Recipe
publicShareStar

Proper Machine Learning Explanations through LIME, using the OptiLIME framework

Collaborators
Reviewers
Total time needed: ~3 hours
Learning Objectives
With this list you will learn approaches to improve the explanations generated with LIME
Potential Use Cases
Improve LIME explanations
Target Audience
INTERMEDIATEPeople with basic knowledge on interpretable machine learning
Go through the following annotated items in order:
VIDEO 1. Proper Machine Learning Explanations through LIME using OptiLIME framework
  • Can we optimize the LIME explanations?
60 minutes
PAPER 2. Understand explaining the predictions of any machine learning models
  • Can we explain blackbox models?
  • Are the explanations useful for evaluating the model ?
30 minutes
VIDEO 3. Understand local interpretable model-agnostic explanations (LIME)
  • Can we explain blackbox models?
25 minutes
PAPER 4. How can we deal with instability associated with LIME explanations?
  • Can explanations generated by a locally interpretable model provide consistent results for the same instance?
30 minutes
PAPER 5. Understand generating robust and stable explanations
  • Can we generate explanations robust to data shifts?
  • Can we generate stable explanations?
30 minutes

Concepts Covered