Fairness in AI (DDSW 2019)

Fairness in AI (DDSW 2019)

• 1.
  Fairness in AI “Facilisdescensus Averno” Henk Griffioen - 2019-05-07
• 2.
  Data Scientists haveto put it the work to not end up in ML hell 2 “The gates of hell are open night and day; Smooth the descent, and easy is the way: But to return, and view the cheerful skies, In this the task and mighty labor lies.” The Works of Virgil (John Dryden)
• 3.
  The impact ofAI on society is not all good. AI can encode and amplify human biases, leading to unfair outcomes at scale 3
• 4.
  Fairness is ahot topic and gaining traction! 4 https://fairmlclass.github.io/
• 5.
  Let’s work throughan example 5
• 6.
  Predict if incomeexceeds $50K/year from 1994 census data with sensitive attributes 6
• 7.
  Getting a predictivemodel is easy! But is it fair? 7
• 8.
  How to measurefairness? 8
• 9.
  Ratio of • probabilityof a positive outcome given the sensitive attribute being true; • probability of a positive outcome given the sensitive attribute being false; is no less than p:100 p%-rule: measure demographic parity 9 40% : 50% = 80%10% : 50% = 20%
• 10.
  Our model isunfair: low probability of high income for black people and women 10
• 11.
  How are wegetting bias in our systems? 11
• 12.
  • Skewed sample:initial bias that compounds over time • Tainted examples: bias in the data caused by humans • Sample size disparity: minority groups not as well represented • Limited features: less informative or accurate data collected on minority groups • Proxies: data implicitly encoding sensitive attributes • … Many reasons why bias is creeping into our systems 12Barocas & Selbst, 2016
• 13.
  Intermezzo: Dutch tax authoritiesin the news 13
• 14.
  Ethnic profiling bythe Dutch tax authorities 14 Profiling people for fraud “A daycare center in Almere sounded the alarm when only non-Dutch parents were confronted with discontinuation of childcare allowances… …The Tax and Customs Administration says that it uses the data on Dutch nationality or non-Dutch nationality in the so-called automatic risk selection for fraud.” https://www.nrc.nl/nieuws/2019/05/20/autoriteit -persoonsgegevens-onderzoekt-mogelijke- discriminatie-door-belastingdienst-a3960840
• 15.
  Is it enoughto leave out data on (second) nationality? 15 …In a response, the Tax and Customs Administration states that the information about the (second) nationality of parents or intermediary is not used in this investigation… “Since 2014, a second nationality with Dutch nationality is no longer included in the basic registration. This has been introduced to prevent discrimination for people with dual nationality.” Is this enough to assure that non-Dutch parents in Almere will not suffer another tax injustice? Towards a fair future?
• 16.
  The model isstill unfair without sensitive data. Biases are still encoded by proxies in the dataset! 16
• 17.
  How can weenforce fairness?
• 18.
  Start out withour normal (biased) classifier… 18 Louppe, 2017
• 19.
  … and addan adversarial classifier as a fairness referee 19 Louppe, 2017
• 20.
  Warming up: trainclassifier to predict income 20
• 21.
  Warming up: trainadversarial to detect unfairness 21 Louppe, 2017
• 22.
  Adversarial training: iterativelytrain classifier and adversarial 22 Louppe, 2017
• 23.
  After enough trainingrounds the classifier gives fair income predictions! 23 https://blog.godatadriven.com/fairness-in-ml
• 24.
  Putting in thework
• 25.
  Demographics parity • Decisionuncorrelated with sensitive attribute Equality of opportunity • Outcome probability is same for different groups No mathematical formulation of fairness. There are many (conflicting) measures 25http://www.ece.ubc.ca/~mjulia/publications/Fairness_Definitions_ Explained_2018.pdf
• 26.
  Many ML fairnessapproaches 26 https://dzone.com/articles/machine-learning- models-bias-mitigation-strategies
• 27.
  • Identify productgoals • What? For whom? • Get the right people in the room • Identify stakeholders • Who might be harmed? How? • Select a fairness approach • How to measure? What interaction? • Analyze and evaluate your system • What decisions are made? • Mitigates issues • Intervention needed? • Monitor continuously and escalation plans • Auditing and transparency • Who certifies your system? Fairness should be part of your product process and larger strategy 27 https://www.slideshare.net/KrishnaramKenthapadi/ fairnessaware-machine-learning-practical- challenges-and-lessons-learned-www-2019-tutorial
• 28.
  Fairness is farfrom being solved and needs active work! 28
• 29.
  Artificial Intelligence atGoogle: Our Principles - Objectives for AI applications 1. Be socially beneficial. 2. Avoid creating or reinforcing unfair bias. 3. Be built and tested for safety. 4. Be accountable to people. 5. Incorporate privacy design principles. 6. Uphold high standards of scientific excellence. 7. Be made available for uses that accord with these principles. Even big companies can’t always practice what they preach 29 https://ai.google/principles/
• 30.
  Data Scientists haveto put it the work to not end up in ML hell 30 “The gates of hell are open night and day; Smooth the descent, and easy is the way: But to return, and view the cheerful skies, In this the task and mighty labor lies.” The Works of Virgil (John Dryden)
• 31.
  An ethics checklistfor data scientists • http://deon.drivendata.org/ Tutorial on fairness for products • sites.google.com/view/wsdm19-fairness-tutorial Community concerned with fairness in ML • www.fatml.org Our blogs • blog.godatadriven.com/fairness-in-ml • blog.godatadriven.com/fairness-in-pytorch Where to go from here? 31