Guillermo Ortiz-Jiménez

profile.jpg
gortizji⚛gmail.com

¡Hola! 👋 I’m a senior research scientist at Google DeepMind. My main area of research is responsible AI where I develop technologies to cultivate a healthier data ecosystem (such as SynthID) and to improve GenAI models. I am also interested in the science of deep learning: understanding why and how neural networks learn the way they do.

I obtained my PhD at EPFL🇨🇭 under the supervision of Pascal Frossard. My PhD research centered on the science of deep learning and investigated several phenomena related to the generalization and robustness of deep neural networks. During my PhD, I interned at Google Research in Zürich and visited Philip Torr’s lab at the University of Oxford🇬🇧 . Prior to that, I obtained my MSc. from TU Delft 🇳🇱 and my BSc. from Universidad Politécnica de Madrid 🇪🇸.

If you’re not sure how to pronounce my name you can watch this video. Here’s a quick tip: the double “ll” in Spanish sounds like a “y” and the “u” in “Gui” is silent.

news

Jan 2024 I started a new job at Google DeepMind.
Dec 2023 We presented Task Arithmetic in the Tangent Space as and Oral in NeurIPS 2024.
Dec 2023 I defended my PhD at EPFL 🎉!

selected publications

  1. On the difficulty of constructing a robust and publicly-detectable watermark
    Jaiden Fairoze ,  Guillermo Ortiz-Jimenez ,  Mel Vecerik ,  Somesh Jha ,  and  Sven Gowal
    In Artificial Intelligence and Statistics (AISTATS) , 2025
  2. LiNeS: Post-training layer scaling prevents forgetting and enhances model merging
    Ke Wang ,  Nikolaos Dimitriadis ,  Alessandro Favero ,  Guillermo Ortiz-Jimenez ,  Francois Fleuret , and 1 more author
    In International Conference on Learning Representations (ICLR) , 2024
  3. Localizing Task Information for Improved Model Merging and Compression
    Ke Wang ,  Nikolaos Dimitriadis ,  Guillermo Ortiz-Jimenez ,  Francois Fleuret ,  and  Pascal Frossard
    In International Conference on Machine Learning (ICML) , 2024
  4. Task Arithmetic in the Tangent Space: Improved Editing of Pre-Trained Models
    Guillermo Ortiz-Jimenez* ,  Alessandro Favero* ,  and  Pascal Frossard
    In Advances in Neural Information Processing Systems (NeurIPS) , 2023
  5. What can linearized neural networks actually say about generalization?
    Guillermo Ortiz-Jimenez ,  Seyed-Mohsen Moosavi-Dezfooli ,  and  Pascal Frossard
    In Advances in Neural Information Processing Systems (NeurIPS) , 2021