Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Youtube Technology

How YouTube's Algorithm Really Works (theatlantic.com) 62

YouTube wants to recommend things people will like, and the clearest signal of that is whether other people liked them. From a report: Pew found that 64 percent of recommendations went to videos with more than a million views. The 50 videos that YouTube recommended most often had been viewed an average of 456 million times each. Popularity begets popularity, at least in the case of users (or bots, as here) that YouTube doesn't know much about. On the other hand, YouTube has said in previous work describing its algorithm that users like fresher content, all else being equal. But it takes time for a post to build huge numbers of views and signal to the algorithm that it's worth promoting. So, the challenge becomes how to recommend "new videos that users want to watch" when those videos are new to the system and low in views. (Finding fresh, potentially hot videos is important, YouTube researchers have written, for "propagating viral content.")

Pew's research reflects this: About 5 percent of the recommendations went to videos with fewer than 50,000 views. The system learns from a video's early performance, and if it does well, views can grow rapidly. In one case, a highly recommended kids' video went from 34,000 views when Pew first encountered it in July to 30 million in August. The behavior of the system was explicable in a few other ways, too, especially as it adapted to making more clicks inside YouTube's system. First, as Pew's software made choices, the system selected longer videos. It's as if the software recognizes that the user is going to be around for a while, and starts to serve up longer fare. Second, it also began to recommend more popular videos regardless of how popular the starting video was.

This discussion has been archived. No new comments can be posted.

How YouTube's Algorithm Really Works

Comments Filter:
  • No not really (Score:3, Insightful)

    by Anonymous Coward on Friday November 09, 2018 @11:15AM (#57617208)

    I look up one video on how to fix something and all i get now are howto videos. I just needed one video to fix it. Its fixed stop sending me fixit videos and let me get back to my cat videos. Right meow

    • by Z00L00K ( 682162 ) on Friday November 09, 2018 @12:43PM (#57617626) Homepage Journal

      That's not the worst with the YT algorithm - the worst is that it suggests videos I already watched like a year ago. Only a rare few of them are worth watching again.

      • the worst is that it suggests videos I already watched like a year ago. No, even worse than that is that I click the "not interested" button, and it says OK, Ill take than into account and then shows me the same damn video choices again Not just stupid, but doubling down on annoyingly stupid!

        And If I have clicked on "No, I don't want that" four or five times, it still sends me the same, 5 year old dross.

        Also, why does the search/home not offer "only today's stuff please"? 10 day old stories about Trump ar

    • As more and more videos are uploaded, the ration of good to Garbage gets worse.

      Eventually you won't be able to find anything good for all the garbage.

    • by Hasaf ( 3744357 )

      An example of a lot worse than that is mentioned in the article. The following is from it and illustrates the problem extremely well:

      My favorite example of how informationally toxic YouTube's algorithm is this:

      Imagine you're high school freshman and got a school assignment about the Federal Reserve.

    • by AmiMoJo ( 196126 )

      It really helps to open the little menu and tell YouTube you are not interested.

      However, the response is delayed. It seems like YouTube has a cron task that regularly sets up new recommendations for you, and it only runs once or twice a day so you have to wait for it.

    • by Askmum ( 1038780 )
      So true. You look at one cat video and the next 10,000 suggestions are cat videos.
      But apart from that, you might think that when you show someone 10 suggestions and he chooses none of them, that person is not really interested in those suggestions so when he reloads you show him something else.
      Noooo. Next 10 suggestions are exactly the same. Sometimes even marking the video as "not interested" doesn't do it, it comes right up again.
      Youtube's algorithm is flawed on a basic level.
  • by taustin ( 171655 ) on Friday November 09, 2018 @11:20AM (#57617228) Homepage Journal

    Very, very poorly.

    Their criteria is what they think they'll make the most money on, and to hell with user preferences. I know this because I can tell it I'm not interested in a particular video, or entire category, then refresh the page, and that video or category will be back. Again and again and again, for months, even though I tell it I'm not interested every time.

    • by Anonymous Coward

      Yea, very true. I'd like to ban Bright Side from appearing on Youtube for me and possible go as far as blocking any video sponsored by Bright Side. Clearly there's no meaningful way to do this. Youtube doesn't care if people post Clickbait content nor do they put any effort into people spamming obviously false content. Then there's just the stuff I'm not interested in. 99.9999% of Youtube is a clusterfuck.

