Large-Scale Weakly-Supervised Content Embeddings for Music Recommendation and Tagging

Qingqing Huang, Aren Jansen, Li Zhang, Daniel P. W. Ellis, Rif A. Saurous, John R. AndersonICASSP 2020 · OpenReview

Abstract

We explore content-based representation learning strategies tailored for large-scale, uncurated music collections that afford only weak supervision through unstructured natural language metadata and co-listen statistics. At the core is a hybrid training scheme that uses classification and metric learning losses to incorporate both metadata-derived text labels and aggregate co-listen supervisory signals into a single convolutional model. The resulting joint text and audio content embedding defines a similarity metric and supports prediction of semantic text labels using a vocabulary of unprecedented granularity, which we refine using a novel word-sense disambiguation procedure. As input to simple classifier architectures, our representation achieves state-of-the-art performance on two music tagging benchmarks.
BibTeX
@inproceedings{huang2020large,
  author = {Qingqing Huang and Aren Jansen and Li Zhang and Daniel P. W. Ellis and Rif A. Saurous and John R. Anderson},
  title = {{Large-Scale Weakly-Supervised Content Embeddings for Music Recommendation and Tagging}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2020},
  url = {https://openreview.net/forum?id=fM8hnN_ssCq},
}

Loading related content...

Figure