From be9623ade13d38335b291a0500f8f770166bb8fc Mon Sep 17 00:00:00 2001 From: Yuchen Pei Date: Thu, 14 Feb 2019 10:11:38 +0100 Subject: minor --- posts/2019-02-14-raise-your-elbo.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/posts/2019-02-14-raise-your-elbo.md b/posts/2019-02-14-raise-your-elbo.md index e692894..a3dc0db 100644 --- a/posts/2019-02-14-raise-your-elbo.md +++ b/posts/2019-02-14-raise-your-elbo.md @@ -112,7 +112,7 @@ The term \"variational\" comes from the fact that we perform calculus of variations: maximise some functional ($L(w, q)$) over a set of functions ($q$). Note however, most of the VI / VB algorithms do not concern any techniques in calculus of variations, but only uses Jensen\'s inequality -/ the fact the $D(p, q)$ reaches minimum when $p = q$. Due to this +/ the fact the $D(q||p)$ reaches minimum when $p = q$. Due to this reasoning of the naming, EM is also a kind of VI, even though in the literature VI often referes to its fully Bayesian version. -- cgit v1.2.3