  • by slothman32 ( 629113 ) * <pjohnjackson@gmail. c o m> on Friday November 09, 2018 @11:24AM (#57617240) Homepage Journal

    This is as good a place as any.
    For the last month or so when I clicked the "not interested" box the row collapsed but when I closed then re-opened YT it was there again.
    I don't like seeing the "live gaming" and other things but they keep coming back.

    • by TechyImmigrant ( 175943 ) on Friday November 09, 2018 @12:03PM (#57617414) Homepage Journal

      >I don't like seeing the "live gaming" and other things but they keep coming back.

      I don't like anime, but when you do like a few Japanese bands, the algorithm has only one mapping in its hash table. Japanese --> anime.

  • by SmaryJerry ( 2759091 ) on Friday November 09, 2018 @11:25AM (#57617246)
    YouTube took a wrong turn somewhere with it's algorithm. Recommendations have less variety than basic on air tv channels. The algorithm always assumes you want something similar to what you've previously watched which is a major mistake. Yes I watched a clip of a late night talk show. But I don't need late night talk shows recommended to me constantly. Google has the same issue with advertising. Yes I bought a product but I don't need to buy multiple of them so stop advertising it to me constantly for a month afterwards.
    • Similarity is by itself not a mistake. But generally the way the Youtube algorythm works is that IF you watch something, that thing will taint your viewer log and permanently mark it. So instead of watching something, you click on a bad video, watch a few minutes, exit, and then that users entire catalog is suddenly "The new recommended feature" for quite some time. And this is where the similarity problem comes in: You then get more bad videos from other users because it has similar tags, without any real

  • I forget who (Score:4, Insightful)

    by rsilvergun ( 571051 ) on Friday November 09, 2018 @11:27AM (#57617252)
    but somebody once observed that some of the brightest minds of our generation are figuring out how to get me to watch cat videos (and the 15 second ad spot before them) instead of curing cancer. To say nothing of all the math wiz's working on quicker High Frequency Transactions for Wall Street.
    • by Anonymous Coward

      but somebody once observed that some of the brightest minds of our generation are figuring out how to get me to watch cat videos (and the 15 second ad spot before them) instead of curing cancer. To say nothing of all the math wiz's working on quicker High Frequency Transactions for Wall Street.

      Because we're all individuals who need the same thing - money in the capitalistic world that we live in - because that's where the glory is. Corporations make money - they give money - they are entities for themselves. Humanity in general doesn't offer money - businesses do. Spend your time - expend your life - rinse repeat. Could you have made the world a better place? Sure! Would you have sacrificed your life doing so? Absolutely.

  • So when I search for "Live Slayer Reign 1986" I get:

    - An instructional video on making a dove tail joint
    - A small theater production of The King and I
    - 4 different videos of stick thin pubescents doing covers of Taylor Swift on out of tune acoustic guitars
    - The last ten minutes of Heavy Metal Parking Lot.
    - A video on how the Federal Reserve is bringing back vampires from DNA
  • The first video mentioned in the TFA is: "Bath Song | +More Nursery Rhymes & Kids Songs - Cocomelon (ABCkidTV)". Of course, children. YouTube's largest unofficial demographic that will watch anything put in front of them. One popular children channel has 6 billion views [bbc.com].
  • Comment removed based on user account deletion
  • Not what we want to watch. Even when we want to watch it, what they recommend is a like a double layer 35" thick crust sauasge, beef, pepperoni pizza. We want it, but we avoid it because it isnot good for us. But they keep shoving it under our noses.
    • Rather like how "youtube creators for change" propaganda videos keep mysteriously popping up into my subscription feed.

      Or how "authoritative news sources" get recommended over channels that have organically grown their subscriber count and views by virtue of practicing actual journalism, rather than what passes for it on cable news.

      Or how "trending" is almost always vapid pointless nonsense often with few views or was originally posted long before new and more relevant content, as if someone has a job

  • Comment removed based on user account deletion
  • Sounds like an iterative function. Let's hope it never explodes.
    • youtube also uses browser fingerprinting and ip address for recommendations, it doesn't seem they included this in their experiment, so each walk would be influenced by previous walks. i find the autoplay on youtube combined with the deep learning for recommendations to be an odd combination...the ai is basically reinforcing itself if the user doesn't do anything. you get a situation where the user can walk away from the computer and the AI assumes the user liked all these videos it played. soundcloud down [soundcloudtomp3.io]

It is easier to write an incorrect program than understand a correct one.

Working...