Overview
This 1-day training course with cover the technical application of ethical AI principles. Techniques for quantification and mitigation of bias and fairness considering the data, model and end users. Frameworks for data scientists to operationalise transparency and explainability. Responsible governance after model deployment
Audience: AI development and engineering team members, including data scientists.
Course Content
- Intro and broad outline
- Bias and fairness
- What is fairness?
- Measuring fairness
- Working towards fairness
- Self –guided notebooks
- Explainability and transparency
- What is explainability and why is it important?
- Implementing explainability
- Self –guided notebooks
- Monitoring and governance
- What are the risks once a model has been deployed in production?
- How do we monitor these risks?
- Governance procedures to mitigate the risks
- Group discussion on scenario
- Operationalising responsible AI development
- Generative AI and its ethical challenges
- Summary and conclusions
