SHAP Clustering

SHAP (SHapley Additive exPlanations) clustering is a technique used to enhance the interpretability of machine learning models by grouping similar model predictions based on their feature importance scores (SHAP values). Current research focuses on improving the computational efficiency of SHAP value calculation, particularly for complex models like neural networks, and on developing methods to assess the faithfulness and self-consistency of the resulting explanations. This work is significant because it addresses the critical need for trustworthy and efficient explainable AI (XAI), enabling better understanding and improved reliability of machine learning models across diverse applications, such as energy consumption prediction and recommendation systems.

Papers

October 9, 2024

A Utility-Mining-Driven Active Learning Approach for Analyzing Clickstream Sequences
Danny Y. C. Wang, Lars Arne Jordanger, Jerry Chun-Wei Lin
Active Learning Cost Model High Utility SHAP Clustering

February 7, 2024

Beyond explaining: XAI-based Adaptive Learning with SHAP Clustering for Energy Consumption Prediction
Tobias Clement, Hung Truong Thanh Nguyen, Nils Kemmerzell, Mohamed Abdelaal, Davor Stjelja
Explainable AI Adaptive Learning Model Interpretability Energy Prediction SHAP Clustering

November 13, 2023

On Measuring Faithfulness or Self-consistency of Natural Language Explanations
Letitia Parcalabescu, Anette Frank
Natural Language Explanation Model Explanation Self Consistency LLM Alignment Faithfulness Test SHAP Clustering

September 5, 2023

Computing SHAP Efficiently Using Model Structure Information
Linwei Hu, Ke Wang
Shapley Value Shapley Additive Explanation Approximation Method SHAP Score SHAP Clustering Model Structure

March 11, 2023

Efficient Computation of Shap Explanation Scores for Neural Network Classifiers via Knowledge Compilation
Leopoldo Bertossi, Jorge E. Leon
Explainable AI Black Box Binary Neural Network Efficient Computation Neural Network Classifier SHAP Score Knowledge Compilation SHAP Clustering Classifier Circuit

November 3, 2022

A $k$-additive Choquet integral-based approach to approximate the SHAP values for local interpretability in machine learning
Guilherme Dean Pelegrina, Leonardo Tomazeli Duarte, Michel Grabisch
Machine Learning Shapley Value Shapley Interaction Local Interpretability SHAP Value SHAP Clustering Integral Regression Choquet Integral

June 9, 2022

On the Bias-Variance Characteristics of LIME and SHAP in High Sparsity Movie Recommendation Explanation Tasks
Claudia V. Roberts, Ehtsham Elahi, Ashok Chandrashekar
Recommendation Task Local Explanation Fusion SHAP Explainable Recommendation Dseg Lime LIME Explanation SHAP Clustering

SHAP Clustering

Papers

A Utility-Mining-Driven Active Learning Approach for Analyzing Clickstream Sequences

Beyond explaining: XAI-based Adaptive Learning with SHAP Clustering for Energy Consumption Prediction

On Measuring Faithfulness or Self-consistency of Natural Language Explanations

Computing SHAP Efficiently Using Model Structure Information

Efficient Computation of Shap Explanation Scores for Neural Network Classifiers via Knowledge Compilation

A $k$-additive Choquet integral-based approach to approximate the SHAP values for local interpretability in machine learning

On the Bias-Variance Characteristics of LIME and SHAP in High Sparsity Movie Recommendation Explanation Tasks