v0.1.2 Fix categorical initialization bug (#15)

* fix categorical init bug

* update gitignore

* update README

.gitignore

@@ -162,3 +162,4 @@ cython_debug/
 .DS_Store
 data
 *.out
+assets

README.md

@@ -1,7 +1,7 @@
 # MixMIL
 Code for the paper: [Mixed Models with Multiple Instance Learning](https://arxiv.org/abs/2311.02455)
 
-Accepted at AISTATS 24 as an oral presentation.
+Accepted at AISTATS 24 as an oral presentation & [Outstanding Student Paper Highlight](https://aistats.org/aistats2024/awards.html).
 
 Please raise an issue for questions and bug-reports.
 ## Installation

mixmil/__init__.py

@@ -1,4 +1,4 @@
 from mixmil.model import MixMIL
 
 __all__ = ["MixMIL", "utils", "likelihood", "posterior", "data", "simulation", "paths"]
-__version__ = "0.1.1"
+__version__ = "0.1.2"

mixmil/utils.py

@@ -53,7 +53,7 @@ def get_lr_init_params(X, Y, b, Fiv):
         penalty="l2",
         multi_class="multinomial",
         solver="lbfgs",
-        n_jobs=8,
+        n_jobs=1,
         verbose=0,
         random_state=42,
         max_iter=1000,
@@ -76,7 +76,7 @@ def get_lr_init_params(X, Y, b, Fiv):
     alpha = Fiv.dot(np.ones((Fiv.shape[1], 1))).dot(alpha) - b.dot(mu_beta)
 
     # init prior
-    var_z = (mu_beta**2 + sd_beta**2).mean().reshape(1, 1)
+    var_z = (mu_beta**2 + sd_beta**2).mean(axis=0, keepdims=True)
 
     return [torch.Tensor(el) for el in (mu_beta, sd_beta, var_z, alpha)]
 

pyproject.toml

@@ -37,9 +37,9 @@ classifiers = [
 ]
 
 [project.urls]
-Homepage = "https://github.com/AIH-SGML/MixMIL"
-"Bug Tracker" = "https://github.com/AIH-SGML/MixMIL/issues"
-Discussions = "https://github.com/AIH-SGML/MixMIL/discussions"
+Homepage = "https://github.com/AIH-SGML/mixmil"
+"Bug Tracker" = "https://github.com/AIH-SGML/mixmil/issues"
+Discussions = "https://github.com/AIH-SGML/mixmil/discussions"
 
 [project.optional-dependencies]
 experiments = ["anndata>=0.8.0", "jupyterlab>=3.0.0"]

tests/test_model.py

@@ -26,6 +26,7 @@ def mock_data_categorical():
 def test_init_with_mean_model_binomial(mock_data_binomial):
     Xs, F, Y = mock_data_binomial
     model = MixMIL.init_with_mean_model(Xs, F, Y, likelihood="binomial", n_trials=2)
+    model.train(Xs, F, Y, n_epochs=3)
     assert isinstance(model, MixMIL)
     assert model.likelihood_name == "binomial"
     assert model.n_trials == 2
@@ -35,6 +36,7 @@ def test_init_with_mean_model_binomial(mock_data_binomial):
 def test_init_with_mean_model_categorical(mock_data_categorical):
     Xs, F, Y = mock_data_categorical
     model = MixMIL.init_with_mean_model(Xs, F, Y, likelihood="categorical")
+    model.train(Xs, F, Y, n_epochs=3)
     assert isinstance(model, MixMIL)
     assert model.likelihood_name == "categorical"
     assert model.n_trials is None