ICML 2025 Spotlight
Privacy Amplification by Structured Subsampling for Deep Differentially Private Time Series Forecasting
Privacy guarantees for forecasting models trained on structured time-series samples.
Oxford, UK
Machine learning research
Jed Guzelkabaagac
DPhil student · University of Oxford
I’m a doctoral student in the EIT CDT in Fundamentals of AI at the University of Oxford.
I’m interested in the mathematical foundations of machine learning, with a particular focus on generative models.
Previously at ETH, MPI-IS, and TUM.
Based atUniversity of OxfordDoctoral Training CentreFundamentals of AIEIT Centre for Doctoral Training
Selected research
Google Scholar ↗IROS 2025 · FMRD Workshop
Label-Efficient Grasp Joint Prediction with Point-JEPA
Self-supervised point-cloud representations for learning grasp joint angles from fewer labels.