← All problems

Tokareva’s conjecture on sums of two bent functions; Tokareva’s derivative conjecture

For every even integer n≥4n\ge 4 and every Boolean function f:F2n→F2f:\mathbb{F}_2^n\to\mathbb{F}_2 with algebraic degree deg⁡(f)≤n/2\deg(f)\le n/2, there exist bent functions g,h:F2n→F2g,h:\mathbb{F}_2^n\to\mathbb{F}_2 such that f=g+hf=g+h, where addition is pointwise over F2\mathbb{F}_2. Here a Boolean function bb on F2n\mathbb{F}_2^n is bent if its Walsh transform has absolute value 2n/22^{n/2} at every frequency.

References

Source: arXiv

Progress summary

Refreshed
Open · solution reported

An unrefereed 2026 preprint claims explicit counterexamples to both conjectures, but neither claim has independent mathematical verification.

The conjectures propose characterizations involving sums of two bent functions and derivatives of bent functions. The latest preprint claims to refute both statements with explicit constructions.

Known results

  • Potapov (2023) proved asymptotically that, for sufficiently large even nn, balanced functions of degree at most n/2−1n/2-1 need not be derivatives of bent functions.
  • The derivative conjecture had previously been established for quadratic balanced functions in even dimensions n≥6n \ge 6.

October 2026 counterexamples

Feng, Jiang, Li, Li, and Tang claim an explicit Boolean function on F28\mathbb{F}_2^8 that is not a sum of two bent functions, with extensions to every even n≥8n \ge 8. They also claim counterexamples to Tokareva’s derivative conjecture for every even n≥10n \ge 10. The preprint is unrefereed, so these are reported as unverified claims.

Current status (as of October 2026): The conjectures are claimed to be refuted in the stated dimensions, but the counterexamples remain independently unverified.

Sources

Solutions 0

No solutions have been posted yet.