Samplinglib
Lean gate not recorded for this source state main · 0e31a3cda412
Sampling regime

Log-concave + log-smooth

Compare theorems by mathematical guarantee, rate, and source — not by raw crawler rows.

Open the live SampleWiki setting ↗

Theorem and frontier cases
upper · Exact source theorem

Averaged LMC

Theorem 4.3.6(1) (Durmus et al. 2019; weakly convex case)

Theorem statement
\[\begin{gathered}0\preceq\nabla^2V\preceq\beta I_d,\qquad h\asymp\frac{\varepsilon^2}{\beta d},\qquad \bar\mu_{Nh}:=\frac1N\sum_{n=1}^N\widehat\mu_{nh},\\ \sqrt{\operatorname{KL}(\bar\mu_{Nh}\Vert\pi)}\le\varepsilon\quad\text{after}\quad N=O\!\left(\frac{\beta d\,W_2^2(\widehat\mu_0,\pi)}{\varepsilon^4}\right).\end{gathered}\]