Women in Data Science Conference - Digital Discrimination: Cognitive Bias in Machine Learning

Tools:
AI Fairness 360 Toolkit: http://aif360.mybluemix.net/
Model Asset Exchange: http://ibm.biz/model-exchange

Sources:

Podcasts
https://leanin.org/podcast-episodes/siri-is-artificial-intelligence-biased
https://art19.com/shows/the-ezra-klein-show/episodes/663fd0b7-ee60-4e3e-b2cb-4fcb4040eef1

COMPAS
https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
https://www.technologyreview.com/s/612775/algorithms-criminal-justice-ai/

Data for Black Lives
Conference Notes: https://docs.google.com/document/d/1E1mfgTp73QFRmNBunl8cIpyUmDos28rekidux0voTsg/edit?ts=5c39f92e

Gender Shades Project
http://news.mit.edu/2018/study-finds-gender-skin-type-bias-artificial-intelligence-systems-0212
https://www.youtube.com/watch?time_continue=1&v=TWWsW1w-BVo
https://www.ajlunited.org/fight

Other resources referenced in this talk:
https://www.nytimes.com/2018/02/12/business/computer-science-ethics-courses.html
https://www.vox.com/science-and-health/2017/4/17/15322378/how-artificial-intelligence-learns-how-to-be-racist
https://www.engadget.com/2019/01/24/pinterest-skin-tone-search-diversity/

Transcript

1. Digital Discrimination
   Cognitive Bias in Machine Learning
   1
   

   View Slide

2. “
   “A cognitive bias is a systematic pattern
   of deviation from norm or rationality in
   judgment. Individuals create their own
   "subjective social reality" from their
   perception of the input.”
   - Wikipedia
   2
   

   View Slide

3. 3
   https://twitter.com/alexisohanian/status/1087973027055316994
   

   View Slide

4. Example of Cognitive
   Bias in Machine
   Learning
   

   View Slide

5. Northpointe’s
   COMPAS
   Algorithm
   5 Image Credit: #WOCinTech
   

   View Slide

6. May 2016 - Northpointe’s COMPAS Algorithm
   http://www.equivant.com/solutions/inmate-classification
   6
   Source:
   https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
   

   View Slide

7. May 2016 - Northpointe’s COMPAS Algorithm
   http://www.equivant.com/solutions/inmate-classification
   7
   Source:
   https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
   

   View Slide

8. May 2016 - Northpointe’s COMPAS Algorithm
   http://www.equivant.com/solutions/inmate-classification
   8
   Source:
   https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
   

   View Slide

9. May 2016 - Northpointe’s COMPAS Algorithm
   http://www.equivant.com/solutions/inmate-classification
   9
   Source:
   https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
   

   View Slide

10. 10
    Source:
    https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
    

    View Slide

11. 11
    Source:
    https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
    

    View Slide

12. View Slide

13. “
    “We have entered the age of automation
    overconfident, yet underprepared. If we fail
    to make ethical and inclusive artificial
    intelligence we risk losing gains made in civil
    rights and gender equity under the guise of
    machine neutrality.” - @jovialjoy
    13
    

    View Slide

14. SOLUTIONS?
    14
    What can we do to combat bias in AI?
    

    View Slide

15. 15
    https://art19.com/shows/the-ezra-klein-show/episodes/663fd0b7-ee60-4e3e-b2cb-4fcb4040eef1
    

    View Slide

16. “
    “Coders are the most empowered
    laborers that have ever existed.”
    - @anildash
    16
    

    View Slide

17. EDUCATION IS
    KEY
    17 Image Credit: #WOCinTech
    

    View Slide

18. 18
    https://www.nytimes.com/2018/02/12/business/computer-science-ethics-courses.html
    

    View Slide

19. QUESTIONS POSED TO
    STUDENTS
    ○ Is the technology fair?
    ○ How do you make sure that the data is not
    biased?
    ○ Should machines be judging humans?
    19
    

    View Slide

20. 20
    https://twitter.com/Neurosarda/status/1084198368526680064
    

    View Slide

21. FIX THE
    PIPELINE?
    21 Image Credit: #WOCinTech
    

    View Slide

22. Rediet Abebe, Black in AI
    “Cognitive bias in machine learning is human bias on
    steroids.” - @red_abebe
    22
    

    View Slide

23. January 2019 - New Search Feature on
    https://www.pinterest.com/
    23
    Source: https://www.engadget.com/2019/01/24/pinterest-skin-tone-search-diversity/
    

    View Slide

24. “
    “By combining the latest in machine
    learning and inclusive product
    development, we're able to directly
    respond to Pinner feedback and
    build a more useful product.”
    - Pinterest
    24
    

    View Slide

25. TOOLS TO
    COMBAT BIAS
    25 Image Credit: #WOCinTech
    

    View Slide

26. TOOL #1:
    AI Fairness 360 Toolkit
    Open Source Library
    

    View Slide

27. http://aif360.mybluemix.net/
    

    View Slide

28. View Slide

29. DEMO
    http://aif360.mybluemix.net/
    

    View Slide

30. https://github.com
    /IBM/AIF360
    AI Fairness 360 Toolkit Public Repo
    30
    

    View Slide

31. http://aif360.mybluemix.
    net/community
    AI Fairness 360 Slack Community
    31
    

    View Slide

32. TOOL #2:
    Model Asset Exchange
    Open Source Pre-Trained Deep Learning
    Models
    

    View Slide

33. http://ibm.biz/model-exchange
    

    View Slide

34. http://ibm.biz/model-exchange
    

    View Slide

35. http://ibm.biz/
    model-exchange
    Model Asset eXchange
    35
    

    View Slide

36. 36
    Photo by rawpixel on Unsplash
    

    View Slide

37. THANKS!
    Slides and Links:
    http://bit.ly/wids-bias
    Any questions?
    You can find me on Twitter at:
    @Mo_Mack
    37
    

    View Slide