← All problems

Quantitative two-chaos fourth-moment theorem

1 year old · traced to

Can the qualitative fourth-moment theorem for sums of two Wiener–Itô integrals of different parity be strengthened to an explicit total-variation rate depending only on the fourth cumulant, and can an analogous theorem be established on Poisson space?

References

Progress summary

Refreshed
Open · progress reported

The Gaussian version is proved with an explicit rate, while the Poisson version works only under extra assumptions and fails without them.

The problem asks whether fourth-moment convergence controls total-variation distance for sums of two Wiener–Itô integrals of different parity, and whether a comparable result holds for Poisson chaoses. No proposer or proposal date was identified.

February 2025 Gaussian theorem

For normalized sums Z=Ip(f)+Iq(g)Z=I_p(f)+I_q(g) with pp and qq of different parity, the theorem proves

dTV(Z,N(0,1))≤2Cp,qκ4(Z).d_{\mathrm{TV}}(Z,N(0,1))\leq 2\sqrt{C_{p,q}\kappa_4(Z)}.

A sharper bound is dTV(Z,N(0,1))≤6κ4(Z)d_{\mathrm{TV}}(Z,N(0,1))\leq\sqrt{6\kappa_4(Z)}. Thus the qualitative theorem and an explicit fourth-cumulant rate are established. Equal parity is genuinely excluded: a counterexample exists for p=1p=1 and q=3q=3.

September 2025 Poisson analogue

For Poisson-chaos sums Zn=Xn+YnZ_n=X_n+Y_n, fourth-moment convergence implies Gaussian convergence when the mixed odd moments satisfy E[XnYn3]→0\mathbb{E}[X_nY_n^3]\to0 and E[Xn3Yn]→0\mathbb{E}[X_n^3Y_n]\to0. A normalized p=1,q=3p=1,q=3 construction has fourth moment 33 but is non-Gaussian, proving these assumptions cannot simply be removed.

Current status (as of September 2025): The Gaussian question is resolved with quantitative total-variation bounds; the Poisson question has only a conditional analogue and is false without the additional mixed-moment hypotheses.

Sources

Solutions 0

No solutions have been posted yet.