Here is the inner workings of the accuracy of the YouTube recommendation engine.
How many times have you found new or interesting content by clicking recommended videos on YouTube? The recommendations may or may not have a majority vote of users, but a lot of thought has gone into developing the algorithm that makes those suggestions to users, a senior executive said on Wednesday.
Reiterating how recommendations play an important role in maintaining a responsible platform, YouTube today shared information on how its recommendation engine works.
âRecommendations today generate a significant portion of the overall YouTube audience, even more than channel subscriptions or search. And we think about it responsibly. Our goal is to help viewers access high quality information while minimizing the chances of them seeing problematic content. Our goal is to have limit content views from recommendations below 0.5% of overall views on YouTube. Said Cristos Goodrow, vice president, Engineering, YouTube.
Unlike other platforms, YouTube does not connect viewers to content through their social network. This basically means that unlike Facebook or Instagram, you won’t see any videos on YouTube that your friends or larger social networks are watching.
YouTube’s recommendation philosophy is based on accurately predicting which videos the viewer wants to watch – based on their own interests and preferences, not who they are connected with.
YouTube has built recommendations on the simple premise of helping people find the videos they want to watch that will make them valuable. Viewers find them at work in two places: the ‘home page’, which appears when you first open YouTube, displaying a mix of personalized recommendations, subscriptions, and the latest news and information. and the âTo be continuedâ panel, which appears when watching a video, providing subsequent content suggestions based on the current video.
Additionally, viewers have controls to manage what they want to share and how much they want to share in order to get a personalized experience on YouTube. For example, viewers who do not want personalized recommendations can choose to delete the history of watched videos.
However, there has been criticism of how recommendations work on YouTube. Earlier this year, a study released by Mozilla found that YouTube’s artificial intelligence engine recommends content that users regret watching, increasing views and serving more ads.
With people using YouTube not only for entertainment but also for news and information, the platform used recommendations to reduce low-quality content from being viewed widely, Goodrow said. He built classifiers to identify and prevent recommendation of racy / violent videos, started downgrading sensationalist content, removed any video showing minors in risky situations, and further expanded how the recommendation system is used. to reduce problematic misinformation and borderline content.