Visualize Your Model Errors! Microsoft Toolkit Identifies and Diagnoses ML Failures

Description

Summary

• Imagine an autonomous vehicle traffic sign detector whose accuracy plummets when dealing with rain or unexpected inputs.
• To address this, a Microsoft research team recently introduced Error Analysis, a responsible AI toolkit for describing and explaining system failures.
• Perturbation Exploration observes prediction changes after applying changes for selected data points.
• Data Exploration Global Exploration Local Exploration Perturbation Exploration It’s hoped the new toolkit will enable researchers and practitioners to more efficiently and accurately identify and diagnose error patterns, an important step in the development of robust ML systems.