Scholar
Sonia Laguna
Google Scholar ID: PljVnCQAAAAJ
PhD student, ETH Zürich
Citations & Impact
Citations
225
H-index
7
i10-index
5
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- October 2025: Speaker at the ETH-UN Forum and Selected Project at the ETH-UN Incubator; June 2025: Four first author publications accepted and to be presented at ICML, MLHC, and IPCAI; April 2025: Four workshop papers accepted at ICLR 2025 with an oral (Best Paper Award) contribution; April 2025: Invited speaker at the Microsoft Cambridge ML Seminar Series; April 2025: Participated in the Spring into Quant 2025 G-Research Program; February 2025: Started a research stay at University of Cambridge.
Research Experience
- During PhD, visiting student at Cambridge University with Prof. Mihaela Van der Shaar, working on alignment and interpretability of LLMs; Research Intern and Student Researcher at Google, developing 3D diffusion-based generative models in the AR&VR team; Co-leader of CSNOW, Computer Science Network of Women at ETH.
Education
- PhD: ETH Zurich, Machine Learning, supervised by Prof. Julia Vogt and Prof. Bernhard Schölkopf; MSc: ETH Zurich, Department of Information Technology and Electrical Engineering; Exchange semester at Harvard University, working on 3D generative models for super-resolution of MR images; BSc: Universidad Carlos III de Madrid, Biomedical Engineering, spent one year at Georgia Institute of Technology, and interned at ETH Zurich as an Amgen Scholar.
Background
- Research interests include generative models, foundation models, representation learning, and machine unlearning. Currently a PhD student in Machine Learning at ETH Zurich, supervised by Prof. Julia Vogt and Prof. Bernhard Schölkopf. Research focuses on developing generative models and gaining better understanding and control of them (diffusion, VAEs, LLMs) through their representations, as well as solving problems on interpretability of machine learning methods. Recently interested in machine unlearning, exploring how models can selectively forget data while retaining general knowledge.
Miscellany
- Interests include generative models, foundation models, representation learning, and machine unlearning.
Co-authors