Pierre Alquier

There

bsky.app/profile/pier...

I agree with this. And it seems related to a question I had on my mind (even though I was not able to ask it very accurately)

Yellow bittern @ Jurong lake garden #birds

Recording from my talk

“𝙳𝚢𝚗𝚊𝚖𝚒𝚌 𝚁𝚊𝚗𝚍𝚘𝚖 𝙿𝚊𝚛𝚝𝚒𝚝𝚒𝚘𝚗𝚜: 𝙰𝚙𝚙𝚕𝚒𝚌𝚊𝚝𝚒𝚘𝚗𝚜, 𝙾𝚙𝚙𝚘𝚛𝚝𝚞𝚗𝚒𝚝𝚒𝚎𝚜, 𝚊𝚗𝚍 𝙲𝚑𝚊𝚕𝚕𝚎𝚗𝚐𝚎𝚜”

at the ICERM Workshop: 𝐍𝐨𝐧𝐩𝐚𝐫𝐚𝐦𝐞𝐭𝐫𝐢𝐜 𝐁𝐚𝐲𝐞𝐬𝐢𝐚𝐧 𝐈𝐧𝐟𝐞𝐫𝐞𝐧𝐜𝐞 -𝐂𝐨𝐦𝐩𝐮𝐭𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐈𝐬𝐬𝐮𝐞𝐬

Video: icerm.brown.edu/video_archiv... #BayesianNonparametrics

On the importance of abstract, from @lateinteraction.bsky.social on X:

Patrice Jeener, le graveur de mathématiques [Documentaire Scientifique IHP] YouTube video by Institut Henri Poincaré

youtu.be/P0x9W9U5e9k?...

Patrice Jeener, artiste-graveur posant devant deux de ses dessins donnés au Palais de la découverte et installés sur le balcon des mathématiques. À l'occasion de l'inauguration de l'exposition permanente de ces dons, le 27 mai 2009.

Vue de 8 dessins de Patrice Jeener sur le balcon des mathématiques au Palais de la découverte (avant sa fermeture pour rénovation).

Nous avons appris la nouvelle du décès de Patrice Jeener, compagnon de longue date du Palais de la découverte et donateur de gravures. Artiste-graveur il y avait trouvé, lorsqu'il était étudiant, matière à transformer des équations en œuvres.
⬇️

Say, if I see this in code written by my students, it can mean two things: either they understood my weak joke, or they simply believed blindly everything I said.

In both cases, they deserve to pass 😇

One can argue itwas not an amazing joke 😅

That was the joke 🤗

In Python, if you don't set a seed yourself, it's automatically set using the current system time. Which is why in this case you get different results every time you run the code.

The code

rd.seed( a = 1 )
rd.seed( a = rd.randint(0,1000) )

will lead to a funny but constant seed.

But using

rd.seed( a = rd.randint(0,1000) )

(without fixing the seed first) will fix the seed randomly.

In order to get reproducible results, many people say we should fix the random seed. However, it is little known that it leads to randomness which is always the same, which is not so random. I found an easy patch, don't hesitate to use it:

import random as rd
rd.seed( a = rd.randint(0,1000) )

Haha. I mean, yes, we pay in SGD in Singapore, and that's really cool 😎

I liked this post, I wonder why 🐦 🤗

Today, I bought a can of iced woolong tea. I was supposed to pay 1.10 SGD but someone forgot 0.10 on the machine, so I only paid 1.00. I feel guilty about it.

If you are the one who forgot 0.10 in the machine, please contact me so that I can reimburse you.

Today, AI solved a problem that I was definitely not able to solve by myself: it helped me going through the AISTATS formatting checker and removing all problems. Some of them were non-trivial (I was not able to find solutions online). AI fixed them while providing pedagogical explanations.

Thanks.

I like these birds so much!

Couple of Javan pond herons @ Bishan park #birds

New short paper forthcoming in Statistics & Probability Letters:

An objective non-local prior for skew-symmetric models.

arxiv.org/abs/2603.08285

This paper develops a Moment-Objective Minimum-Discrepancy (MOOMIN) Prior for testing symmetry against skew-symmetric alternatives.

Learning Functions of Halfspaces We give an algorithm that learns arbitrary Boolean functions of $k$ arbitrary halfspaces over $\mathbb{R}^n$, in the challenging distribution-free Probably Approximately Correct (PAC) learning model, ...

Well, this seems like a big deal. arxiv.org/abs/2603.087...

"This is the first algorithm that can PAC learn even intersections of two halfspaces in time 2^o(n)."

Devenir professeur à l'ESSEC || ESSEC | ESSEC Africa 📢 ESSEC Business School – Africa Campus is hiring As part of the development of its academic programs, ESSEC Business School – Africa campus is recruiting a full-time Professor in Finance (PMP). We ...

Our campus in Morocco is hiring.

www.linkedin.com/posts/esseca...

Numerical Analysis School 2026 | yannig goude Ouverture des inscriptions pour l'édition 2026 de l’École d’été d’analyse numérique, organisée conjointement par le CEA, EDF et Inria. Cet événement se tiendra au CEA Liten (Bourget du lac) et portera...

I forward a post by Yannig Goude on LinkedIn about an exciting event.

www.linkedin.com/posts/yannig...

Excellent

Postdoctoral Research Associate in Theoretical Computer Science and Algorithms Full time, fixed term for two years (with the possibility of extension) Located on the Camperdown Campus, University of Sydney Base Salary Academic Level A $109,263 - $121,054 + 17% superannuation Abo...

On the postdoc job market? "As foretold" some time ago, the Sydney Algorithms and Computing Theory group at @sydneycompsci.bsky.social is hiring postdocs (2 years, renewable) in Theoretical Computer Science and Algorithms!

usyd.wd105.myworkdayjobs.com/en-GB/USYD_E... #TCSSky

Little heron @ Kallang river #birds

Didn't know that this one had a name, but it's a cute one. There's a nice "polynomial-style" proof (cf. SoS hierarchies etc.), a nice "martingale-style" proof, and probably some others too. Neat!

Chinese pond heron @ Bishan park #birds

I don't really understand why the statement requires numbers a_1 ... a_n ? The proof seems to be valid for any rv A in the interval [m,M] right? Am I missing something?

Streets of Ho Chi Minh city