Contextual AI Explainer tutorials

This directory houses the Jupyter notebooks which demonstrate how to use each explanation algorithm found in xai.explainer.

• lime_tabular_explainer.ipynb:

  • Explanation algorithm: LIME tabular explainer

  • Target model: RandomForest

  • Dataset: Wisconsin breast cancer

• lime_text_explainer.ipynb:

  • Explanation algorithm: LIME text explainer

  • Target model: Naive Bayes

  • Dataset: 20newsgroups

• lime_text_explainer_with_keras.ipynb:

  • Explanation algorithm: LIME text explainer

  • Target model: Convolutional Neural Network

  • Dataset: 20newsgroups

• shap_tabular_explainer.ipynb:

  • Explanation algorithm: SHAP kernel explainer

  • Target model: RandomForest

  • Dataset: Wisconsin breast cancer

Tutorial

• LIME Tabular Explainer via XAI
  • Step 1: Import libraries
  • Step 2: Train a model on a sample dataset
  • Step 3: Instantiate the explainer
  • Step 4: Build the explainer
  • Step 5: Generate some explanations
  • Step 6: Save and load the explainer
• LIME Text Explainer via XAI
  • Step 1: Import libraries
  • Step 2: Load dataset and train a model
  • Step 3: Instantiate the explainer
  • Step 4: Build the explainer
  • Step 5: Generate some explanations
  • Step 6: Save and load the explainer
• Contextual AI LIME Text ExplainerFactory with Keras
  • Step 1: Import libraries
  • Step 2: Load dataset and train a model
  • Prepare the embedding matrix
  • Define the model
  • Train the model
  • Step 3: Instantiate the explainer
  • Step 4: Build the explainer
  • Step 5: Generate some explanations
  • Step 6: Save and load the explainer
• SHAP Kernel Explainer for Tabular Data via Contextual AI
  • Step 1: Import libraries
  • Step 2: Train a model on a sample dataset
  • Step 3: Instantiate the explainer
  • Step 4: Build the explainer
  • Step 5: Generate some explanations
  • Step 6: Save and load the explainer