Bias in AI: Fundamental Trade-Offs and Algorithms


Vishnu Boddeti

May 3, 2022

VishnuBoddeti

Progress In Machine Learning

Speech Processing
Image Analysis
Natural Language Processing
Physical Sciences



Key Driver
Data, Compute, Algorithms

State-of-Affairs

(report from the real-world)
"Tay, Microsoft's AI chatbot, gets a crash course in racism from Twitter"




"FaceApp's creator apologizes for the app's skin-lightening 'hot' filter"

"Facial recognition is accurate, if you're a white guy"

  • Boulamwini and Gebru, "Gender Shades:Intersectional Accuracy Disparities in Commercial Gender Classification," FAT 2018
Real world machine learning systems are effective but,


are biased,


violate user's privacy and


not trustworthy.

Research Agenda



Build machine learning systems that are fair and trustworthy.

Research Questions



    • What are the exact fundamental limits and trade-offs between utility and fairness?


    • How can we design machine learning systems to achieve these fundamental limits and trade-offs?

Trade-Offs in Algorithmic Fairness

  • B. Sadeghi, S. Dehdashtian, V.N. Boddeti, "Characterizing the Fundamental Trade-offs in Learning Invariant Representations," TMLR 2022

Notions of Algorithmic Fairness

  • There are many notions of fairness (21 definitions):
    • Demographic Parity
    • Equalized Odds
    • Equality of Opportunity

  • Estimate the fundamental limits and trade-offs for all definitions of fairness.

Fairness Under Practically Relevant Scenarios

  • Application Domain: Face Recognition
  • Summary: Next generation of machine learning systems have to be designed with security/privacy/fairness constraints.
Thank You