Summer internship positions in the Trustworthy Machine Learning group in 2025

1. Privacy-preserving machine learning (multiple positions)

Supervisor: Antti Honkela + PhD students and postdocs

Background: machine learning, mathematics, programming skills (Python, PyTorch and/or JAX)

Privacy-preserving machine learning studies the privacy properties of machine learning, and develops learning methods that can operate while guaranteeing privacy of the data subjects. These methods can be applied to solving predictive learning problems on private data but also for creating provably anonymised data sets.

Our work on developing privacy-preserving algorithms builds upon differential privacy, that allows mathematically provable privacy. We apply it in the context of various modern machine learning methods, including Bayesian methods, deep learning and federated learning.

In addition to developing privacy-preserving algorithms, we also develop and apply new privacy attacks to empirically test the privacy of machine learning methods.

In this project you will participate in developing, analysing and applying new privacy-preserving machine learning methods. Depending on your background and interests, the work will combine working on the mathematical theory of differential privacy, development of privacy-preserving algorithms or new privacy attacks, as well as implementation and application of the developed methods in different applications.

You will join an active community of ~10 researchers with diverse backgrounds working in the Research Programme in Privacy-Preserving and Secure AI at the Finnish Center for Artificial Intelligence FCAI. We are also affiliated with the DataLit project and the European Lighthouse on Secure and Safe AI (ELSA).

The topic is suitable for a Master's thesis topic.

Examples of related MSc theses done in the group:

Marlon Tobaben. Hyperparameters and neural architectures in differentially private deep learning. MSc Thesis, University of Helsinki, 2022.

Ossi Räisä. Differentially Private Metropolis–Hastings Algorithms. MSc Thesis, University of Helsinki, 2021.

Examples of our recent related papers:

Ossi Räisä, Stratis Markou, Matthew Ashman, Wessel P. Bruinsma, Marlon Tobaben, Antti Honkela, Richard E. Turner. Noise-Aware Differentially Private Regression via Meta-Learning. In Advances in Neural Information Processing Systems 38 (NeurIPS 2024) (2024).

Ossi Räisä, Joonas Jälkö, Antti Honkela. Subsampling is not Magic: Why Large Batch Sizes Work for Differentially Private Stochastic Optimisation. In Proceedings of the 41st International Conference on Machine Learning (ICML 2024) (2024).

Ossi Räisä, Joonas Jälkö, Samuel Kaski, and Antti Honkela. Noise-Aware Statistical Inference with Differentially Private Synthetic Data. In Proceedings of the 26th International Conference on Artificial Intelligence and Statistics (AISTATS 2023) (2023).

Marlon Tobaben, Aliaksandra Shysheya, John F Bronskill, Andrew Paverd, Shruti Tople, Santiago Zanella-Beguelin, Richard E Turner, and Antti Honkela. On the Efficacy of Differentially Private Few-shot Image Classification. Transactions on Machine Learning Research, December 2023.

Antti Koskela, Marlon Tobaben, and Antti Honkela. Individual Privacy Accounting with Gaussian Differential Privacy. In Proceedings of the 11th International Conference on Learning Representations (ICLR 2023) (2023).