1: \begin{abstract}

2: The Polyak--\L{}ojasiewicz (PL) constant of a function $f \colon \R^d\to\R$ characterizes the best exponential rate of convergence of gradient flow for $f$, uniformly over initializations.

3: Meanwhile, in the theory of Markov diffusions, the log-Sobolev (LS) constant plays an analogous role, governing the exponential rate of convergence for the Langevin dynamics from arbitrary initialization in the Kullback--Leibler divergence. We establish a new connection between optimization and sampling

4: by showing that the low temperature limit $\lim_{t\to 0^+} t^{-1} \cls(\mu_t)$ of the LS constant of $\mu_t \propto \exp(-f/t)$ is exactly the PL constant of $f$, under mild assumptions.

5: In contrast, we show that the corresponding limit for the Poincar\'e constant is the inverse of the smallest eigenvalue of $\nabla^2 f$ at the minimizer.

6: \end{abstract}

7: