From b8390753993f1e04124d7763ab751e0bfd45438a Mon Sep 17 00:00:00 2001 From: Arne Diehl Date: Sat, 7 May 2022 18:15:46 +0200 Subject: [PATCH] =?UTF-8?q?Fix=20Nystr=C3=B6m=20Typo?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- pyro/contrib/gp/models/sgpr.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/pyro/contrib/gp/models/sgpr.py b/pyro/contrib/gp/models/sgpr.py index 90688400ee..f07b14f46c 100644 --- a/pyro/contrib/gp/models/sgpr.py +++ b/pyro/contrib/gp/models/sgpr.py @@ -19,7 +19,7 @@ class SparseGPRegression(GPModel): the covariance matrix :math:`k(X, X)` will require a lot of computational steps to compute its inverse (for log likelihood and for prediction). By introducing an additional inducing-input parameter :math:`X_u`, we can reduce computational cost - by approximate :math:`k(X, X)` by a low-rank Nymstr\u00F6m approximation :math:`Q` + by approximate :math:`k(X, X)` by a low-rank Nystr\u00F6m approximation :math:`Q` (see reference [1]), where .. math:: Q = k(X, X_u) k(X_u,X_u)^{-1} k(X_u, X).