Teaching Your Models to Play Fair | Global AI Student Conf

It is very important to ensure fairness while building an AI system that can scale to a large number of users. Thus, I plan to first talk about how fairness is important while building AI apps. I would then go on to talk about how FairLearn helps us in doing so specifically with the easy to use dashboard interface. I plan to show how FairLearn could be used to assess model fairness and how it does so. I would then talk about mitigation strategies with FairLearn so as to remove the biases with state of the art models and compare multiple models to perform this successfully. As time persists I would also show live demos about using FairLearn and Azure ML to assess and mitigate fairness in an AI system.

Transcript

1. Teaching Your Models
   to Play Fair
   Rishit Dagli
   @rishit_dagli
   

   View Slide

2. View Slide

3. View Slide

4. View Slide

5. View Slide

6. View Slide

7. • Interpretability
   • Privacy
   • Security
   

   View Slide

8. View Slide

9. View Slide

10. View Slide

11. View Slide

12. View Slide

13. View Slide

14. View Slide

15. View Slide

16. View Slide

17. • Property of a person does not affect outcome
    

    View Slide

18. Motivation
    

    View Slide

19. View Slide

20. View Slide

21. Source: Ziyuan Zhong
    

    View Slide

22. View Slide

23. View Slide

24. View Slide

25. View Slide

26. • Dashboards
    

    View Slide

27. • Dashboards
    

    View Slide

28. • Dashboards
    

    View Slide

29. • Dashboards
    

    View Slide

30. • Dashboards
    

    View Slide

31. • Dashboards
    

    View Slide

32. • Dashboards
    

    View Slide

33. • Dashboards
    • Mitigation strategies
    

    View Slide

34. • Dashboards
    • Mitigation strategies
    

    View Slide

35. View Slide

36. View Slide

37. bit.ly/fairness-deck
    

    View Slide

38. View Slide

39. Thank You
    Rishit Dagli
    @rishit_dagli
    

    View Slide