And “My Wave” learned to predict the development of the user’s musical tastes
Yandex Music rebranded for the first time since 2014; the logo and corporate identity of the service were updated. In addition, a number of new features have appeared. Thus, the smart recommendation system “My Wave” began to use deep neural models to understand how users’ musical interests develop. According to the creators of the service, this allows listeners to discover three times more new favorite artists than they find on their own. By analyzing the history of listening, the sequence of music tracks and other factors, “My Wave” is now able to grasp the direction of development of a person’s musical taste and predict his hobbies – this is similar to the work of YandexGPT, but in the context of music.
Key sections of the service have also been updated: the main screen of “Yandex Music” is now completely personalized, it now has a quick transition to “Collection”, listening history and “My Wave” for your favorite genres, and sections “In Style” or “Met” have been added in “My Wave”, where you can find performers whom the listener discovered thanks to the system of personal recommendations. And in the updated “Collection” section, a special “My Wave” for music collected by the user has started working: the system offers tracks from all music sections of the user’s collection, and also supplements them with new ones, similar in style and sound.
Yandex Music has long been more than just a huge catalog of tracks. Having launched “My Wave” in 2021, we wanted to learn how to accurately recommend to listeners what suits them, in order to save them from having to independently search for music that they will like. And we succeeded: according to almost half of Russian Internet users who listen to music online, we are more accurate than others in their musical preferences. Now we want to go further – using recommendations to make music an endless source of encouragement and inspiration.
Alexandra Sagalovich, head of Yandex Music