Parisa successfully defends her Master's!

First, a huge congratulations to the now Parisa Salmani MSc for successfully defending their thesis entitled Bias and Fairness in Transfer Learning.

Parisa’s research demonstrated, for the first time, that the use of transfer learning (a common technique used in all sorts of applications of machine learning) can lead to the introduction of new demographic bias and hence introduce discrimination against particular sub-populations. Parisa demonstrated this across two domains when comparing models that were trained from scratch to perform the same task.

Excellent work and a great milestone!