David Rousseau

Particle Physicist at @IJClab @ATLASexperiment at @CERN.
#AI and physics.
I have left this officially na$i platform, you know where, same handle.

[David Rousseau @dhpmrou]

Is uncertainties handling the ultimate roadblock for a wider adoption of AI in science ?
AI and the Uncertainty Challenge in Fundamental Physics, 27-Nov-1-Dec
with sessions on Uncertainty Quantification, Simulation Based Inference and more, https://t.co/OD2n5XphNE

[David Rousseau @dhpmrou]

Je viens de faire mon #eXit! L’exode de X est massif. Ne perdez pas un seul de vos abonnés. Grâce à #HelloQuitX j’ai inscrit 2265 nouveaux passagers pour partir vers #BlueSky & #Mastodon. Retrouvez automatiquement vos communautés le #20Janvier via https://t.co/7uQizPlefx !

[David Rousseau @dhpmrou]

@uberJJ @bengrossbg That’s good one. And easy to demonstrate.
x^x=e^(x ln x).
ln x loses against any power of x so lim x ln x when x going to 0 is 0.
Si lim x^ x when x goes to 0 is 1.

[David Rousseau @dhpmrou]

Just created the “boomers @ Neurips” group in #neurips Whova app. I had to.

[David Rousseau @dhpmrou]

@zack_overflow Bellard has also briefly held the record number of pi decimals computed. This was done using just a somewhat beefed up commercial PC.

[David Rousseau @dhpmrou]

@Rainmaker1973 He did not "save" them...
Ducks have always nested in the holes of large trees and have always managed very well on their own!

[David Rousseau @dhpmrou]

I’m moving to the other BS side, same handle

[David Rousseau @dhpmrou]

@EricNewcomer @MaximZiatdinov Because there is 1.5 teleoperator for each Waymo car on the road ?
Because Waymo’s revenue is dismal?

[Aishik Ghosh @Aishik_Ghosh_]

Career update 📢—I’m on the job market for faculty positions! My work focuses on advancing particle physics 💥 and astrophysics 🌌 by creating innovative, robust statistical methods with AI 🤖. If you know of any relevant openings, please reach out!

[Gregor Kasieczka @GregorKasieczka]

Today ends the #ML4Jets conference in Paris.

Great to see the progress in the field from one year to the next, with no saturation in sight. Below some @quunihh students & collaborators (and, of course, beautiful #Paris)

Many thanks to AnjaB & @dhpmrou et al for hosting this!

[David Rousseau @dhpmrou]

@balazskegl Wow !
Can it do an uncertainty loaded competition as well ?

https://t.co/9DWOYMKMPS

[David Rousseau @dhpmrou]

@KyleCranmer @3scorciav @soldni @JFPuget @SandesaraJay @glouppe @Aishik_Ghosh_ But it is hypothetical, we don’t know how to do this, right ?
At least with the soup you have a single model for inference instead of an ensemble.

[David Rousseau @dhpmrou]

@SandesaraJay @soldni @3scorciav @JFPuget @KyleCranmer @glouppe @Aishik_Ghosh_ Indeed, not a big effect on total resources, but still a significant simplification for model use.
Ah but then how do we study model stat uncertainty )

[David Rousseau @dhpmrou]

@3scorciav @soldni @JFPuget @KyleCranmer @SandesaraJay @glouppe @Aishik_Ghosh_ No no no, it has to be a single average model, not an average of models, that’s the whole point !

[David Rousseau @dhpmrou]

@soldni @3scorciav @JFPuget @KyleCranmer @SandesaraJay @glouppe @Aishik_Ghosh_ Asimov model, from Isaac Asimov's "Franchise" short story, where one citizen is picked to be the perfect average American, simplifying the convoluted election process considerably.

[David Rousseau @dhpmrou]

@soldni @3scorciav @JFPuget We physicists would call it an Asimov model.
@KyleCranmer @SandesaraJay @glouppe @Aishik_Ghosh_

[Jay Sandesara @SandesaraJay]

Excited to share the first results using Neural Simulation-Based Inference (NSBI) techniques applied to @ATLASexperiment data! We measure the elusive off-shell Higgs boson with 3.1x better observation sensitivity than standard (histogram) analysis techniques! A thread: (1/N)

[Claudius Krause @ClaudiusKrause]

I'm happy and proud to present the results of the #CaloChallenge!

https://t.co/zff4dDz8jH

A huge thank you to everyone who contributed to this effort in the past year(s)!
1/12

[Kyle Cranmer @KyleCranmer]

But making it happen in an actual experiment at the LHC took a long time. Check out the thread by @Aishik_Ghosh_ about his 6-year journey. And shoutout to @dhpmrou for being an awesome advisor that saw the potential for these techniques early on
https://t.co/Gop7V3agEN

[Aishik Ghosh @Aishik_Ghosh_]

A thread on quantum interference and the 6 year adventure it took me through

In experimental particle physics we sift out events from signal (eg. Higgs) processes vs the orders of magnitude more events from background processes…

[David Rousseau @dhpmrou]

@martinmbauer …Check the video…