Multi-Task Music Representation Learning from Multi-Label Embeddings
Authors:
- 1 September 2019
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Abstract
This paper presents a novel approach to music representation learning. Triplet loss based networks have become popular for representation learning in various multimedia retrieval domains. Yet, one of the most crucial parts of this approach is the appropriate selection of triplets, which is indispensable, considering that the number of possible triplets grows cubically. We present an approach to harness multi-tag annotations for triplet selection, by using Latent Semantic Indexing to project the tags onto a high-dimensional space. From this we estimate tag-relatedness to select hard triplets. The approach is evaluated in a multi-task scenario for which we introduce four large multi-tag annotations for the Million Song Dataset for the music properties genres, styles, moods, and themes.Keywords
All Related Versions
This publication has 17 references indexed in Scilit:
- Music Retrieval and Recommendation2015
- FaceNet: A unified embedding for face recognition and clustering2015
- Improved music feature learning with deep neural networks2014
- A survey of music similarity and recommendation from music context dataACM Transactions on Multimedia Computing, Communications, and Applications, 2013
- Feature learning and deep architectures: new directions for music informaticsJournal of Intelligent Information Systems, 2013
- Social Tagging and Music Information RetrievalJournal of New Music Research, 2008
- A Large-Scale Evaluation of Acoustic and Subjective Music-Similarity MeasuresComputer Music Journal, 2004
- Artist detection in music with Minnowmatch2002
- A music similarity function based on signal analysis2001
- Indexing by latent semantic analysisJournal of the American Society for Information Science, 1990