VK unveils unified AI model that builds recommendations from user data across its apps
VK has introduced a new AI model that aggregates user activity data from its various services into a single neuro-profile, improving content recommendations and ad effectiveness. The technology is already deployed in VK Video, VK Clips, and VKontakte, boosting watch time in VK Clips by 5.5%.
VK presented a new AI model that collects data on user actions from different company services and combines them into a unified neuro-profile. This allows algorithms to better understand what is interesting to a particular user. The technology is already used in VK Video, VK Clips, and VKontakte, helping to recommend content more accurately and increase advertising efficiency. After its implementation, the time spent watching content in VK Clips increased by 5.5%.
Source: Rusbase (RB.RU) —
original