update tutorials

docs/source/notebooks/1.1-Binary_HGF.md

@@ -183,7 +183,7 @@ from pyhgf.distribution import HGFDistribution
 from pyhgf.response import first_level_binary_surprise
 ```
 
-### two-level model
+### Two-level model
 #### Creating the model
 
 ```{code-cell} ipython3
@@ -196,7 +196,7 @@ hgf_logp_op = HGFDistribution(
 ```
 
 ```{note}
-The data is being passed to the distribution when the instance is created.
+The data is being passed to the distribution when the instance is created, so we won't use the `observed` argument in our PyMC model.
 ```
 
 ```{code-cell} ipython3
@@ -207,29 +207,12 @@ slideshow:
 ---
 with pm.Model() as two_levels_binary_hgf:
 
+    # Set a prior over the evolution rate at the second level.
     omega_2 = pm.Uniform("omega_2", -3.5, 0.0)
 
-    pm.Potential(
-        "hgf_loglike",
-        hgf_logp_op(
-            omega_1=jnp.inf,
-            omega_2=omega_2,
-            continuous_precision=jnp.nan,
-            binary_precision=jnp.inf,
-            rho_1=0.0,
-            rho_2=0.0,
-            pi_1=0.0,
-            pi_2=1.0,
-            mu_1=jnp.inf,
-            mu_2=0.5,
-            kappa_1=1.0,
-            omega_3=jnp.nan,
-            rho_3=jnp.nan,
-            pi_3=jnp.nan,
-            mu_3=jnp.nan,
-            kappa_2=jnp.nan
-        ),
-    )
+    # Call the pre-parametrized HGF distribution here.
+    # All parameters are set to their default value, except omega_2.
+    pm.Potential("hgf_loglike", hgf_logp_op(omega_2=omega_2))
 ```
 
 #### Visualizing the model
@@ -278,7 +261,7 @@ hgf_mcmc.plot_trajectories();
 hgf_mcmc.surprise()
 ```
 
-### three-level model
+### Three-level model
 #### Creating the model
 
 ```{code-cell} ipython3
@@ -295,10 +278,6 @@ hgf_logp_op = HGFDistribution(
 )
 ```
 
-```{note}
-The data is passed to the distribution when the instance is created.
-```
-
 ```{code-cell} ipython3
 ---
 editable: true
@@ -307,30 +286,13 @@ slideshow:
 ---
 with pm.Model() as three_levels_binary_hgf:
 
+     # Set a prior over the evolution rate at the second and third levels.
     omega_2 = pm.Uniform("omega_2", -4.0, 0.0)
     omega_3 = pm.Normal("omega_3", -11.0, 2)
 
-    pm.Potential(
-        "hgf_loglike",
-        hgf_logp_op(
-            omega_1=jnp.inf,
-            omega_2=omega_2,
-            omega_3=omega_3,
-            continuous_precision=jnp.nan,
-            binary_precision=jnp.inf,
-            rho_1=0.0,
-            rho_2=0.0,
-            rho_3=0.0,
-            pi_1=0.0,
-            pi_2=1.0,
-            pi_3=1.0,
-            mu_1=jnp.inf,
-            mu_2=0.5,
-            mu_3=0.0,
-            kappa_1=1.0,
-            kappa_2=1.0,
-        ),
-    )
+    # Call the pre-parametrized HGF distribution here.
+    # All parameters are set to their default value except omega_2 and omega_3.
+    pm.Potential("hgf_loglike", hgf_logp_op(omega_2=omega_2, omega_3=omega_3))
 ```
 
 #### Visualizing the model

docs/source/notebooks/1.3-Continuous_HGF.md

@@ -190,7 +190,7 @@ from pyhgf.distribution import HGFDistribution
 from pyhgf.response import first_level_gaussian_surprise
 ```
 
-### two-level model
+### Two-level model
 #### Creating the model
 
 +++
@@ -216,29 +216,13 @@ The data has been passed to the distribution in the cell above when the function
 ```{code-cell} ipython3
 with pm.Model() as two_level_hgf:
 
-    # omegas priors
+    # Set a prior over the evolution rate at the first level.
     omega_1 = pm.Normal("omega_1", -10, 2.0)
 
+    # Call the pre-parametrized HGF distribution here.
+    # All parameters are set to their default value except omega_1, omega_2 and mu_1.
     pm.Potential(
-        "hgf_loglike",
-        hgf_logp_op(
-            omega_1=omega_1,
-            omega_2=-2.0,
-            continuous_precision=1e4,
-            binary_precision=jnp.nan,
-            rho_1=0.0,
-            rho_2=0.0,
-            pi_1=1e4,
-            pi_2=1e1,
-            mu_1=timeserie[0],
-            mu_2=0.0,
-            kappa_1=1.0,
-            omega_3=jnp.nan,
-            rho_3=jnp.nan,
-            pi_3=jnp.nan,
-            mu_3=jnp.nan,
-            kappa_2=jnp.nan
-        ),
+        "hgf_loglike", hgf_logp_op(omega_1=omega_1, omega_2=-2.0, mu_1=1.0)
     )
 ```
 
@@ -305,28 +289,13 @@ hgf_logp_op = HGFDistribution(
 ```{code-cell} ipython3
 with pm.Model() as three_level_hgf:
 
+    # Set a prior over the evolution rate at the first level.
     omega_1 = pm.Normal("omega_1", -10, 2.0)
 
+    # Call the pre-parametrized HGF distribution here.
+    # All parameters are set to their default value except omega_1, omega_2, omega_3 and mu_1.
     pm.Potential(
-        "hgf_loglike",
-        hgf_logp_op(
-            omega_1=omega_1,
-            omega_2=-2.0,
-            omega_3=-2.0,
-            continuous_precision=1e4,
-            binary_precision=jnp.nan,
-            rho_1=0.0,
-            rho_2=0.0,
-            rho_3=0.0,
-            pi_1=1e4,
-            pi_2=1e1,
-            pi_3=1e1,
-            mu_1=timeserie[0],
-            mu_2=0.0,
-            mu_3=0.0,
-            kappa_1=1.0,
-            kappa_2=1.0,
-        ),
+        "hgf_loglike", hgf_logp_op(omega_1=omega_1, omega_2=-2.0, omega_3=-2.0, mu_1=1.0)
     )
 ```
 

docs/source/notebooks/2-Using_custom_response_functions.md

@@ -271,27 +271,7 @@ with pm.Model() as sigmoid_hgf:
     omega_2 = pm.Normal("omega_2", -2.0, 2.0)
 
     # the main HGF distribution
-    pm.Potential(
-        "hgf_loglike",
-        hgf_logp_op(
-            omega_1=jnp.inf,
-            omega_2=omega_2,
-            continuous_precision=jnp.nan,
-            binary_precision=jnp.inf,
-            rho_1=0.0,
-            rho_2=0.0,
-            pi_1=0.0,
-            pi_2=1e4,
-            mu_1=0.0,
-            mu_2=0.5,
-            kappa_1=1.0,
-            omega_3=jnp.nan,
-            rho_3=jnp.nan,
-            pi_3=jnp.nan,
-            mu_3=jnp.nan,
-            kappa_2=jnp.nan
-        ),
-    )
+    pm.Potential("hgf_loglike", hgf_logp_op(omega_2=omega_2))
 ```
 
 ```{code-cell} ipython3