Upgrade to Pro — share decks privately, control downloads, hide ads and more …

The AI 'Black Box'

The AI 'Black Box'

As humans we're inherently biased. Sometimes it's explicit and other times it's unconscious, but as we move forward with technology how do we keep our biases out of the algorithms we create? We need to have a conversation about how AI 'Black Box' should be governed and ask who is responsible for overseeing the ethical standards of these algorithms; and the need for gender balance — having access to large and diverse points of view helps to train algorithms to maintain the principle of fairness.

Carla Vieira

August 16, 2019
Tweet

More Decks by Carla Vieira

Other Decks in Technology

Transcript

  1. The Artificial
    Intelligence “Black-box”
    Carla Vieira
    @carlaprvieira
    Ilustração: Hanne Mostard

    View full-size slide

  2. About me
    Carla Vieira
    Information Systems – USP
    Artificial Intelligence Evangelist
    Community Manager perifaCode
    @car lap r vie ir a @car lap r v@h o tmail.co m

    View full-size slide

  3. Tech Conferences

    View full-size slide

  4. bias
    data
    privacy law
    ethics

    View full-size slide

  5. We need to talk less about Artificial
    Intelligence hype …
    … and more about how we are using this
    technology.

    View full-size slide

  6. #1 Google Photos

    View full-size slide

  7. #2 Gender Shades Article
    http://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a.pdf

    View full-size slide

  8. “Whether AI will help us reach our
    aspirations or reinforce the unjust
    inequalities is ultimately up to us.”
    Joy Buolamwini

    View full-size slide

  9. #4 Google’s Algorithm

    View full-size slide

  10. • 46% false positives for
    African American
    • African American authors are
    1.5 times more likely to be
    labelled “offensive”
    https://homes.cs.washington.edu/~msap/pdfs/sap2019risk.pdf

    View full-size slide

  11. Software COMPAS
    #5 COMPAS Software

    View full-size slide

  12. https://www.research.ibm.com/artificial-intelligence/trusted-ai/diversity-in-faces/

    View full-size slide

  13. Artificial Intelligence needs to
    learn from the real world.
    Creating a smart computer is
    not enough, you need to teach
    it the right thing.
    https://about.google/stories/gender-balance-diversity-important-to-
    machine-learning/?hl=pt-BR

    View full-size slide

  14. Gender Gap in Artificial Intelligence
    “Only 22% of AI professionals globally are female,
    compared to 78% who are male.”
    (The Global Gender Gap Report 2018 - p.28)

    View full-size slide

  15. Human Bias Technology
    Bias

    View full-size slide

  16. Even though these decisions affect humans, to
    optimize task performance ML models often
    become too complex to be intelligible to
    humans: black-box models

    View full-size slide

  17. BLACK BOX
    INPUT OUTPUT

    View full-size slide

  18. BLACK BOX
    INPUT OUTPUT

    View full-size slide

  19. JUSTICE MATH

    View full-size slide

  20. https://www.artificiallawyer.com/2019/06/04/france-bans-judge-analytics-5-years-in-prison-for-rule-breakers/
    "This new law is a complete
    shame for our democracy.”
    Louis Larret Chahine
    Co-founder PREDICTICE

    View full-size slide

  21. https://edition.cnn.com/2019/05/14/tech/san-francisco-facial-recognition-ban/index.html

    View full-size slide

  22. How to open this black-box?
    TRANSPARENCY
    EXPLAINABLE ARTIFICIAL
    INTELLIGENCE
    (XAI)
    TRUST

    View full-size slide

  23. XAI intends to create a new suite of
    ML techniques that produce more
    interpretable ML models

    View full-size slide

  24. Accuracy vs. Interpretability trade-off

    View full-size slide

  25. Pre-modelling
    explainability
    Goal
    Understand/describe data
    used to develop models
    Methodologies
    • Exploratory data analysis
    • Dataset description
    standardization
    • Dataset summarization
    • Explainable feature
    engineering
    Explainable
    modelling
    Goal
    Develop inherently more
    explainable models
    Methodologies
    • Adopt explainable model
    family
    • Hybrid models
    • Joint prediction and
    explanation
    • Architectural adjustments
    • Regularization
    Post-modelling
    explainability
    Goal
    Extract explanations to
    describe pre-developed
    models
    Methodologies
    • Perturbation mechanism
    • Backward propagation
    • Proxy models
    • Activation optimization
    Explainability

    View full-size slide

  26. Post-modelling explainability
    The proposed taxonomy of the post-hoc explainability methods including the four aspects of
    target, drivers, explanation family, and estimator.

    View full-size slide

  27. Post-modelling explainability
    first a perturbation model is used to obtain perturbed versions of the input sequence. Next,
    associations between input and predicted sequence are inferred using a causal inference model.
    Finally, the obtained associations are partitioned and the most relevant sets are selected.

    View full-size slide

  28. http://www.portaltransparencia.gov.br/download-de-dados

    View full-size slide

  29. If we want AI to really benefit
    people, we need to find a way to
    get people to trust it.

    View full-size slide

  30. https://brasil.io/home
    https://serenata.ai/
    https://colaboradados.github.io/

    View full-size slide

  31. Less machines that are going to take our
    jobs and more about what technology can
    actually achieve…

    View full-size slide

  32. The choices we are making today
    about Artificial Intelligence are going
    to define our future.

    View full-size slide

  33. Thank you!
    Carla Vieira
    @carlaprvieira
    [email protected]

    View full-size slide

  34. Useful links
    − AI NOW
    − Racial and Gender bias in Amazon Rekognition
    − Diversity in faces (IBM)
    − Google video – Machine Learning and Human Bias
    − Visão Computacional e Vieses Racializados
    − Machine Bias on Compas
    − Machine Learning Explainability Kaggle
    − Predictive modeling: striking a balance between accuracy and interpretability

    View full-size slide

  35. Useful links
    −Racismo Algorítmico em Plataformas Digitais: microagressões e
    discriminação em código
    −Metrics for Explainable AI: Challenges and Prospects
    −The Mythos of Model Interpretability
    −Towards Robust Interpretability with Self-Explaining Neural
    Networks
    −The How of Explainable AI: Post-modelling Explainability

    View full-size slide