Lightnews — Scholar-powered news

Meet our second Platinum sponsor for the Paris event!
CodeGlass helps software teams look inside their running #JuliaLang applications, with profiler and time-travel analysis tools. Wanna trace performance and memory back to the lines of code that really matter? Check them out 👀

An advertisement for the software company CodeGlass: "bringing clarity to Julia's runtime"

1 6

Reposted by Guillaume Dalle

JuliaCon 2025 @juliacon.bsky.social · 15d

JuliaHub has sponsored every single JuliaCon since 2014, and this year is no different! Check out their JuliaCon Global 2025 Pittsburgh Sponsor Talk, and look out for them at JuliaCon Local Paris in a few weeks!

youtu.be/zCUQfE-MGgM

JuliaHub Sponsor Talk | Viral B. Shah | JuliaCon Global 2025

YouTube video by The Julia Programming Language

youtu.be

1 2

Guillaume Dalle @gdalle.bsky.social · 16d

@sophie.huiberts.me will be asked to consult on the scientific aspects

1 3

Guillaume Dalle @gdalle.bsky.social · 16d

Don't worry, I can still write "Dantzig Queen" about linear programming

1 3

Guillaume Dalle @gdalle.bsky.social · 16d

The geometric interpretation I linked to on Wikipedia is quite nice, it relies on the hyperplane separation theorem which is immediately intuitive to anyone who has met a convex set.

But I don't think it beats the death metal song linked above
bsky.app/profile/adam...

Sebastian Karcher @adam42smith.bsky.social · 16d

Do I have news for you: www.youtube.com/watch?v=Muu3...

AARDVARKS – Farkas’ Lemma

YouTube video by AARDVARKS Official

www.youtube.com

2

Guillaume Dalle @gdalle.bsky.social · 16d

This is incredible. Honestly, compared to that, all my creations are mainstream. Even the Zotero love song.

1

Guillaume Dalle @gdalle.bsky.social · 16d

After more years than I care to admit, I am pleased to announce that I finally know what the fuck Farkas' lemma was trying to tell me all along.
en.wikipedia.org/wiki/Farkas%...

Farkas' lemma - Wikipedia

en.wikipedia.org

2 9

Guillaume Dalle @gdalle.bsky.social · 16d

Speaking as a Parisian, despite our proverbial bad weather, it is quite rare to see clouds hovering that low

1 1

Reposted by Guillaume Dalle

Pierre Alquier @pierrealquier.bsky.social · 18d

This is not OK.

I don't submit often to NeurIPS, but I reviewed papers for this conference almost every year. As a reviewer, why would I spend time trying to give a fair opinion on papers if it's what happens in the end???

Noa Garcia @noagarciad.bsky.social · 20d

This is the metareview

3 11 51

Guillaume Dalle @gdalle.bsky.social · 19d

But does it have cool songs about it too?
www.youtube.com/watch?v=rHnRs54hv4o

Compile (The \LaTeX Song)

YouTube video by Joe Breen Math

www.youtube.com

1

Guillaume Dalle @gdalle.bsky.social · 19d

As someone who has organized two (very minor) conferences this year, I concede this would be a major headache, and I commend the chairs of satellite events for taking on the challenge. But then again, the sheer scale of NeurIPS already stretches the physical boundaries of "feasibility".

1 1

Guillaume Dalle @gdalle.bsky.social · 19d

My dream is that instead of having 3 or 4 interchangeable and domain-agnostic conferences (NeurIPS, ICML, ICLR, AAAI), we could have a handful of domain-specific venues (some are starting to emerge like LoG or RLC).

1 1

Guillaume Dalle @gdalle.bsky.social · 19d

That we agree on.

2

Reposted by Guillaume Dalle

Mathieu Besançon @matbesancon.bsky.social · 20d

@neuripsconf.bsky.social decisions and stats are out. Definitely sad that a great initiative was set up but NOT what it should be: a third option to avoid 5000 authors burning kerosene crossing the world to present 8 pages

1 1 10

Guillaume Dalle @gdalle.bsky.social · 19d

And rejecting papers for space reasons while a big alternative venue is not fully recognized seems a bit counterproductive

1 2

Reposted by Guillaume Dalle

EurIPS Conference @euripsconf.bsky.social · 19d

It may only be a band-aid, but we have just announced our new "Salon des Refusés" sessions for papers rejected due to space constraints: bsky.app/profile/euri...

EurIPS Conference @euripsconf.bsky.social · 19d

Congratulations to everyone who got their @neuripsconf.bsky.social papers accepted 🎉🎉🎉

At #EurIPS we are looking forward to welcoming presentations of all accepted NeurIPS papers, including a new “Salon des Refusés” track for papers which were rejected due to space constraints!

4 5 14

Guillaume Dalle @gdalle.bsky.social · 19d

That’s definitely true, and I might not have quoted the exact right post for my point. Still, breaking up the big conferences into topic-specific areas might also mitigate the relentless resubmission lottery? Part of the problem is that each new paper gives rise to more than one review cycle

1 1

Guillaume Dalle @gdalle.bsky.social · 19d

Awesome idea!

1