4 ms·
I think this arises because we can't really model intentions as they involve irrational, subjective, qualitative preferences. .. That's to say, I completely ag
by michael-ax 7y ago
I think this arises because we can't really model intentions as they involve irrational, subjective, qualitative preferences. .. That's to say, I completely agree with you.
- dredds 7y agoTo further the point, the algos are training YT admins based on suspect user metrics. Did anyone ask for; "What did you think of this video?" 5-star feature, or "Recent Posts" ie. "Hi guys, i know i haven't done a video recently blah blah". It appears they are repeating the same mistakes made on G+.
- Robadob 7y agoThe android app "Google Opinions Rewards" occasionally asks me to rate a video i've recently watched, then follows it with a binary choice of recommendation, which I subsequently have to rate the quality of the recommendation. I'd assume that data is being used for training somewhere.