Learning to rank music tracks using triplet loss

Abstract

Most music streaming services rely on automatic recommendation algorithms to exploit their large music catalogs. These algorithms aim at retrieving a ranked list of music tracks based on their similarity with a target music track. In this work, we propose a method for direct recommendation based on the audio content without explicitly tagging the music tracks. To that aim, we propose several strategies to perform triplet mining from ranked lists. We train a Convolutional Neural Network to learn the similarity via triplet loss. These different strategies are compared and validated on a large-scale experiment against an auto-tagging based approach. The results obtained highlight the efficiency of our system, especially when associated with an Auto-pooling layer.

References in corpus (8)

• FaceNet: A Unified Embedding for Face Recognition and Clustering
• In Defense of the Triplet Loss for Person Re-Identification
• Working hard to know your neighbor's margins: Local descriptor learning loss
• Automatic tagging using deep convolutional neural networks
• End-to-end learning for music audio tagging at scale
• Sampling Matters in Deep Embedding Learning
• Explaining Deep Convolutional Neural Networks on Music Classification
• Large-Scale Music Annotation and Retrieval: Learning to Rank in Joint Semantic Spaces