← All problems

Bernstein and Erdős conjectures for Hermite-weighted Lagrange interpolation

For every integer n≥1n\ge 1 and every set of distinct interpolation nodes x1,…,xn∈Rx_1,\ldots,x_n\in\mathbb{R}, determine whether the Bernstein and Erdős characterizations for weighted Lagrange interpolation with Hermite weight w(x)=e−x2w(x)=e^{-x^2} on R\mathbb{R} hold, together with the corresponding de Boor--Pinkus intertwining theorem relating the nodes to the relevant local extrema of the weighted Lebesgue function. The supplied source names these conjectures and the associated theorem, but does not state their full conclusions explicitly.

References

Source: arXiv

Progress summary

Refreshed
Open · solution reported

A new unrefereed preprint claims to prove the conjectures in the Hermite-weighted setting, but the claim has not been independently checked.

The problem asks whether the Bernstein and Erdős interpolation characterizations extend to the Hermite-weighted whole-line setting, along with the associated de Boor–Pinkus intertwining theorem. A newer preprint claims all of these results, despite an earlier preprint identifying a serious flaw in a previous proof.

Known results

  • The classical Bernstein conjecture was solved by T. Kilgore in 1978.
  • The classical Erdős conjecture was subsequently solved by de Boor and Pinkus.
  • For Hermite weights, Szokol’s 2025 preprint gives an n=3n=3 counterexample to the derivative-matrix and intertwining arguments used in an earlier claimed proof, without disproving the conjectures themselves.

October 2026 claimed proof

Béla Nagy, Patrícia Szokol, and Szilárd Gy. Révész claim the Bernstein and Erdős conjectures and the corresponding de Boor–Pinkus theorem for Hermite-weighted interpolation. The preprint is unrefereed, and no independent mathematical assessment or verification was found.

Current status (as of November 2026): A preprint claims the Hermite-weighted conjectures are proved, but the claim is unverified; the earlier proof was reported invalid.

Sources

Solutions 0

No solutions have been posted yet.