AcousticRadioz: Dynamic artist-based radio using manifold learning embeddings

Raphael Shuhendler, M.Sc. student at the department of Industrial

7 June 2022, 14:00 PM, Room 206& via zoom

Abstract:
Streaming services have become a central platform for music consuming. Often, they rely on large archives on which playlists are created. While many playlists are handcrafted, automatic playlist generation is a rising star, as it lets people explore new music easily. This paper focuses on a content-based embedding process constructed using raw-audio features from AcousticBrainz, and used for generating artist-based radios. The novelty of our work lies in the creation of artist-based embedding process using a manifold learning technique, denoted by REF-DM; and by its scalability in terms of computational complexity. REF-DM mutually organizes the seed artist songs into several clusters as well as the rest of the songs in the dataset. The radio is dynamically generated using the distances in the embedded space, while also taking diversity into account. Empirical results are evaluated by the MELON playlist dataset, a novel public dataset for playlists. Our evaluations show that REF-DM followed by our proposed radio generation algorithm, yields a more accurate and diverse sequence when compared to other dimension reduction embedding techniques.

Bio:
Raphael Shuhendler, M.Sc. student (fast track program) at the Department of Industrial Engineering and management in Tel Aviv University, specializing in business intelligence and data science. Raphael holds a B.Sc. degree in Industrial Engineering from Tel Aviv University. His research is supervised by Dr. Neta Rabin