A Machine Learning based Music Retrieval and Recommendation System

A Machine Learning based Music Retrieval and Recommendation System Naziba Mostafa, Yan Wan, Unnayan Amitabh, Pascale Fung Human Language Technology Center Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong nmostafa@connect.ust.hk, ywanad@connect.ust.hk, uamitabh@connect.ust.hk, pascale@ece.ust.hk Abstract In this paper, we present a music retrieval and recommendation system using machine learning techniques. We propose a query by humming system for music retrieval that uses deep neural networks for note transcription and a note-based retrieval system for retrieving the correct song from the database. We evaluate our query by humming system using the standard MIREX QBSH dataset. We also propose a similar artist recommendation system which recommends similar artists based on acoustic features of the artists’ music, online text descriptions of the artists and social media data. We use supervised machine learning techniques over all our features and compare our recommendation results to those produced by a popular similar artist recommendation website. Keywords: query by humming, similar artist recommendation, music information retrieval, machine learning 1. Introduction Faster computational speed and increasing number of on- line users have resulted in a dramatic increase in music consumption. It is getting more and more difficult for the general public, especially non-experts, to find and retrieve music from the millions of songs available online. A lot of research is being done these days to find efficient mu- sic retrieval and recommendation methods. One music re- trieval method that is gaining a lot of popularity these days due to its convenient usage is query by humming, which is a content-based music retrieval method that can retrieve melodies using users’ hummings as queries. This allows users to find old songs that they only remember the tune of or retrieve obscure songs heard in public places. The Mu- sic Information Retrieval (MIR) community has also been doing a lot of work on automatic recommendation systems ranging from the content-based methods to social tagging and similarity networks (Cohen and Fan, 2000; Hong et al., 2008). One of the key research topics in this area that has gained a lot of traction is automatic similar artist recom- mendation. Currently, there are several musical retrieval and similar artist recommendation apps. There are apps such as Sound- Hound, MusixMatch etc, that can retrieve songs using hum- ming as a query, and websites such as All Music Guide (AMG)1 and last.fm2 that give similar artist recommenda- tion. However, accuracy and efficiency of these music re- trieval and recommendation systems still leave a lot of room for improvement. Therefore, we are planning to create a holistic music retrieval and recommendation system using machine learning techniques. The biggest challenges of a query by humming system include i) queries sung by users often vary from the ac- tual melody in pitch, tempo etc. so the melodic similarity matching must be done at a more abstract level in order to get meaningful results, ii) background noise is often present in users’ queries which also makes it harder to identify …

Content truncated to 500 words