Dreamus Company Develops Innovative Technology That Automatically Creates 'Plee' When Mood Is Entered

Development of AI Language Model-Based Music Recommendation Technology
Technology to Automatically Create Playlists via Natural Language Search
To Be Applied to Some Services of Music Platform FLO Within the Year

Dreamus Company, which operates the music platform FLO, announced on the 13th that it has independently developed a music recommendation technology based on an artificial intelligence (AI) language model. By leveraging the company's AI technological capabilities and vast music data, this AI technology automatically generates playlists when users describe the mood of the songs they want to listen to. It will be applied to some of FLO's services within the second half of this year.

The technology newly developed by Dreamus Company utilizes a ‘Joint Embedding Architecture’ that simultaneously learns audio and text, enabling AI to generate playlists matching the text through natural language search. This allows users to receive playlist recommendations for their desired mood simply by entering natural language. For example, if a user says, ‘powerful idol songs to listen to during aerobic exercise,’ the system creates a playlist composed of exciting idol songs suitable for workouts.

Previously, when searching with text on the platform, there was a limitation where only songs mentioning the searched words in the title or lyrics, or already registered playlists, were provided using a search engine. To address the lack of natural language technology data, the technology was implemented by matching audio and text through AI learning. Instead of just listening to playlists like ‘good music for commuting’ created by the curation team or other users when searching for ‘commute,’ users can receive playlists tailored to specific situations and moods, such as ‘songs that calmly comfort you when you feel gloomy going to work on a holiday.’

Kwon Oh-beom, head of the recommendation technology unit at Dreamus Company, said, “We plan to enhance recommendation curation by utilizing the newly developed AI music recommendation technology and improve the quality of users’ music listening experiences in various situations.”