MindSpore XAI Documents

MindSpore XAI is an interpretable AI tool, for comparing and selecting interpretation methods that are most appropriate for a specific scenario.

Currently, most deep learning models are black-box models with good performance but poor explainability. MindSpore XAI provides a variety of explanation and decision methods to help you better understand, trust, and improve models. It also evaluates the explanation methods from various dimensions.

Code repository address: <https://gitee.com/mindspore/xai>

Installation

• MindSpore XAI Installation

Guide

• Using CV Explainers
• Using CV Benchmarks
• Using Tabular Explainers
• Using TabSim Data Simulator
• Using TB-Net Whitebox Recommendation Model

API References

• mindspore_xai.explainer
• mindspore_xai.benchmark
• mindspore_xai.tool
• mindspore_xai.visual

FAQ

• Troubleshooting