Predicting Diagonalizability of a Mean Matrix
Wu and Santhanam asked whether one can determine, from an increasing i.i.d. sample of binary random matrices, whether the unknown mean matrix is diagonalizable, while making only finitely many errors almost surely. Answered affirmatively over both R and C.
- Result
- Proved(see note)
- Status
- Resolved
- AI contribution
- AI-discovered
- Method
- Argument
- Field
- Statistical learning theory
- Posed by
- Yuheng Wu, Narayana Santhanam
- Year posed
- 2024
- Years open
- 2y
- Solved
- 2026-08-11
- Model
- GPT-5.6 Sol Ultra
- Vendor
- OpenAI
- Collaborators
- Jinze Zhao
- Verification
- Unreviewed
- Publication
- Preprint
- Significance
- 8 / 100
- Disclosed cost
- —
- Wikipedia
- No dedicated article
What was actually shown
The general principle is the interesting part: every semialgebraic property of a bounded fixed-dimensional mean parameter is eventually almost surely predictable. Against merely integrable matrix laws it fails from dimension two.
What the AI did
A disclosure section of its own: "The proof strategy and counterexample were produced by OpenAI's GPT-5.6 Sol Ultra through Codex in response to prompts from the author. Codex was also used to revise the exposition and prepare the LaTeX manuscript. The author selected the problem, directed the interactions and revisions, and is the sole named author."
Verification
A preprint days old. The paper says so itself: "This disclosure is not a substitute for independent expert mathematical review."
Source
- PaperarXiv