Irene Cannistraci

Postdoctoral Researcher, ETH Zurich, CS Department

Postdoctoral Researcher and Lecturer at ETH Zurich in the CS department, working on representation learning, multimodal machine learning, foundation models, and post-training with applications in healthcare. I received my PhD with honors from Sapienza University in 2025, and my work has been recognized with an ICLR 2024 Spotlight (top 5%). I am an ELLIS member and co-organizer of the UniReps workshop series at NeurIPS and the ELLISxUniReps speaker series.

Feel free to reach out at irene.cannistraci[at]inf.ethz.ch

Latest News & Updates

Dec 2025

publication

Two new works available on arXiv!

Two new preprints submitted: 'Structure is Supervision: Multiview Masked Autoencoders for Radiology' and 'You Only Train Once: Differentiable Subset Selection for Omics Data'.

Dec 2025

organization

UniReps 2025 Workshop Co-Organizer and Panel Moderator

Co-organized and moderated the panel session of UniReps at NeurIPS2025 with Sara Hooker, Ahmad Beirami, and Meenakshi Khosla.

Nov 2025

award

ELLIS Member

Proud to announce my membership in the ELLIS Society, joining a community of leading European AI researchers.

May 2025

organization

Launched the ELLISxUniReps Speaker Series

Monthly event exploring how and why neural models develop similar internal representations, and what this means for learning, alignment, and reuse.

Selected Publications

View all publications

TOAST: Transformer Optimization using Adaptive and Simple Transformations

Irene Cannistraci, Simone Antonelli, Emanuele Palumbo, Thomas Sutter, Emanuele Rodolà, Bastian Rieck, Julia Vogt

Under Review 2025

View paper

From Bricks to Bridges: Product of Invariances to Enhance Latent Space Communication

Irene Cannistraci, Luca Moschella, Marco Fumero, Valentino Maiorca, Emanuele Rodolà

ICLR 2024 2024 Spotlight · Top 5%

View paper