Location
Online
Event Website
https://hicss.hawaii.edu/
Start Date
3-1-2023 12:00 AM
End Date
7-1-2023 12:00 AM
Description
A high-velocity paradigm shift towards Explainable Artificial Intelligence (XAI) has emerged in recent years. Highly complex Machine Learning (ML) models have flourished in many tasks of intelligence, and the questions have started to shift away from traditional metrics of validity towards something deeper: What is this model telling me about my data, and how is it arriving at these conclusions? Previous work has uncovered predictive models generating explanations contrasting domain experts, or excessively exploiting bias in data that renders a model useless in highly-regulated settings. These inconsistencies between XAI and modeling techniques can have the undesirable effect of casting doubt upon the efficacy of these explainability approaches. To address these problems, we propose a systematic, perturbation-based analysis against a popular, model-agnostic method in XAI, SHapley Additive exPlanations (Shap). We devise algorithms to generate relative feature importance in settings of dynamic inference amongst a suite of popular machine learning and deep learning methods, and metrics that allow us to quantify how well explanations generated under the static case hold. We propose a taxonomy for feature importance methodology, measure alignment, and observe quantifiable similarity amongst explanation models across several datasets.
Recommended Citation
Dineen, Jacob; Kridel, Don; Dolk, Daniel; and Castillo, David, "Unified Explanations in Machine Learning Models: A Perturbation Approach" (2023). Hawaii International Conference on System Sciences 2023 (HICSS-56). 2.
https://aisel.aisnet.org/hicss-56/da/big_data_and_analytics/2
Unified Explanations in Machine Learning Models: A Perturbation Approach
Online
A high-velocity paradigm shift towards Explainable Artificial Intelligence (XAI) has emerged in recent years. Highly complex Machine Learning (ML) models have flourished in many tasks of intelligence, and the questions have started to shift away from traditional metrics of validity towards something deeper: What is this model telling me about my data, and how is it arriving at these conclusions? Previous work has uncovered predictive models generating explanations contrasting domain experts, or excessively exploiting bias in data that renders a model useless in highly-regulated settings. These inconsistencies between XAI and modeling techniques can have the undesirable effect of casting doubt upon the efficacy of these explainability approaches. To address these problems, we propose a systematic, perturbation-based analysis against a popular, model-agnostic method in XAI, SHapley Additive exPlanations (Shap). We devise algorithms to generate relative feature importance in settings of dynamic inference amongst a suite of popular machine learning and deep learning methods, and metrics that allow us to quantify how well explanations generated under the static case hold. We propose a taxonomy for feature importance methodology, measure alignment, and observe quantifiable similarity amongst explanation models across several datasets.
https://aisel.aisnet.org/hicss-56/da/big_data_and_analytics/2