AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Recommendations Models That Don't Create Biases in the Market
One thing i have noticed is that the recommendations seem to promote a price creep, like a budget creep. We are very careful about what data we add to the models. So for example, crime data, its not something that we would put into the model because we believe that it might create some inherent biaces in the data which we don't want it to have. Sometimes there is always an element of exploration and exploitation when you're doing machine learning. And sometimes, you know, certainly for me, i can tell you my own story. The region, the area that i was looking at to buy a home myself. And so i ended u buying a home in the ingreater seat