PhD Thesis

Oramas, S. (2017). Knowledge Extraction and Representation Learning for Music Recommendation and Classification. PhD thesis, Universitat Pompeu Fabra, Barcelona, Spain.

Short Abstract

In this thesis, we address the problems of classifying and recommending music present in large collections. We focus on the semantic enrichment of descriptions associated to musical items (e.g., artists biographies, album reviews, metadata), and the exploitation of multimodal data (e.g., text, audio, images). To this end, we first focus on the problem of linking music-related texts with online knowledge repositories and on the automated construction of music knowledge bases. Then, we show how modeling semantic information may impact musicological studies and helps to outperform purely text-based approaches in music similarity, classification, and recommendation. Next, we focus on learning new data representations from multimodal content using deep learning architectures, addressing the problems of cold-start music recommendation and multi-label music genre classification, combining audio, text, and images. We show how the semantic enrichment of texts and the combination of learned data representations improve the performance on both tasks.

Datasets

Knowledge bases

Software

  • ELVIS System that integrates different entity linking tools, enriching their output and providing high confident entity disambiguations. https://github.com/sergiooramas/elvis
  • TARTARUS System to perform and evaluate deep learning experiments on classification and recommendation from different data modalities and their combination. https://github.com/sergiooramas/tartarus
  • MEL API and demo website for a Music Entity Linking system that disambiguate musical entities to MusicBrainz. http://mel.mtg.upf.edu

Journal Publications

Peer-reviewed Conference Papers

Other conference presentations

Tutorials and Challenges