Adding heteroskedastic tests

doc/source/notebooks/advanced/heteroskedastic.pct.py

@@ -22,7 +22,7 @@
 #
 #
 # ## Heteroskedastic Regression
-# This notebooks shows how to construct a model which uses multiple (2) GP latent functions to learn both the location and the scale and of the Gaussian likelihood distribution. It does so by connecting a **Multi-Output Kernel**, which generates multiple GP latent functions, to a **Heteroskedastic Likelihood**, which maps the latent GPs into a single likelihood.
+# This notebooks shows how to construct a model which uses multiple (2) GP latent functions to learn both the location and the scale of the Gaussian likelihood distribution. It does so by connecting a **Multi-Output Kernel**, which generates multiple GP latent functions, to a **Heteroskedastic Likelihood**, which maps the latent GPs into a single likelihood.
 #
 # The generative model is described as:
 #

tests/gpflow/likelihoods/test_heteroskedastic_constant_variance.py

@@ -0,0 +1,102 @@
+# Copyright 2017-2020 the GPflow authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import numpy as np
+import pytest
+import tensorflow as tf
+import tensorflow_probability as tfp
+
+import gpflow
+from gpflow.likelihoods.heteroskedastic import HeteroskedasticTFPDistribution
+
+tf.random.set_seed(99012)
+
+
+class Data:
+    g_var = 0.345
+    # toData.Y data: shape [N, 1]
+    Y = np.c_[0.5, 0.4, 1.0].T
+    N, _ = Y.shape
+    # single "GP" (for the mean):
+    f_mean = np.c_[0.4, 0.7, 0.9].T
+    f_var = np.c_[0.7, 0.2, 1.3].T
+    equivalent_f2 = np.log(np.sqrt(0.345))
+    F2_mean = np.full((N, 1), equivalent_f2)
+    f2_var = np.zeros((N, 1))
+    # stacked N x 2 arraData.Ys
+    F2_mean = np.c_[f_mean, F2_mean]
+    F2_var = np.c_[f_var, f2_var]
+
+
+def test_log_prob():
+    """
+    heteroskedastic likelihood where the variance parameter is alwaData.Ys constant
+     giving the same answers for variational_expectations, predict_mean_and_var,
+      etc as the regular Gaussian  likelihood
+    """
+    l1 = gpflow.likelihoods.Gaussian(variance=Data.g_var)
+    l2 = HeteroskedasticTFPDistribution(tfp.distributions.Normal)
+    np.testing.assert_allclose(
+        l1.log_prob(Data.f_mean, Data.Y),
+        l2.log_prob(Data.F2_mean, Data.Y),
+    )
+
+
+def test_variational_expectations():
+    # Create likelihoods
+    l1 = gpflow.likelihoods.Gaussian(variance=Data.g_var)
+    l2 = HeteroskedasticTFPDistribution(tfp.distributions.Normal)
+    np.testing.assert_allclose(
+        l1.variational_expectations(Data.f_mean, Data.f_var, Data.Y),
+        l2.variational_expectations(Data.F2_mean, Data.F2_var, Data.Y),
+    )
+
+
+def test_predict_mean_and_var():
+    l1 = gpflow.likelihoods.Gaussian(variance=Data.g_var)
+    l2 = HeteroskedasticTFPDistribution(tfp.distributions.Normal)
+    np.testing.assert_allclose(
+        l1.predict_mean_and_var(Data.f_mean, Data.f_var),
+        l2.predict_mean_and_var(Data.F2_mean, Data.F2_var),
+    )
+
+
+@pytest.mark.skip("Conditional mean is not implemented in heteroskedastic likelihood")
+def test_conditional_mean():
+    l1 = gpflow.likelihoods.Gaussian(variance=Data.g_var)
+    l2 = HeteroskedasticTFPDistribution(tfp.distributions.Normal)
+    np.testing.assert_allclose(
+        l1.conditional_mean(Data.f_mean),
+        l2.conditional_mean(Data.F2_mean),
+    )
+
+
+@pytest.mark.skip("Conditional variance is not implemented in heteroskedastic likelihood")
+def test_conditional_variance():
+    l1 = gpflow.likelihoods.Gaussian(variance=Data.g_var)
+    l2 = HeteroskedasticTFPDistribution(tfp.distributions.Normal)
+    np.testing.assert_allclose(
+        l1.conditional_variance(Data.f_mean),
+        l2.conditional_variance(Data.F2_mean),
+    )
+
+
+@pytest.mark.skip("Currently broken as it returns the sum over outputs when given multiple outputs")
+def test_predict_log_density():
+    l1 = gpflow.likelihoods.Gaussian(variance=Data.g_var)
+    l2 = HeteroskedasticTFPDistribution(tfp.distributions.Normal)
+    np.testing.assert_allclose(
+        l1.predict_log_density(Data.f_mean, Data.f_var, Data.Y),
+        l2.predict_log_density(Data.F2_mean, Data.f2_var, Data.Y),
+    )