has not changed that much, apart from the sharp increase in security restrictions. At the time, my parents were able to stay till the plane to Chicago left (two hours late) and watch it take off from the terraces of the airport. This time, we were just unable to enter the airport beyond the parking lot and watched the plane take off (one hour late) from a flight tracker! But overall there was the same bittersweet feeling of seeing one’s kid move (far) away for a major and exciting step in their professional life. (When I called my mom to watch for the plane flying west straight across Normandy, very close to our family roots, she reminded me of that and also of my grand-parents watching my plane flying by… Without a flight tracker! I actually remember spotting Mont Saint-Michel on that trip.) Fare well, Dr. R, and see you soon!
~

back to W

Posted in pictures, Statistics, Travel, University life with tags , , , , , , , , , , , , on October 3, 2021 by xi'an

gone South [jatp]

Posted in Mountains, pictures, Statistics, Travel, University life, Wines with tags , , , , , , , , on March 27, 2021 by xi'an

flow contrastive estimation

Posted in Books, Statistics with tags , , , , , , , , on March 15, 2021 by xi'an

On the flight back from Montpellier, last week, I read a 2019 paper by Gao et al. revisiting the MLE estimation of a parametric family parameter when the normalising constant Z=Z(θ) is unknown. Via noise-contrastive estimation à la Guttman & Hyvärinnen (or à la Charlie Geyer). Treating the normalising constant Z as an extra parameter (as in Kong et al.) and the classification probability as an objective function and calling it a likelihood, which it is not in my opinion as (i) the allocation to the groups is not random and (ii) the original density of the actual observations does not appear in the so-called likelihood.

“When q appears on the right of KL-divergence [against p],  it is forced to cover most of the modes of p, When q appears on the left of KL-divergence, it tends to chase the major modes of p while ignoring the minor modes.”

The flow in the title indicates that the contrastive distribution q is estimated by a flow-based estimator, namely the transform of a basic noise distribution via easily invertible and differentiable transforms, for instance with lower triangular Jacobians. This flow is also estimated directly from the data but the authors complain this estimation is not good enough for noise contrastive estimation and suggest instead resorting to a GAN version where the classification log-probability is maximised in the model parameters and minimsed in the flow parameters. Except that I feel it misses the true likelihood part. In other words, why on Hyperion would estimating all θ, Z=Z(θ), and α at once improve the estimation of Z?

The other aspect that puzzles me is that (12) uses integrated classification probabilities (with the unknown Z as extra parameter), rather than conditioning on the data, Bayes-like. (The difference between (12) and GAN is that here the discriminator function is constrained.) Esp. when the first expectation is replaced with its empirical version.

%d bloggers like this: