Convergence of the log-Sobolev constant in the central limit theorem

From papers

Let μP(Rd)\mu\in\mathcal{P}(\mathbb{R}^d) be an isotropic distribution, meaning that

xdμ(x)=0andxxTdμ(x)=Id.\int x\,d\mu(x)=0\quad\text{and}\quad\int xx^T\,d\mu(x)=I_d.

Let νn\nu_n denote the normalized nn-fold convolution sequence from the central limit theorem, so that νn\nu_n converges in law to the standard Gaussian distribution, and suppose that CLS(μ)<C_{LS}(\mu)<\infty. The log-Sobolev convergence conjecture.

limnCLS(νn)=1.\lim_{n\longrightarrow\infty}C_{LS}(\nu_n)=1.

Since CLS(γ)=1C_{LS}(\gamma)=1 for the standard Gaussian γ\gamma, this would show that the log-Sobolev constant detects convergence to Gaussianity along the central limit theorem. The analogous convergence for the Poincaré constant is known, while the stated log-Sobolev convergence is proposed here and its status is not established in the supplied text.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Primary source

Thomas A. Courtade and Edric Wang, “Subadditivity of the log-Sobolev constant on convolutions”, arXiv:2508.19648 (2025).

Solutions 0

No solutions have been posted yet.