The popular short video application TikTok (TikTok overseas version) has attracted hundreds of millions of users worldwide to continue to indulge in it, thanks in large part to algorithms. The company’s executives recently disclosed how this “killer” algorithm works, including how to use machine learning to maximize user engagement and how to break the shackles of the “information cocoon room”.
TikTok’s algorithm can recommend users to watch the next video, it helps the app gain a lot of popularity and tens of billions of dollars in valuation. Under pressure from the United States, ByteDance is preparing to sell TikTok’s US business, and the algorithm has become one of the focal points of potential transactions. TikTok executives said in a conference call with reporters that they are disclosing details about algorithms and data operations to eliminate various rumors and misunderstandings about the company.
Michael Beckerman, TikTok’s vice president of public policy in the United States, said: “Even though we have just been established for two years, we have long-term goals. We have no chance to grow in the golden age of the Internet, when technology companies would not do anything wrong Things. We grew up in an era of technological frenzy, and people were skeptical about platforms, how they reviewed content, and how their algorithms worked.”
TikTok executives provided reporters with a virtual tour to visit its new “Transparency Center” in Los Angeles. The center will provide people with an area to demonstrate computer modules that show how TikTok’s algorithms and data operations work.
TikTok’s algorithm uses machine learning to determine the content that users are most likely to participate in and provide them with more content by finding similar videos, or videos liked by people with similar user preferences. When users open TikTok for the first time, they will see 8 popular videos with different trends, music and themes. After that, the algorithm will continue to provide users with 8 updated videos based on the videos they watch or participate in and what the users are doing.
The algorithm recognizes videos that are similar to the video watched by the user based on video information (which may include details such as subtitles, tags, or sounds). It also takes into account user device and account settings, including data such as language preferences, country/region settings, and device type.
When TikTok has collected enough user data, the app can map user preferences based on similar users’ preferences and group them into “clusters”. At the same time, it also groups videos based on similar themes, such as “basketball” or “rabbit”.
Using machine learning, TikTok’s algorithm can provide users with videos based on their proximity to other user clusters and their favorite content. TikTok’s logic aims to avoid redundant content that may annoy users, such as seeing multiple videos with the same music or multiple videos from the same creator.
TikTok admits that the reason why it is able to lock users’ preferences so effectively means that its algorithms can generate “information cocoons” and strengthen users’ existing preferences instead of showing them more diverse content and broadening their horizons. Or provide them with uninteresting content.
TikTok said that they are studying “information cocoons”, including how long they can last and how users encounter them in order to better break them if necessary. Since the “information cocoon room” reinforces conspiracy theories, hoaxes, and other misinformation, TikTok’s product and policy team researches which accounts and video information (topics, tags, subtitles, etc.) may be related to misinformation.
Videos or creators linked to incorrect information are sent to the company’s global content reviewers so that TikTok can manage the content before it is distributed to users on the recommendation page.
TikTok also introduced the company’s latest initiatives in data, privacy and security. The company stated that it is trying to classify and prevent incidents on its platform before they occur, by detecting patterns of problems before they spread.
TikTok Chief Security Officer Roland Cloutier (Roland Cloutier) said that it plans to hire more than 100 data, security and privacy experts in the United States by the end of the year. He also said that the company will establish a monitoring, response and investigation response center in Washington, DC to actively detect and respond to critical events in real time.
Beckman said that TikTok’s transparency efforts are to position the company as a leader in Silicon Valley. He said: “We want to take the lead and show more how this app works. For us, we are newcomers and we want to do this because we have nothing to hide. We and the legislators The more conversations and meetings they have, the more they know about our products and the more satisfied they feel.”
For more such interesting article like this, app/softwares, games, Gadget Reviews, comparisons, troubleshooting guides, listicles, and tips & tricks related to Windows, Android, iOS, and macOS, follow us on Google News, Facebook, Instagram, Twitter, YouTube, and Pinterest.