Scholar
Joan Serrà
Google Scholar ID: sZLj96sAAAAJ
Sony AI
Citations & Impact
Citations
6,277
H-index
32
i10-index
63
Publications
20
Co-authors
121
list available
Publications
Resume (English only)
Academic Achievements
- Involved in several research projects, co-invented over 20 patents, and co-authored over 150 publications, many of them highly cited and/or in top tier venues. Recent publications include:
- - 'Towards blind data cleaning: a case study in music source separation' (preprint 2025)
- - 'Automatic music sample identification with multi-track contrastive learning' (preprint 2025)
- - 'Leveraging Whisper embeddings for audio-based lyrics matching' (preprint 2025)
- - 'Attribution-by-design: ensuring inference-time provenance in generative music systems' (preprint 2025)
- - 'System and method for attributing an output of a generative artificial intelligence (AI) system' (patent 2025)
- - 'Enhancing neural audio fingerprint robustness to audio degradation for music identification' (ISMIR 2025)
- - 'A comprehensive real-world assessment of audio watermarking algorithms: will they survive neural codecs?' (INTERSPEECH 2025)
- - 'Large-scale training data attribution for music generative models via unlearning' (NeurIPS 2025)
- - 'Supervised contrastive learning from weakly-labeled audio segments for musical version matching' (ICML 2025)
Research Experience
- Machine learning researcher at Telefónica R&D (2015-2019); AI researcher and manager at Dolby Laboratories (2019-2024).
Education
- MSc and PhD: 2006-2011 in machine learning for audio at the Music Technology Group of Universitat Pompeu Fabra; Postdoc: 2011-2015 in artificial intelligence at IIIA-CSIC.
Background
- Research interests: machine learning, with a focus on audio and multimedia analysis, synthesis, and retrieval. Brief bio: staff research scientist and team lead at Sony AI since 2024.
Miscellany
- Occasionally acts as reviewer or area chair for some venues (provided articles are free access/charge), and gives talks and lectures on subjects of interest, lately mainly related to representation learning and generative modeling.
