Back to search
2105.11406

Sufficiently dense Kuramoto networks are globally synchronizing

Martin Kassabov, Steven H. Strogatz, Alex Townsend

correcthigh confidence
Category
Not specified
Journal tier
Strong Field
Processed
Sep 28, 2025, 12:56 AM

Audit review

The paper proves that for the homogeneous Kuramoto model on any connected graph with modified connectivity μ̃ > 3/4, the only stable equilibrium is the all‑in‑phase state, yielding μc ≤ 0.75 for loopless graphs. The proof uses moment inequalities for ρ1 and ρ2, a self‑loop/twinning reduction, and a final bound (Theorem 5) that rules out any nontrivial stable equilibria . By contrast, the candidate’s argument hinges on asserting that any equilibrium’s phases can be placed in a single interval of length < π (so that all maximum–neighbor phase differences have nonpositive sine). That assertion is false: a configuration can have all pairwise geodesic distances < π yet fail to lie inside any semicircle, invalidating the key sign argument and the ensuing two‑case dichotomy. Consequently, the model’s proof is flawed, even though parts of it (e.g., that any two‑cluster 0/π equilibrium is a saddle) are correct.

Referee report (LaTeX)

\textbf{Recommendation:} minor revisions

\textbf{Journal Tier:} strong field

\textbf{Justification:}

The manuscript addresses a timely and actively studied question: a connectivity threshold that guarantees global synchronization in the homogeneous Kuramoto model, independent of graph topology. It improves the best known upper bound to μc ≤ 0.75 through a clean and self-contained argument based on inequalities for the first two circular moments combined with a self-loop (twinning) reduction. The results are significant and likely to influence subsequent work on dense networks. I recommend minor revisions for clarity (especially around the self-loop/twinning device and the optimization step establishing the |ρ2| ≥ 1/2 bound), but the core results and proofs appear correct and well motivated.