A comparison of explainable artificial intelligence methods in the phase classification of multi-principal element alloys

Abstract

We demonstrate the capabilities of two model-agnostic local post-hoc model interpretability methods, namely breakDown (BD) and shapley (SHAP), to explain the predictions of a black-box classification learning model that establishes a quantitative relationship between chemical composition and multi-principal element alloys (MPEA) phase formation. We trained an ensemble of support vector machines using a dataset with 1,821 instances, 12 features with low pair-wise correlation, and seven phase labels. Feature contributions to the model prediction are computed by BD and SHAP for each composition. The resulting BD and SHAP transformed data are then used as inputs to identify similar composition groups using k-means clustering. Explanation-of-clusters by features reveal that the results from SHAP agree more closely with the literature. Visualization of compositions within a cluster using Ceteris-Paribus (CP) profile plots show the functional dependencies between the feature values and predicted response. Despite the differences between BD and SHAP in variable attribution, only minor changes were observed in the CP profile plots. Explanation-of-clusters by examples show that the clusters that share a common phase label contain similar compositions, which clarifies the similar-looking CP profile trends. Two plausible reasons are identified to describe this observation: (1) In the limits of a dataset with independent and non-interacting features, BD and SHAP show promise in recognizing MPEA composition clusters with similar phase labels. (2) There is more than one explanation for the MPEA phase formation rules with respect to the set of features considered in this work.

On the importance of interpretable machine learning predictions to inform clinical decision making in oncology

“…Insights into feature effects on model predictions from the Shapley additive explanation (SHAP) analysis. The CP plots provide a counterfactual interpretation to quantify feature effects and offer clear visualization to investigate the relationships between model responses and features (62). However, the approach is limited to displaying information for one feature at a time.…”

How to Make the Most of Local Explanations: Effective Clustering Based on Influences

“…With the rise of explainability, ML research looks beyond simply explaining the machine learning model. Several papers in the last year have covered use cases combining machine learning explainability and clustering to find relationships between instances [3,10]. Based on a COVID-19 dataset, [3] tries to identify better clusters based on KernelSHAP values.…”

Predicting and Evaluating Decoring Behavior of Inorganically Bound Sand Cores, Using XGBoost and Artificial Neural Networks

“…This work uses machine learning methods to model cause-effect relationships in the context of decoring inorganically bound sand cores. There are various examples of gaining insights from data in many technology areas, for example, material analysis [7], laser beam welding [8], or injection molding [9]. The basic idea is to train machine learning models using annotated data.…”