Two Case Studies: Production ML infrastructure and Recommendation Engines - ML 072
Adventures in Machine Learning
00:00
Coursera - Is There a Way to Sort Out the Best Content?
The data set is exactly what you would expect. We are looking to surface the best possible tiles of, let's say, the top 20 pieces of content re-rows. And then algorithmically, we'll be looking at finding some sort of implicit score associated with this content. How do you find people to record if they don't have enough data? If there's not enough users that have interacted with that, are we going to drown out recommendations by just effectively looking at what's the most popular stuff?"
Transcript
Play full episode
