VibeMathedMath problems solved by AI

Large Hypercube Intervals in Bruhat Order

How large can a Bruhat interval in SnS_n that is a poset hypercube be? Using a permutation pattern suggested by AlphaEvolve, the authors exhibit hypercube intervals of dimension O(nlogn)O(n \log n) for nn a power of 2, matching the largest possible dimension up to a constant - in the problem circle around the combinatorial invariance conjecture for Kazhdan-Lusztig polynomials.

Result
Proved(see note)
Status
Partial result
AI contribution
AI co-developed
Method
Construction
Field
Coxeter combinatorics
Posed by
Year posed
Years open
Solved
2026-01-03
Model
AlphaEvolve
Vendor
Google DeepMind
Collaborators
Jordan Ellenberg, Nicolas Libedinsky, David Plaza, José Simental, Geordie Williamson
Verification
Unreviewed
Publication
Preprint
Significance
10 / 100
Disclosed cost
Wikipedia
No dedicated article

What was actually shown

Asymptotically optimal for powers of 2; the exact extremal answer for general n stays open.

What the AI did

AlphaEvolve, searching evolutionarily rather than exhaustively, "produced a pattern which performed well for the n tested, and which we show works well for general n" - the agent found the construction, the humans proved it works in general.

Source

arXiv

Discussion