English
Victor Reis
73 posts
Victor Reis
@vetohaze
Senior researcher @MSFTResearch. Previously postdoc @the_IAS, PhD @uwcse
Bellevue, WA Katılım Temmuz 2020
136 Takip Edilen406 Takipçiler
Wow nice! The 3D-printed geometry reminds me of @prof_grimmer 's optimization courses, likely 3 or 4 years ago when I was at Hopkins.
Victor Reis@vetohaze
I spent a year of my PhD stuck on a 2002 problem of Schechtman. GPT 5.5-Pro helped me finish: vector balancing for zonotopes (shadows of a cube)! For any zonotope Z ⊂ ℝᵈ, v₁,...,vₙ ∈ Z, there are signs x₁,...,xₙ ∈ {-1, 1} with x₁v₁+...+xₙvₙ ∈ O(√d) Z, sharp. [1/4]
English
I spent a year of my PhD stuck on a 2002 problem of Schechtman. GPT 5.5-Pro helped me finish: vector balancing for zonotopes (shadows of a cube)!
For any zonotope Z ⊂ ℝᵈ, v₁,...,vₙ ∈ Z, there are signs x₁,...,xₙ ∈ {-1, 1} with x₁v₁+...+xₙvₙ ∈ O(√d) Z, sharp. [1/4]
English
A harder question is still open: can any d-dimensional zonotope be O(1)-approximated by another zonotope with O(d) segments?
Paper: arxiv.org/abs/2605.23866 [4/4]
English
GPT 5.5-Pro couldn't one-shot it, but it was useful in translations from operator theory to convex geometry, and helped me bound entropy numbers by suggesting Carl's inequality and Gordon's escape theorem (alongside hallucinations), neither of which I was familiar with. [3/4]
English
Victor Reis retweetledi
Victor Reis retweetledi
"We are not trying to meet some abstract production quota of definitions, theorems and proofs. The measure of our success is whether what we do enables people to understand and think more clearly and effectively about math."—Bill Thurston. What it means for AI for math debate:⤵️
English
I remember seeing a video of some discussion between famous mathematicians (Vitali Milman, Gromov, don’t remember who else) and what stuck in my mind is Milman saying that the computer-aided proof of the 4CT is just evidence that the 4CT is an answer to a bad question
Greg Burnham@GregHBurnham
I've heard from mathematicians that if an answer is "ugly", then you might be asking the wrong question. So is square-packing somehow the wrong question?
English
Victor Reis retweetledi
2026 marks half a century since Phil Ochs’ death, yet his lyrics are more relevant now than perhaps even he could have imagined.
currentaffairs.org/news/the-long-…
English
"Well, there really is no need to travel to these evil lands
And though the list grows larger you must try to understand
And try hard not to worry if someday you should hear
That the whole world is off limits—visit Disneyland this year" (Phil Ochs, 1964)
White Papers Policy Institute@WhitePapersPol
@StateDept A visualization for anyone who is curious:
English
Our Algorithms group at MSR Redmond is hiring interns in privacy, generative modeling and theory:
apply.careers.microsoft.com/careers/job/19…
apply.careers.microsoft.com/careers/job/19…
apply.careers.microsoft.com/careers/job/19…
This summer my amazing intern Siyue Liu and I wrote a paper accepted to ITCS 2026: arxiv.org/pdf/2511.18546
English
Victor Reis retweetledi
Decoupling via Affine Spectral-Independence: Beck-Fiala and Koml\'os Bounds Beyond Banaszczyk. arxiv.org/abs/2508.03961
Polski
Victor Reis retweetledi
Boaz Klartag, Joseph Lehec: Affirmative Resolution of Bourgain's Slicing Problem using Guan's Bound arxiv.org/abs/2412.15044 arxiv.org/pdf/2412.15044
English
Our Algorithms group at Microsoft Research is hiring interns in differential privacy, reasoning abilities of LLMs, and theory:
jobs.careers.microsoft.com/global/en/job/…
jobs.careers.microsoft.com/global/en/job/…
jobs.careers.microsoft.com/global/en/job/…
English
Victor Reis retweetledi
"Faster Integer Programming," by Don Monroe, discusses a #paper by @vetohaze (@The_IAS) and Thomas Rothvoss (@UW) which proved a new upper bound on the time required to solve for any integer program. bit.ly/3z4ou5M
English
Victor Reis retweetledi
Integer linear programming (ILP) helps solve certain optimization problems, like airline crew scheduling, that deal in whole numbers. Researchers recently created an ILP algorithm that can run faster than any other. @science_eye reports: quantamagazine.org/researchers-ap…
English
I gave a 15 min talk about subspace flatness at the IAS yesterday:
youtube.com/watch?v=apL9X-…
YouTube
English