Question from the audience: where to look first to mitigate unfairness: the data or the model?
The observation is that the model tends to accentuate the biases in the input data.
🔆 Roundtables going full steam ahead at #Mais22.
Great discussions on Fairness and Algorithmic harms, Interpretability, Robustness, Security and Privacy, Ethics, and Accountability
⏳ The countdown for #MAIS22 is on! We're really excited about the program we've put together with keynotes, roundtables, panel, contributed talks and posters.
🧵 Here is a thread with the highlights of the program that you'll see tomorrow Saturday at @Mila_Quebec#ResponsibleAI