Popcorn Hack 1

D.

Data Driven Bias

  • Data doesn’t reflect real world values
  • Explicit Data Bias: what you directly interact with
  • Implicit Data Bias: what it collects based on your activity

Algorithmic Bias

  • Bias in the features of the algorithm itself
  • Can come from data that ins’t biased

Human Driven Bias

  • The creators of the program have subconciously developed biases in the program

Popcorn Hack #2

This algorithm is Data bias and this may have occurred because the developers didn’t account for people with darker skin. This may have been because the testers of the facial recognition have fairer skin, but they did not do this to discriminate or decrease their target audience.

Popcorn Hack #3

  • TikTok and other social media algorithms
  • Phone tracks steps, location, etc.

Mitigation Strategies

  • Diverse testing and hiring (divese data)
  • Pre-processing and in-processing
  • Post-processing (after you have completed your program)
  • Input Correction, Output correction, Classifer Correction