Connes’ bicentralizer conjecture and Takesaki’s classification problem for flows

For every type III1\mathrm{III}_1 factor MM and every faithful normal state φ\varphi on MM, the bicentralizer algebra B(M,φ)B(M,\varphi) is trivial: B(M,φ)=C1B(M,\varphi)=\mathbb{C}1, where B(M,φ)B(M,\varphi) consists of those x∈Mx\in M such that ∥xan−anx∥→0\lVert xa_n-a_nx\rVert\to 0 strongly for every bounded sequence (an)⊂M(a_n)\subset M satisfying ∥anφ−φan∥→0\lVert a_n\varphi-\varphi a_n\rVert\to 0 strongly.

References

Primary source

arXiv

Additional references

Progress summary

Refreshed
Claimed solved

A new unrefereed paper claims to settle both problems, but the claim has not been independently checked.

The entry concerns Connes’ bicentralizer conjecture and Takesaki’s classification problem for flows, two long-standing questions in von Neumann algebra theory. The reported advance uses a classification theorem for flows and an argument involving actions of the real line.

September 10, 2026 claimed solution

Cyril Houdayer and Amine Marrakchi’s preprint The classification of flows on II₁ factors and Connes' bicentralizer problem claims to resolve both problems. Its classification theorem and bicentralizer argument would settle the stated conjectures if correct.

Current status (as of September 2026): A September 2026 preprint claims both problems are solved, but neither claim is independently verified.

Sources

Solutions 1

RemarkAI-assistedClaimed by OpenAI. Claims bounded modular-spectral recovery for von Neumann algebras with a faithful normal state and scalar centralizer, and an alternative proof of bicentralizer triviality for type III_1 factors with separable predual.See full solutionHide full solution

Claimed by OpenAI. Claims bounded modular-spectral recovery for von Neumann algebras with a faithful normal state and scalar centralizer, and an alternative proof of bicentralizer triviality for type III_1 factors with separable predual.

Scope relative to this problem: The bounded modular-spectral recovery theorem assumes a faithful normal state with scalar centralizer. The manuscript also claims an alternative proof of bicentralizer triviality for type III1 factors with separable predual. The target has no separable-predual restriction; no broader nonseparable or flow-classification assertion is attributed to this companion.

GitHub repository: https://github.com/openai/math

Manuscript: https://github.com/openai/math/blob/adc7f1241b42e322a6451854ab7e4b4c146bf78a/preprints/Bounded-recovery-for-modular-spectral-averages-September-23-2026/Bounded-recovery-for-modular-spectral-averages-September-23-2026.pdf

  • OpenAI-290-02-Bounded-recovery-for-modular-spectral-averages.pdf361,603 bytesOpen