ELSA Workshop on Privacy-Preserving Machine Learning

Blue and white calendar in ELSA style with little yellow icons representing two people talking. Text in image: ELSA Workshop Privacy-Preserving Machine Learning @ Bertinoro, Italy. March 16 - 21st 2025

Save the Date: March 17 – 21, 2025

ELSA Work Package 2 is inviting ELSA and ELLIS members to join them for an inspiring workshop on

  • privacy in machine learning
  • privacy in federated machine learning

Besides providing a space for discussion about recent developments, the workshop aims to be a space which enables collaboration amongst attendees.

About the event

What? Workshop for ELSA and ELLIS members

When? March 17 – 21, 2025

Where?  Bertinoro University Center (CEUB), Bertinoro, Italy

This workshop is inviting members of ELSA and the ELLIS network. Applications have closed. If you want to know about our events in advance, we recommend to follow us on LinkedIn or sign up for our newsletter.

Workshop Outline

The workshop will provide the framework for our attendees to exchange knowledge and ideas, to learn from each other, and to network.

Each day will consist of three to four sessions with coffee and lunch breaks in between. Please find the agenda below:

Monday
  • “The Promise and Pitfalls of Public Data in Private ML” – presentation by Gautam Kamath, University of Waterloo
  • “Privacy for Correlated Data” – presentation by Amartya Sanyal, University of Copenhagen
  • “The advantage of threat models in Privacy Preserving Machine Learning” – presentation by Giovanni Cherubin, Microsoft Research – Cambridge
  • (“ε, δ) Considered Harmful — Best Practices for Reporting Differential Privacy Guarantees” – presentation by Antti Honkela, University of Helsinki
  • Poster Session
  • Panel Discussion: DP Practice
Tuesday
  • “LLM’s unintended memories  (attacks)” – presentation by Yves-Alexandre de Montjoye, Imperial College London
  • “Non-biased Membership Inference Attacks Assessment on LLMs with Ex-Post Dataset Construction” – presentation by Nicolas Anciaux, INRIA
  • “Privacy of AI systems: the point of view of a Data Protection Authority   (regulation, DP practice)” – presentation by
    Nicolas Berkouk, CNIL
  • “Auditing Privacy Risks of LLM-Generated Synthetic Text” – presentation by Santiago Zanella-Beguelin, Microsoft Research – Cambridge
  • ELSA & friends – talk
  • Panel Discussion: Attacks and Auditing
Wednesday
  • “Purifying Approximate Differential Privacy with Randomized Postprocessing” – talk by Yu-Xiang Wang, University of California Santa Barbara
  • “Information-theoretic structures for privacy and fairness” – presentation by Catuscia Palamidessi, INRIA
  • “A QIF approach to hypothesis testing in differential privacy” – presentation by Annabelle McIver, Macquarie University
  • “An Extension of the Adversarial Threat Model in Quantitative Information Flow” – Parastoo Sadeghi, UNSW
  • Visit to Ravenna
Thursday
  • Interactive program for all participants (more details coming soon)
  • “Privacy in Machine Learning: The gap between theory and practice” – Natasha Fernandes, Macquarie University
  • Panel discussion: Privacy Theory
Friday
  • “Workshop follow-up: Written report on challenges in privacy-preserving ML” – by Antti Honkela, University of Helsinki
  • Coffee & lunch
  • End of workshop

Travel easily: use the ELSA Mobility Fund

Travelling to our workshops is supported by our mobility program. Make sure to have a look at our Mobility Fund Website and check whether you are eligible! Check out the Mobility Fund.

If you have any questions regarding the mobility fund, please contact elsa-coordination@cispa.de.