PeopleofSuperintelligence

Researcher

Pascal Frossard

Pascal Frossard is a full professor at EPFL and an IEEE Fellow, and his AI work includes two papers on image perturbations that fool deep neural networks. He earned EPFL degrees in 1997 and 2000 and worked at IBM's TJ Watson Research Center from 2001 to 2003. In 2003 he joined EPFL and founded the LTS4 laboratory. With co-authors, he wrote DeepFool and universal adversarial perturbations, which were published at CVPR 2016 and CVPR 2017. As of October 2026, he is Unit Director of ELLIS Unit Lausanne.

Works
934
Citations
28,368
h-index
61

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 1997–2000

    EPFL degrees in electrical engineering

    He earns an M.Sc. in 1997 and a PhD in 2000, both in electrical engineering at EPFL.

  2. 2001–2003

    IBM research before EPFL

    At IBM's TJ Watson Research Center he works from 2001 to 2003. In 2003 he joins EPFL and founds the LTS4 laboratory.

  3. 2003–2008

    Early SNSF and IBM awards

    In 2003 the Swiss National Science Foundation awards him a professorship. IBM follows with a Faculty Award in 2005 and an Exploratory Stream Analytics Innovation Award in 2008.

  4. 2016–2017

    Adversarial perturbation papers

    Written with Seyed-Mohsen Moosavi-Dezfooli and Alhussein Fawzi, DeepFool computes perturbations that fool deep networks, in the CVPR 2016 proceedings. Universal adversarial perturbations, written with Seyed-Mohsen Moosavi-Dezfooli, Alhussein Fawzi and Omar Fawzi, cause natural images to be misclassified with high probability, in the CVPR 2017 proceedings.

  5. 2016–2017

    Best paper and Google awards

    He wins an IEEE Signal Processing Magazine Best Paper Award in 2016. In 2017 he wins a Google Faculty Award.

  6. 2020

    Full professor and ELLIS unit

    He is named a full professor at EPFL in 2020. Among the founding units of the ELLIS network, the Lausanne unit is established in 2020.

  7. 2023–2025

    Co-directing Swiss AI centers

    In 2023 he becomes academic co-director of the EPFL AI Center, in 2024 of the Swiss Data Science Center, and in 2025 of the Swiss National AI Institute.

  8. As of October 2026

    Leading ELLIS Unit Lausanne

    As Unit Director of ELLIS Unit Lausanne, he is also a full professor at EPFL and an IEEE Fellow.

Sources and credits

Photo: We have no published photo of Pascal Frossard yet. Send us a free-licensed photo.

  1. Short bio - LTS4 - EPFL epfl.ch. Checked 15 October 2026.
  2. Pascal Frossard Named as Full Professor (STI) - EcoCloud ecocloud.epfl.ch. Checked 15 October 2026.
  3. Prof. Pascal Frossard - LTS4 - EPFL epfl.ch. Checked 15 October 2026.
  4. DeepFool: A Simple and Accurate Method to Fool Deep Neural Networks - CVPR 2016 Open Access Repository openaccess.thecvf.com. Checked 15 October 2026.
  5. [1511.04599] DeepFool: a simple and accurate method to fool deep neural networks arxiv.org. Checked 15 October 2026.
  6. Universal Adversarial Perturbations - CVPR 2017 Open Access Repository openaccess.thecvf.com. Checked 15 October 2026.
  7. [1610.08401] Universal adversarial perturbations arxiv.org. Checked 15 October 2026.
  8. ELLIS Unit Lausanne: Advancing AI at the Intersection of Science, Health, and Society ellis.eu. Checked 15 October 2026.
  9. Opening of the new EPFL AI Center: A Hub for AI in Switzerland - EPFL actu.epfl.ch. Checked 15 October 2026.