Sophie Huiberts
@sophie.huiberts.me
1.1K followers 390 following 440 posts
CNRS researcher in linear programming
Posts Media Videos Starter Packs
sophie.huiberts.me
wait that actually makes sense. the next largest float64 number after 1e22 is like 1e6 higher, which is way more than 2pi. no use in talking about forward or backward error
Reposted by Sophie Huiberts
manoelhortaribeiro.bsky.social
Computer Science is no longer just about building systems or proving theorems--it's about observation and experiments.

In my latest blog post, I argue it’s time we had our own "Econometrics," a discipline devoted to empirical rigor.

doomscrollingbabel.manoel.xyz/p/the-missin...
sophie.huiberts.me
and 3: the Distinguisher in the forest tells us how to use an untrusted quantum computer
www.youtube.com
sophie.huiberts.me
Just learned that cryptographers are a lot more creative than the rest of us with their videos. Sharing my three favorites:

1: lattice-based zero-knowledge proofs rap
www.youtube.com
Reposted by Sophie Huiberts
primaldual.bsky.social
Some news about active set methods.
sophie.huiberts.me
SODA notifications are in the inboxes, and this paper will be in the conference :)

This is my second paper using extended formulations to prove running time lower bounds. I am surprised that people have been sleeping on this angle, but happy to fill in the gap where needed
An unconditional lower bound for the active-set method in convex quadratic maximization
We prove that the active-set method needs an exponential number of iterations in the worst-case to maximize a convex quadratic function subject to linear constraints, regardless of the pivot rule used...
arxiv.org
sophie.huiberts.me
(to be honest, Nils and Yann did most of the work during those weeks)
sophie.huiberts.me
The paper came from a lucky get-together. Nils and Yann had a problem without solution, Eleon and I had a construction without a purpose. Together we had a 3 week paper writing session
sophie.huiberts.me
SODA notifications are in the inboxes, and this paper will be in the conference :)

This is my second paper using extended formulations to prove running time lower bounds. I am surprised that people have been sleeping on this angle, but happy to fill in the gap where needed
An unconditional lower bound for the active-set method in convex quadratic maximization
We prove that the active-set method needs an exponential number of iterations in the worst-case to maximize a convex quadratic function subject to linear constraints, regardless of the pivot rule used...
arxiv.org
Reposted by Sophie Huiberts
chadtopaz.bsky.social
My forthcoming book is already appearing as a "Hot New Release" on Amazon (race relations section) and I AM TRYING NOT TO FREAK OUT. It doesn't actually release until May! I wrote this optimistic and broadly accessible book for you and hope you will consider pre-ordering.*

(* And/or sharing this!)
Unlocking Justice: The Power of Data to Confront Inequity and Create Change
Amazon.com: Unlocking Justice: The Power of Data to Confront Inequity and Create Change: 9780691276151: Topaz, Chad M.: Books
www.amazon.com
sophie.huiberts.me
True but thats money people know how to spend. Cost for remote are both more varied (with more money generally meaning better quality streams) and harder to find money for. Should it be funded from the IRL registration fees? Should there be remote registration fees, how high should they be? not easy
sophie.huiberts.me
link valid for 24h, afterwards you may dm me for an invite
sophie.huiberts.me
cool! is it a server formed by a group of people who already knew each other beforehand?
sophie.huiberts.me
That does sound like a way to become unpopular 😅 How do you think conferences should approach the costs involved in making the workshop be remotely accessible? I've heard that streaming it can be quite expensive
Reposted by Sophie Huiberts
curiouscoding.nl
We have a discord server with a bunch of folks across the world and it's been a lot of fun. It's more for private communication than public outreach though (at least in its current form).
Reposted by Sophie Huiberts
mkoeppe.bsky.social
A long-term approach: Become an ally of the local members of the community who cannot participate in seminars, workshops, conferences, social events in person because of health or other restrictions. Once events are made remote-accessible for locals, they can also be open for remote participants.
sophie.huiberts.me
i'm doing the youtube thing, but i am curious about other means as well
sophie.huiberts.me
question for academics who dont fly:
besides bluesky, how do you maintain your connection with your overseas community?
Reposted by Sophie Huiberts
matbesancon.bsky.social
BTW if you're at the conference and looking for your next research adventure in computational optimization, come say hi!
matbesancon.bsky.social
My first @juliacon.bsky.social in too many years, very excited to see what people are building, using, combining, experimenting
gdalle.bsky.social
JuliaCon Local Paris starts tomorrow 🤩 Which part of our amazing program are you most excited about?
#JuliaLang @juliacon.bsky.social @julialang.org
sophie.huiberts.me
diet problems with 10-50 variables and roughly double that many constraints. they use a paid solver product to solve a lot of LPs for a lot of users