JMLXAIProject - LIME and SHAP for Julia
This package implements the explainable AI methods LIME and SHAP using XAIBase.jl. LIME and SHAP are model-agnostic explainable AI methods, so they can be used to explain different kinds of models. The JMLXAIproject package provides explanations for image inputs and visualizes them as heatmaps.
| Input | Output - Lime | Output - Shap |
|---|---|---|
 |  |  |
Getting started
To familiarise yourself with the package, there are sample implementations of LIME & SHAP in the documentation. On the other hand, there is the subfolder "demo" in the JMLXAIProject package folder, from which "lime_demo.jl" can be executed, which runs LIME or SHAP.
To use the package please add the following code to your environment
using Pkg
Pkg.add(url="https://github.com/e-strauss/JML_XAI_Project.jl")Documentation for JMLXAIProject.