Fix pm.DensityDist in notebooks/scripts by replacing with pm.Potential

This is done because of the bug described in arviz-devs/arviz#1279. The commit also changes a few parameters in the MLDA .py example to match the ones in the equivalent notebook.
alan-turing-institute · Sep 14, 2020 · c98bdb8 · c98bdb8
1 parent 586d78f
commit c98bdb8
Show file tree

Hide file tree

Showing 5 changed files with 21 additions and 21 deletions.
diff --git a/docs/source/notebooks/MLDA_benchmarks_tuning.ipynb b/docs/source/notebooks/MLDA_benchmarks_tuning.ipynb
@@ -231,9 +231,9 @@
     "            # convert m and c to a tensor vector\n",
     "            theta = tt.as_tensor_variable(parameters)\n",
     "\n",
-    "            # use a DensityDist (use a lamdba function to \"call\" the Op)\n",
+    "            # use a Potential for the likelihood\n",
     "            ll = logl[j]\n",
-    "            pm.DensityDist('likelihood', lambda v: ll(v), observed={'v': theta})\n",
+    "            pm.Potential('likelihood', ll(theta))\n",
     "\n",
     "        coarse_models.append(cmodel)\n",
     "      \n",
@@ -248,8 +248,8 @@
     "        # Convert m and c to a tensor vector\n",
     "        theta = tt.as_tensor_variable(parameters)\n",
     "\n",
-    "        # use a DensityDist (use a lamdba function to \"call\" the Op)\n",
-    "        pm.DensityDist('likelihood', lambda v: logl[-1](v), observed={'v': theta})\n",
+    "        ## use a Potential for the likelihood\n",
+    "        pm.Potential('likelihood', logl[-1](theta))\n",
     "        \n",
     "    return model, coarse_models, true_parameters"
    ]

diff --git a/docs/source/notebooks/MLDA_multilevel_groundwater_flow.ipynb b/docs/source/notebooks/MLDA_multilevel_groundwater_flow.ipynb
@@ -352,9 +352,9 @@
     "        # convert m and c to a tensor vector\n",
     "        theta = tt.as_tensor_variable(parameters)\n",
     "\n",
-    "        # use a DensityDist (use a lamdba function to \"call\" the Op)\n",
+    "        # use a Potential for the likelihood\n",
     "        ll = logl[j]\n",
-    "        pm.DensityDist('likelihood', lambda v: ll(v), observed={'v': theta})\n",
+    "        pm.Potential('likelihood', ll(theta))\n",
     "\n",
     "    coarse_models.append(model)\n"
    ]
@@ -599,8 +599,8 @@
     "    # Convert m and c to a tensor vector\n",
     "    theta = tt.as_tensor_variable(parameters)\n",
     "\n",
-    "    # use a DensityDist (use a lamdba function to \"call\" the Op)\n",
-    "    pm.DensityDist('likelihood', lambda v: logl[-1](v), observed={'v': theta})\n",
+    "    # use a Potential for the likelihood\n",
+    "    pm.Potential('likelihood', logl[-1](theta))\n",
     "\n",
     "    # Initialise an MLDA step method object, passing the subsampling rate and\n",
     "    # coarse models list\n",

diff --git a/docs/source/notebooks/MLDA_multilevel_groundwater_flow_blocked.ipynb b/docs/source/notebooks/MLDA_multilevel_groundwater_flow_blocked.ipynb
@@ -352,9 +352,9 @@
     "        # convert m and c to a tensor vector\n",
     "        theta = tt.as_tensor_variable(parameters)\n",
     "\n",
-    "        # use a DensityDist (use a lamdba function to \"call\" the Op)\n",
+    "        # use a Potential for the likelihood\n",
     "        ll = logl[j]\n",
-    "        pm.DensityDist('likelihood', lambda v: ll(v), observed={'v': theta})\n",
+    "        pm.Potential('likelihood', ll(theta))\n",
     "\n",
     "    coarse_models.append(model)\n"
    ]
@@ -557,8 +557,8 @@
     "    # Convert m and c to a tensor vector\n",
     "    theta = tt.as_tensor_variable(parameters)\n",
     "\n",
-    "    # use a DensityDist (use a lamdba function to \"call\" the Op)\n",
-    "    pm.DensityDist('likelihood', lambda v: logl[-1](v), observed={'v': theta})\n",
+    "    # use a Potential for the likelihood\n",
+    "    pm.Potential('likelihood', logl[-1](theta))\n",
     "\n",
     "    # Initialise an MLDA step method object, passing the subsampling rate and\n",
     "    # coarse models list\n",

diff --git a/docs/source/notebooks/mlda/MLDA_multilevel_groundwater_flow.py b/docs/source/notebooks/mlda/MLDA_multilevel_groundwater_flow.py
@@ -113,24 +113,24 @@
 # Set random field parameters
 field_mean = 0
 field_stdev = 1
-lamb_cov = 0.1
+lamb_cov = 0.05
 
 # Set the number of unknown parameters (i.e. dimension of theta in posterior)
 nparam = 3
 
 # Number of draws from the distribution
-ndraws = 1000
+ndraws = 6000
 
 # Number of burn-in samples
-nburn = 500
+nburn = 6000
 
 # MLDA and Metropolis tuning parameters
 tune = True
 tune_interval = 100
 discard_tuning = True
 
 # Number of independent chains
-nchains = 2
+nchains = 4
 
 # Subsampling rate for MLDA
 nsub = 5
@@ -273,9 +273,9 @@ def perform(self, node, inputs, outputs):
         # convert m and c to a tensor vector
         theta = tt.as_tensor_variable(parameters)
 
-        # use a DensityDist (use a lamdba function to "call" the Op)
+        # use a Potential for the likelihood
         temp = logl[j]
-        pm.DensityDist('likelihood', lambda v, ll=temp: ll(v), observed={'v': theta})
+        pm.Potential('likelihood', temp(theta))
 
     coarse_models.append(model)
 
@@ -300,8 +300,8 @@ def perform(self, node, inputs, outputs):
     # Convert m and c to a tensor vector
     theta = tt.as_tensor_variable(parameters)
 
-    # use a DensityDist (use a lamdba function to "call" the Op)
-    pm.DensityDist('likelihood', lambda v: logl[-1](v), observed={'v': theta})
+    # use a Potential for the likelihood
+    pm.Potential('likelihood', logl[-1](theta))
 
     # Initialise an MLDA step method object, passing the subsampling rate and
     # coarse models list

diff --git a/pymc3/sampling.py b/pymc3/sampling.py
@@ -971,7 +971,7 @@ def _iter_sample(
 
     try:
         step.tune = bool(tune)
-        if hasattr(step, 'reset_tuning'):
+        if hasattr(step, "reset_tuning"):
             step.reset_tuning()
         for i in range(draws):
             stats = None