(no title)
soundwave106 | 7 years ago
It seems like for these systems, there are often multiple teams developing the algorithms, with multiple algorithms in play behind the recommendation. Each algorithm has their own style of metrics system. Some of them are probably relatively simple with not very sophisticated AI/ML (or even no AI/ML at all) behind it. And, there is a lot of internal political conflict preventing better analysis of said algorithms to refine them.
In other words, you're absolutely right.
No comments yet.