Welcome!
About
I am a professor for the theory of machine learning in Julius-Maximilians-Universität Würzburg, currently focusing on explainable AI. I am the principal investigator of the NIM-ML project, funded by ANR. I am also an ELLIS member. Until March 2024, I was an associate professor in the Probability and Statistics team from the J. A. Dieudonné laboratory of Université Côte d'Azur, and part of Maasai, an Inria team located in Sophia-Antipolis. Previously, I was a postdoctoral researcher in the Max Planck Institute for Intelligent Systems (Tübingen, Germany) working with Ulrike von Luxburg. Even before, I was a PhD student in the Sierra Inria team in Paris, under the supervision of Sylvain Arlot and Gérard Biau.
News
April 2: new preprint: we show that CAM-based methods can provably highlight unseen parts of an image
April 4: talking at the Theory of Interpretable AI Seminar
May 2-4: AISTATS 2024 in Valencia
June 17: talking at the Journées Fondements Maths de l'IA (Paris)
Past news
April 3-5: MASCOT-NUM 2024 in Giens
March 27-28: HyCHA 2024 in Saclay
February 9: talk at the SAMM seminar
February 7: new preprint. What happens when classical post-hoc interpretability methods meet transformers?
November 29: a new preprint, in which we ask ourselves whether larger ensembles are always better?
November 30 - December 1: second edition of the NWI workshop
November 24: Sophia Summit
November 6: new preprint: we propose FRED, a new explainer for NLP applications with some guarantees
October 11-12: 4th Inria - DFKI workshop (Berlin)
July 23-29: in Honolulu for ICML, presenting two papers:
On the Robustness of Text Vectorizers (Samuel has a nice Twitter thread summing up the paper)
July 3-7: in Brussels for the 54th Journées de Statistiques de la SFdS
June 22: talk at the ASCAI online seminar
June 5: talk in Centrale Marseille
June 2: new preprint: what happens when everyone gets recourse?
May 22 - June 4 visiting the TML group in Tübingen
April 25-27: in Valencia for AISTATS, we presented an in-depth analysis of Anchors for text data
April 18: talking at the Machine Learning and Signal Processing seminar in ENS Lyon
March 13: talking at the GdR-ISIS (CNAM, Paris)
February 2-3: keynote speaker at the SMPGD 2023 conference (Ghent)
January 18: Martin Charachon's PhD defense (Centrale-Supélec)
January 6: talking at the Institute of Statistical Mathematics (Tokyo)
November 23-25, 2022: Sophia Summit, participating to a panel on Explainable AI
November 17-18, 2022: organizing the 1st Nice Workshop on Interpretability
November 9-11, 2022: visiting Korteweg-de Vries Institute for Mathematics (Amsterdam)
October 3-8, 2022: talking at the Non-Linear and High Dimensional Inference workshop (Paris)
September 5-7, 2022: talking at the ELISE Theory Workshop on Machine Learning Fundamentals (Sophia-Antipolis)
July 25, 2022: (virtual) talk at the Simons
July 12, 2022: Jonas Wacker's PhD defense
June 2022: new preprint: an in-depth analysis of Anchors, a post-hoc explainability method
June 2022: SMACE got accepted @ECML, congrats to Gianluigi on his first paper!
May 2022: Charbel Yahchouchi started his internship
May 29-June 4, 2022: I visited the TML group in Tuebingen
April 4-5, 2022: giving a talk in Université Paul Sabatier (Toulouse)
March 2022: Arthur Assad is starting his internship, welcome!
March 10-11, 2022: visiting University of Copenhagen, thanks Simon Bartels for the invite!
March 28-30, 2022: presenting How to scale hyperparameters for quickshift image segmentation at AISTATS 2022
February 2022: new preprint: scaling the hyperparameters for quickshift
February 1, 2022: talk in Eurecom
November 2021: I attended Sophia Summit
November 2021: new preprint: we propose SMACE, a new method for the explainability of composite AI systems
October 11: talk at the GdR ISIS (Paris)
October 7: talk in the LABRI (Bordeaux)
September 2021: Mariana succesfully defended her internship, with the top grade. Congrats!
July 2021: the JCJC project NIM-ML, whose main goal is to develop new interpretability methods, received funding from ANR! Job offers are coming shortly.
June 10, 2021: chairing the Interpretability and Fairness of Machine Learning session of the Journées de Statistiques
May 2021: our paper What does LIME really see in images? has been accepted to ICML 2021!
May 6, 2021: talk to the Orsay Probability and Statistics seminar and to the Artificial Intelligence and Geometry seminar
April 2021: Gianluigi Lopardo and Mariana Chaves are starting their internships, welcome!
April 14, 2021: I will be presenting our paper An Analysis of LIME for Text Data to the AISTATS 2021 conference
March 26, 2021: talk to the Maasai seminar
Contact
Office: Professorship for the Theory of Machine Learning
John-Skilton-Straße 8a
97074 Würzburg
GERMANY
Email: damien.garreauatuni-wuerzburg.de (replace at by @)
Phone: +49 931 31-89206