Kakao Concludes 3rd Melon Music Recommendation Competition 'Kakao Arena'… Participants Up 80% vs. 1st Edition

Kakao hosts 'Kakao Arena', a machine learning problem-solving competition platform, every year. This year's 3rd competition was themed 'Music Recommendation for Melon Users' and involved developing models that generate similar playlists based on input playlist titles, tags, and track lists. A total of 364 teams and 1,143 individuals participated in this competition, representing an 80% increase from the 1st competition and a 120% increase from the 2nd competition. Participants experimented with various models such as BERT4REC, MMCF, and DAE, preparing for the competition after work hours while balancing their jobs. Kakao plans to apply the results of this competition to its services and intends to continue contributing to the development of the AI ecosystem by releasing data from major services like blogs and news, and making competition outputs open source.