9 ms·
This quote raises questions: "The scraped data, as well as the YouTube API, provides usa view of the recommendations presented to an anonymous account.
by joe_the_user 7y ago
This quote raises questions:
"The scraped data, as well as the YouTube API, provides usa view of the recommendations presented to an anonymous account. In other words, the account has not ”watched” any videos, retaining the neutral baseline recommendations, described in further detail by YouTube in their recent paper that explains the inner workings of the recommendation algorithm[38]. One should note that the recommendations list provided to a user who has an account and who is logged into YouTube might differ from the list presented to this anonymous account."
Many discussions of radicalization talk about recommendations coming after someone has watched a number of videos.
This situation is that after N videos, the algorithm starts to see a pattern, a niche, that the user fits into and then recommends more niche content. Once the user watches the niche content, just more niche content is recommended, or that-niche and mainstream content is recommended and the mainstream content is boring.
In a sense, I don't see how a blind recommendation system could escape this situation. It seems more an inherent problem of blind recommendation systems.
What could be a better approach?
* Better education in the US and the world so people have BS detectors?
* Instead of unlabeled recommendations, the algorithm categorizes what you're seen, categorizes what it's recommending and let's the user say what category they want and remove or add categories for recommendation.
* Have more ordinary search tool and a better categorization system (Youtube is a horrible black-box now, search is terrible, recommendations is terrible, I get videos from other sites mostly).
- pitay 7y agoHave the recommendation system mix it up, put niches that are somewhat similar in the recommendations not quite as often, put things a bit further away less often, etc. This should have the algorithm recommend counterpoints which are clearly visible to the user. This is assuming things that are partisan one way are similar to things that are partisan in the other way and counter partisan content. Avoid censorship, censorship is too easily abused.
- pjc50 7y agoWe should not classify "refusal to ever recommend X because you know X to be untrue or defamatory" as censorship.
- pitay 7y agoI didn't mention anything specific, so I am not sure what you are replying to. Your example can be dealt with better in other ways, such as putting a prominent statement with a brief description of why something is untrue and a link to further information. For YouTube something like a popup that appears on the video which has to be manually closed, and also the information appearing under the video. Censorship in general is terrible, and abused very regularly. Even on the really small scale I see little mafias pop up all the time, such as on forums, where anyone a bit critical of the select group gets banned and their posts deleted. On a larger scale is Putin announcing that criticism of the state is to be censored, citing other European laws do just that and punish people for what they say. If someone is censoring someone it is just extremely suspicious, and I don't blame anyone for going 'it is censored, therefor it is likely true', because the censor has removed the information that could be used as evidence for whatever it is. "When you tear out a man's tongue, you are not proving him a liar, you're only telling the world that you fear what he might say." - George R.R. Martin, A Clash of Kings
- golergka 7y agoIt's not censorship, but this position makes you not simply a neutral platform, but an opinionated editor.
- pjc50 7y agoThe platform is not and cannot be neutral; they necessarily remove porn and copyright infringing material, plus whatever they are legally obliged to do in various countries. An endless series of judgement calls. Moreover, we're talking about the recommendation algorithm. The only neutral recommendation algorithm is a random number generator. Everything else is making a decision that video X is "better" than video Y for user Z. What it means to be "better" is difficult to analyse, but Google themselves must be constantly reviewing that question.