113247 (1) [Avatar] Offline
I've seen recommender systems discussed as good alternatives to kernel-based and memory-based approaches, but there don't seem to be any good resources out there about how to build and tune one. It would be great if you could add a section here on building a recommender system with a deep neural network, perhaps on the Netflix challenge dataset. Some discussion of implicit user data vs explicit, and how you might adjust tuning to account for that would be helpful as well.
258096 (1) [Avatar] Offline

This is a fascinating topic -- thanks for bringing it up.

Unfortunately, this is not discussed in the book. Although the video course doesn't follow the book exactly, i.e. there is extra content, this is beyond the scope of either the book or the video course. If I come across some good resources or develop a case study, I can integrate that into the second part of the course. At the moment, I don't have plans to.

All the best,