Understanding SHAP for Interpretable Machine Learning: A Tutorial and Hands-on Workshop

[N-Nieto]

Title

Understanding SHAP for Interpretable Machine Learning: A Tutorial and Hands-on Workshop

Responsible person(s)

Nicolás Nieto (n.nieto@fz-juelich.de) 1,2,
Federico Raimondo (f.raimondo@fz-juelich.de) 1
Vera Komeyer (v.komeyer@fz-juelich.de) 1,2,3

1 - Institute of Neuroscience and Medicine (INM-7: Brain and Behaviour), Research Centre Jülich, Jülich, Germany
2 - Institute of Systems Neuroscience, Heinrich Heine University Düsseldorf, Düsseldorf, Germany
3 - Department of Biology, Faculty of Mathematics and Natural Sciences, Heinrich Heine University Düsseldorf, Düsseldorf, Germany

Format

Tutorial + Hands-on

Timeframe

2 hours (Tutorial) + 4 hours (Hands-on)

Description

Scope

This tutorial and hands-on aim to provide participants with a comprehensive understanding of SHAP for interpreting machine learning models. The scope includes an explanation and exploration of SHAP principles, practical implementations, and considerations for model interpretation. Participants will gain proficiency in leveraging SHAP values to enhance the explainability of machine learning models across various scenarios, including dealing with unbalanced data, collinear features and the nuanced relationship between causality and correlation.

Proposed Timeline

Tutorial (2 hours)

Introduction to SHAP (15 mins)

· Overview of SHAP principles
· Importance of explainability in machine learning

Why Use SHAP? (15 mins)

· Identifying scenarios for the SHAP application
· Understanding the benefits and limitations of SHAP
· Differentiation of SHAP usage to other XAI tools

When to Use It? And When Not to Use It? (15 mins)

· Providing guidance on when in a project to use SHAP (and when not to)
· Highlighting scenarios (and stages of a project) where SHAP may not be suitable

How to Use SHAP (15 mins)

· Integrating SHAP into machine learning workflows

Caveats: Collinear Features / Causality <-> Correlation / Unbalanced data (20mins)

· Exploring common pitfalls and best practices

Model Interpretation and Reporting (30 mins)

· Translating SHAP outputs into meaningful model interpretations
· Guidelines for reporting SHAP results in research papers
· What we can say about our models with the help of SHAP and what not.

Hands-on Workshop (4 hours)

Practical Implementation (2 hours)

· Participants work through hands-on exercises using SHAP
· Real-world examples to reinforce understanding

Group Discussions and Q&A (1 hour)

· Addressing participant queries and concerns
· Facilitating group discussions on SHAP application challenges

Project Work (1 hour)

· Participants apply SHAP to their datasets or models
· Guidance and feedback from instructors

By the end of the course, participants will have a foundation in SHAP, enabling them to effectively evaluate the cases when applying this technic and how to communicate the interpretability of machine learning models in their respective domains.

Requirements

· Meeting room with tables (preferably FZJ, Building 14.6y, INM-7 meeting room Room 2033 and 2034)
· Participants with laptops and Anaconda or Miniconda installed
· Basic knowledge in Python programming and (supervised) Machine Learning
· Maximum of 20 participants (arrangements can be made for a larger group with a bigger room)