small change in kl div

fvgp/gp.py

@@ -501,7 +501,9 @@ def update_gp_data(
         #######prepare noise covariances##########
         ##########################################
         if noise_variances is not None and callable(self.noise_function):
-            warnings.warn("Noise function and measurement noise provided. noise_variances set to None", stacklevel=2)
+            warnings.warn("Noise function and measurement noise provided. \
+            This can happen if no measurement noise was provided at initialization.\
+            New noise_variances set to None", stacklevel=2)
             noise_variances = None
 
         if noise_variances is None:
@@ -1883,9 +1885,9 @@ def gp_kl_div(self, x_pred, comp_mean, comp_cov, x_out=None):
         ----------
         x_pred : np.ndarray
             A numpy array of shape (V x D), interpreted as  an array of input point positions.
-        comp_mean : np.array
+        comp_mean : np.ndarray
             Comparison mean vector for KL divergence. len(comp_mean) = len(x_pred)
-        comp_cov : np.array
+        comp_cov : np.ndarray
             Comparison covariance matrix for KL divergence. shape(comp_cov) = (len(x_pred),len(x_pred))
         x_out : np.ndarray, optional
             Output coordinates in case of multitask GP use; a numpy array of size (N x L),
@@ -1906,8 +1908,8 @@ def gp_kl_div(self, x_pred, comp_mean, comp_cov, x_out=None):
 
         res = self.posterior_mean(x_pred, x_out=None)
         gp_mean = res["f(x)"]
-        gp_cov = self.posterior_covariance(x_pred, x_out=None)["S"]
-
+        gp_cov = self.posterior_covariance(x_pred, x_out=None)["S"] + np.identity(len(x_pred))*1e-9
+        comp_cov = comp_cov+np.identity(len(comp_cov))*1e-9
         return {"x": x_pred,
                 "gp posterior mean": gp_mean,
                 "gp posterior covariance": gp_cov,
@@ -1949,9 +1951,9 @@ def gp_kl_div_grad(self, x_pred, comp_mean, comp_cov, direction, x_out=None):
 
         gp_mean = self.posterior_mean(x_pred, x_out=None)["f(x)"]
         gp_mean_grad = self.posterior_mean_grad(x_pred, direction=direction, x_out=None)["df/dx"]
-        gp_cov = self.posterior_covariance(x_pred, x_out=None)["S"]
+        gp_cov = self.posterior_covariance(x_pred, x_out=None)["S"] + np.identity(len(x_pred))*1e-9
         gp_cov_grad = self.posterior_covariance_grad(x_pred, direction=direction, x_out=None)["dS/dx"]
-
+        comp_cov = comp_cov + np.identity(len(comp_cov)) * 1e-9
         return {"x": x_pred,
                 "gp posterior mean": gp_mean,
                 "gp posterior mean grad": gp_mean_grad,

fvgp/gpMCMC.py

@@ -125,7 +125,7 @@ def run_mcmc(self, n_updates=10000,
                 "distribution var": np.var(x[int(len(x) - (len(x) / 10)):], axis=0)}
 
     ###############################################################
-    def _jump(self, x_old, obj, prior, likelihood):  # pragma: no cover
+    def _jump(self, x_old, obj, prior_eval, likelihood):  # pragma: no cover
         x_star = x_old.copy()
         if callable(obj.prop_dist):
             print("obj indices: ", obj.indices)
@@ -139,11 +139,11 @@ def _jump(self, x_old, obj, prior, likelihood):  # pragma: no cover
             likelihood_star = self.log_likelihood_function(x_star, self.args)
             if np.isnan(likelihood_star): likelihood_star = -np.inf
             metr_ratio = np.exp(prior_evaluation_x_star + likelihood_star -
-                                prior - likelihood)
+                                prior_eval - likelihood)
             if np.isnan(metr_ratio):  metr_ratio = 0.
             if metr_ratio > np.random.uniform(0, 1, 1):
                 x = x_star.copy()
-                prior = prior_evaluation_x_star
+                prior_eval = prior_evaluation_x_star
                 likelihood = likelihood_star
                 jump_trace = 1.
                 print("accepted")
@@ -156,7 +156,7 @@ def _jump(self, x_old, obj, prior, likelihood):  # pragma: no cover
         print("old x  :", x_old)
         print("new x  :", x)
         input()
-        return x, prior, likelihood, jump_trace
+        return x, prior_eval, likelihood, jump_trace
 
     ###############################################################