Marco Federici

PhD candidate (advised by Patrick Forré)


Google scholar   Github   Twitter  

I am a PhD student at the Amsterdam Machine Learning Lab (AMLab) supervised by Patrick Forré.

My research focuses on how to use information theoretical approaches to improve model robustness when only little labled data is available and in the context of distribution shifts.

Selected Publications

  1. NeurIPS
    An Information-theoretic Approach to Distribution Shifts
    Federici, Marco, Tomioka, Ryota, and Forré, Patrick
    In 35th Conference on Neural Information Processing Systems (NeurIPS) Dec 2021
  2. ICLR
    Learning Robust Representations via Multi-View Information Bottleneck
    Federici, Marco, Dutta, Anjan, Forré, Patrick, Kushman, Nate, and Akata, Zeynep
    In 8th International Conference on Learning Representations, ICLR 2020 Dec 2020