publish or perish [or move to .005]

Posted in Books, pictures, Statistics, University life with tags , , , , on October 24, 2017 by xi'an

A series of articles in the Sciences et Médecine part of Le Monde reproduced coverages found elsewhere on the debates running within the scientific community on improving the quality of scientific papers. Through reproducible experiments and conclusions. And on using new bounds for the p-value, the solution to all woes! The article borrows a lot from the Nature proposal [discussed quite a lot here in the past weeks] and does not provide particularly insightful views. It however contains a coverage (rightmost columns) on a peer community approach called PubPeer, which was launched by two neuroscientists, Brandon Stell and Boris Barbour, both at CNRS, towards sharing comments on published papers. Mostly to criticise the methodology used in these papers. Or to point out multiple usages of the same graphs. Or doctoring of pictures. In the vast majority of cases, the papers are in biology and the comments not addressed by the authors of the papers. (With this exception of a discussion of the Nature paper covering the call for new bounds on p-values. Nature paper that had the appealing feature of calling for an end to `one-size-fits-all’ thresholds.) Creating a platform for discussing papers from a journal is already hard enough (as shown with the closure of Series B’log!), hence running a global discussion forum for all journals sounds hard to manage and foster. By which I mean it is difficult to fathom the impact of the discussions on the published papers and the journals where they are published, given the reticence of said journals to engage into reassessments of published papers…


random matrix advances

Posted in pictures, Statistics with tags , , , , on October 23, 2017 by xi'an

Le Monde puzzle [open problem]

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

What should have been the last puzzle in Le Monde competition turned out to be an anticlimactic fizzle on how many yes-no questions are needed to identify an integer between 1 and 1025=2¹⁰+1 and an extension to replies possibly being lies

What is much more exciting is that voting puzzle #1021 got cancelled because the authors of this puzzle thought the cascading majority rule would produce the optimal solution and it does not! (As exhibited by my R code.) So here is an open problem to ponder about! (And another puzzle in the pipeline to complete the competition.)

postdocs positions in Uppsala in computational stats for machine learning

Posted in Kids, pictures, Statistics, Travel, University life with tags , , , , , , , , , , , on October 22, 2017 by xi'an

Lawrence Murray sent me a call for two postdoc positions in computational statistics and machine learning. In Uppsala, Sweden. With deadline November 17. Definitely attractive for a fresh PhD! Here are some of the contemplated themes:

(1) Developing efficient Bayesian inference algorithms for large-scale latent variable models in data rich scenarios.

(2) Finding ways of systematically combining different inference techniques, such as variational inference, sequential Monte Carlo, and deep inference networks, resulting in new methodology that can reap the benefits of these different approaches.

(3) Developing efficient black-box inference algorithms specifically targeted at inference in probabilistic programs. This line of research may include implementation of the new methods in the probabilistic programming language Birch, currently under development at the department.


Posted in Books, Kids, pictures, Travel with tags , , , , , , , on October 22, 2017 by xi'an

As our son is doing an internship in Luxembourg City this semester, we visited him last weekend and took the opportunity to visit the Museum of Modern Art (or MUDAM) there. The building itself is quite impressive, inserted in the walls of the 18th Century Fort Thüngen designed by Vauban, with a very luminous and airy building designed by Ming Pei. The main exhibit at the MUDAM is a coverage of the work on Su-Mei Tse, an artist from Luxembourg I did not know but whom vision I find both original and highly impressive, playing on scales and space, from atoms to planets… With connections to Monet’s nympheas. And an almost raw rendering of rock forms that I appreciate most particularly!

The bottom floor also contains an extensive display of the political drawings of Ad Reinhardt, who is more (?) famous for his black-on-black series…

end of my dream! [aka no free lunch]

Posted in pictures with tags , , , , , , , on October 21, 2017 by xi'an

In June, last June, I received a collection of nominal tickets for the National Theatre next door, Les Gémeaux, which animates a rich and diverse scene, attracting to my suburb spectators from all over Paris. Scene that I do not attend as much as I would like, except for a few contemporary dance shows now and then… As I had not ordered these tickets, I was rather pleasantly surprised and wondered at the source of this princely gift, ruling out close family and anonymous sponsors! And the size of it was way too much for a commercial teaser. The most likely explanation was thus a mistake send to the wrong X. Nonetheless, being at home for the first concert, I went to check whether or not my allotted seats were occupied. And unsurprisingly they were. (I did not push the cheek as far as engaging into a discussion with my homonym!) Missing an Austrian theme with Mozart and Haydn. (But I have regular tickets for ext weekend!)

Les Rouquins

Posted in pictures, University life, Wines with tags , , , , , , , on October 20, 2017 by xi'an