Laura Ibáñez-Martínez is a PhD student and researcher at the Music Technology Group (MTG), Universitat Pompeu Fabra (UPF). Her work focuses on controllable music generation from disentangled representations, and she carries out her research under the supervision of Martín Rocamora and Xavier Serra.
Research Interests
- Controllability: developing generative music systems that support fine-grained, time-varying control, with a focus on timbre transfer and MIDI-to-audio generation.
- Representation Learning & Disentanglement: learning disentangled representations of musical structure and timbre, enabling predictable control over individual musical attributes.
- Interpretability: analysing what musical information is encoded in learned representations through probing methods beyond standard downstream evaluation.
- Adaptability & Usability: investigating how generative music systems can be adapted to personal data under realistic constraints and integrated into real-world music production and creative workflows.
Publications
- Arhan Vohra, Choenden Kyirong, Laura Ibáñez-Martínez, Martín Rocamora. “Ghost in the Encoder: Decodable Artist Identity Representations in Lyrics-to-Song Generation.” International Society for Music Information Retrieval Conference (ISMIR), 2026.
- Laura Ibáñez-Martínez, Roser Batlle-Roca, Xavier Serra, Martín Rocamora. “MusGU+: Toward a Musician-Centered Evaluation Framework and Discovery Tool for Generative Music AI.” Conference on AI Music Creativity (AIMC), 2026. [Paper] [Repository] [Website]
- Laura Ibáñez-Martínez, Chukwuemeka Nkama, Andrea Poltronieri, Xavier Serra, Martín Rocamora. “Evaluating Disentangled Representations for Controllable Music Generation.” IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026. [Paper] [Repository]
- Roser Batlle-Roca, Laura Ibáñez-Martínez, Xavier Serra, Emilia Gómez, Martín Rocamora. “MusGO: A Community-Driven Framework for Assessing Openness in Music-Generative AI.” International Society for Music Information Retrieval Conference (ISMIR), 2025. [Paper] [Repository] [Website]
Professional Experience
- Music Technology Group (MTG), UPF — PhD Researcher (Barcelona, Spain, 2024–present)
- PIXELYNX — Machine Learning Engineer (California, US – hybrid from Berlin, Germany, 2023–2024)
- AudioShake — Machine Learning Engineer, freelance (California, US – remote, 2023)
- Pixtunes GmbH — Junior Machine Learning Engineer (Berlin, Germany, 2019–2021)
Education
- Universitat Pompeu Fabra (UPF) — PhD in Information and Communication Technologies (2024–present)
-
Universidad Nacional de Educación a Distancia (UNED) —
M.Sc. in Artificial Intelligence Research (2021–2023)
- Laura Ibáñez-Martínez. “MIDI-AudioLDM: MIDI-Conditional Text-to-Audio Synthesis Using ControlNet on AudioLDM.” Master’s thesis, 2023. [Thesis]
-
Universitat Politècnica de Catalunya (UPC) —
B.Sc. in Industrial Electronics and Automatic Control Engineering (2013–2019)
- Erasmus+ Programme, Technische Universität Berlin (2019)
- Escola Superior de Música de Catalunya (ESMUC) — Bachelor of Music – Classical and Contemporary Music Performance, Guitar (2013–2017)
Festivals & Hackathons
Talks
- “Explorando IA con prácticas artísticas” — Sala de Arte Joven, as part of the exhibition Materia Afectiva (2025)
- “IA y Música: MIDI-AudioLDM” — Volumens Festival (2023)
Other AI-related Projects
- Materia Afectiva: Hotlines — AI conversational agent developed for the exhibition (2025). [Press]
- Cellar Song for Five Voices — AI-generated text, voice and image (2021). [Project]
- THISPERSONDOESNOTEXIST — AI-generated image (2021). [Project]
- bod [包家巷] for PW-Magazine — AI-generated image (2020). [Project]