131–140 of 156 posts
Earlier quoted context omitted.
It provides more information. It's shallow and non-descript because it's an ad is the argument. I don't know if I believe that here. It's a blurry line with sponsored content.
Well then it's not "short-hand".
Earlier quoted context omitted.
By "the field", you surely mean the academic field. In the industry, we run controlled experiments to validate all the time. Recommender systems is one of the few areas in ML where almost all of the knowledge is contained in industry, not academia.
That was my thinking - anything of value is product-specific and behind closed doors. It's not my field, but something I see come up from time to time that seems weirdly over-represented in ML articles.
With all of this technology applied, I am still disappointed by Netflix's recommendations – to the point of just giving up and doing something else.
Do you think part of this is that Netflix has assumed zero effort from user model? My experience has been that Netflix does an ok job of recommendations, but fails at overall discovery experience. There is no way for me to drive or view content from different angles easily. I end up googling for expert opinions or hitting up rotten tomatoes to get better reviews. Netflix knows a ton about me and their content, but se…
TikTok, on the other hand, has way more data. Things like time-to-swipe, shares, comments presumably form the basis of some sentiment metric.
Earlier quoted context omitted.
TikTok
Why TikTok in particular? What is the engineering story behind TikTok's recommendation system? How did they get it right?
https://www.eugenewei.com/blog/2020/9/18/seeing-like-an-algo...
This doesn't make much sense to me since a recommendation is rarely needed instantly. Why not spend, say, 10 s constructing a better recommendation while the user is doing something else, during which the recommendation can simply be blank. Obviously if the user requests a recommendation on first visit, you're out of luck, but I'm thinking the typical use case is for a recommendation after the primary reason for visiting has been completed.
Earlier quoted context omitted.
I have as well, and your comment matches my experience more than the article does. Different teams own different systems, and there's basically no intersection between "things that require a ton of data/computation" and "things that must be computed online".
Yep. The author, as a peddler of recommendations solutions, has an incentive to convince people that this problem is very complicated, and they should hire a consultant. In practice, good old Matrix Factorization works really well. Can you beat it with a huge team and tons of GPU hours to train fancy neural nets? Probably. Can you set up a nightly MF job on a single big machine and serve results quickly? Sure can.
https://www.linkedin.com/pulse/personalized-recommendations-...
Earlier quoted context omitted.
This is a pretty misinformed take when it’s publicly known that YouTube was already doing this (learn from what the user is watching and for how long) the year Bytedance was founded (2012): https://blog.youtube/news-and-events/youtube-now-why-we-focu...
Somehow they're doing it better. At least subjectively, people complain more about the YouTube algo's performance than tiktok. For the latter, the most common complaint is that it's too good.
Earlier quoted context omitted.
Yep. The author, as a peddler of recommendations solutions, has an incentive to convince people that this problem is very complicated, and they should hire a consultant. In practice, good old Matrix Factorization works really well. Can you beat it with a huge team and tons of GPU hours to train fancy neural nets? Probably. Can you set up a nightly MF job on a single big machine and serve results quickly? Sure can.
I thought the two-tower embedding model was now the go-to approach for building real-time recommendation engines? https://www.linkedin.com/pulse/personalized-recommendations-...
If you want cross entropy loss, use word2vec.
Follow it up with a lambdarank ranker.
Is there any recommendation system people we actually happy with? They all seem to suck in my experience
YouTube, TikTok, and Twitter all work well for me.
I know that they probably optimize for ads etc. but if they actually showed me videos I would like to see, then I would spend more time on the platform.
Earlier quoted context omitted.
Somehow they're doing it better. At least subjectively, people complain more about the YouTube algo's performance than tiktok. For the latter, the most common complaint is that it's too good.
TikToks are shorter. So in any given session, a user will give a lot more feedback to TikTok than to YouTube.