Skip to main content
Antti Koskela

Antti Koskela

Machine Learning Researcher

Espoo, Finland

New Image

Biography

I reveived my PhD in Mathematics from the University of Innsbruck, Austria, where I focused on numerical methods for time-dependent partial differential equations. During a post doc period at the University of Helsinki, I started research on privacy-preserving machine learning, in particular on differential privacy. I joined Nokia Bell Labs in December 2021 and my research interests are around deep learning methods and differential privacy.

I regularly review for conferences such as NeurIPS (top reviewer 2023), ICML, ICLR and AISTATS.

Education
  • PhD, Mathematics, University of Innsbruck, Austria, 2014.
  • MSc, Engineering Physics and Mathematics, Aalto University, Finland, 2010.
Selected Articles and Publications
  • Practical differentially private hyperparameter tuning with subsampling
    A Koskela, T Kulkarni
    Thirty-seventh Conference on Neural Information Processing Systems 2023
  • Improving the privacy and practicality of objective perturbation for differentially private linear learners
    R Redberg, A Koskela, YX Wang
    Thirty-seventh Conference on Neural Information Processing Systems 2023
  • Individual Privacy Accounting with Gaussian Differential Privacy
    A Koskela, M Tobaben, A Honkela
    International Conference on Learning Representations 2023
  • Numerical Accounting in the Shuffle Model of Differential Privacy
    A Koskela, M Heikkilä, A Honkela
    Transactions on Machine Learning Research 2023
Memberships

On this page