Motivating (not quite) example: a non-degenerate Gaussian is log-concave, and Bernstein von Mises says well-identified smooth models are approximately Gaussian, but I think in a sense too weak to apply a bounded-perturbation-of-log-concave inequality (eg h
2,574 followers
657 followers
17,381 followers
23,421 followers
Book #otd "Analysis and Geometry of Markov Diffusion Operators" by Dominique Bakry, Ivan Gentil, Michel Ledoux Euclidean/Spherical/Hyperbolic heat semigroups! https://t.co/XnOXFvKF4s https://t.co/MUEHp3smOQ