Understand explainable principles
The concept of explainability in AI is crucial for data scientists, auditors, and business decision makers. It enables these stakeholders to understand and justify the decisions made by AI systems and the reasoning behind them. In terms of Microsoft’s responsible AI principles, explainability covers three principles:
- Fairness, or the ability for the system to make decisions that don’t discriminate or apply a bias toward groups or individuals based on identifiers such as gender, race, religion, or sexual orientation
- Transparency in understanding how a model arrived at its result
- Securing the data inputs and outputs to protect the privacy of both organization and personal data
Explainability is vital for ensuring compliance with company policies, industry standards, and government regulations. For data scientists, it involves being able to explain how they achieved specific levels of accuracy and what factors influenced the outcome. Auditors...
