Prof. Dr. Grégoire Montavon
Research Junior Group Lead
Prof. Dr. Grégoire Montavon is a Guest Professor at the Freie Universität Berlin and a Research Group Lead in the Berlin Institute for the Foundations of Learning and Data (BIFOLD). He received a Masters degree in Communication Systems from École Polytechnique Fédérale de Lausanne in 2009, and a Ph.D. degree in Machine Learning from the Technische Universität Berlin in 2013.
His current research focuses on methods of explainable AI (XAI) for deep neural networks and unsupervised learning, and on closing the gap between existing XAI methods and practical desiderata. This includes using XAI to build more trustworthy machine learning models and using XAI to extract actionable insights from complex datasets.
Jointly with his colleagues, he contributed to Layer-Wise Relevance Propagation (LRP), an efficient method for explaining the predictions of large deep neural networks. He and his co-authors also contributed to the "Neuralization-Propagation" framework which rewrites popular unsupervised learning models as functionally equivalent neural networks for explainability purposes, and higher-order extensions of LRP (BiLRP and GNN-LRP) which enable the identification of joint features contributions in models with product structures.
- 2013 Dimitris N. Chorafas Award
- 2020 Pattern Recognition Best Paper Award
- 2022 Digital Signal Processing Best Paper Award
- Explainable AI
- Machine Learning
- Data Science
Scholar, European Laboratory for Learning and Intelligent Systems (ELLIS), 2021–now
Thomas Schnake, Farnoush Rezaei Jafaria, Jonas Lederer, Ping Xiong, Shinichi Nakajima, Stefan Gugler, Grégoire Montavon, Klaus-Robert Müller
Towards Symbolic XAI -- Explanation Through Human Understandable Logical Relationships Between Features
Jacob Kauffmann, Jonas Dippel, Lukas Ruff, Wojciech Samek, Klaus-Robert Müller, Grégoire Montavon
The Clever Hans Effect in Unsupervised Learning
Farnoush Rezaei Jafari, Grégoire Montavon, Klaus-Robert Müller, Oliver Eberle
MambaLRP: Explaining Selective State Space Sequence Models
Simon Letzgus, Klaus-Robert Müller, Grégoire Montavon
XpertAI: uncovering model strategies for sub-manifolds
Frederick Klauschen, Jonas Dippel, Philipp Keyl, Philipp Jurmeister, Michael Bockmayr, Andreas Mock, Oliver Buchstab, Maximilian Alber, Lukas Ruff, Grégoire Montavon & Klaus-Robert Müller
Explainable artificial intelligence in pathology
Publication Highlight – Pruning Clever-Hans strategies
Hidden Clever Hans effects can undermine the reliability of AI models. The paper “Preemptively pruning Clever-Hans strategies in deep neural networks” introduces a method that corrects biases in neural networks without prior knowledge of faulty features.
Call for XAI-Papers!
Two research groups associated with BIFOLD take part in the organization of the 2nd World Conference on Explainable Artificial Intelligence. Each group is hosting a special track and has already published a Call for Papers. Researchers are encouraged to submit their papers by March 5th, 2024.
DSP Best Paper Prize
BIFOLD researchers Prof. Klaus-Robert Müller, Prof. Wojciech Samek and Prof. Grégoire Montavon were honored by the journal Digital Signal Processing (DSP) with the 2022 Best Paper Prize. The DSP mention of excellence highlights important research findings published within the last five years.
2020 pattern recognition best paper award
A team of scientists from TU Berlin, Fraunhofer Heinrich Hertz Institute (HHI) and University of Oslo has jointly received the 2020 “Pattern Recognition Best Paper Award” and “Pattern Recognition Medal” of the international scientific journal Pattern Recognition. The award committee honored the publication “Explaining Nonlinear Classification Decisions with Deep Taylor Decomposition” by Dr. Grégoire Montavon and Prof. Dr. Klaus-Robert Müller from TU Berlin, Prof. Dr. Alexander Binder from University of Oslo, as well as Dr. Wojciech Samek and Dr. Sebastian Lapuschkin from HHI.
Using machine learning to combat the coronavirus
A joint team of researchers from TU Berlin and the University of Luxembourg is exploring why a spike protein in the SARS-CoV-2 virus is able to bind much more effectively to human cells than other coronaviruses. Google.org is funding the research with 125,000 US dollars.
An overview of the current state of research in BIFOLD
Since the official announcement of the Berlin Institute for the Foundations of Learning and Data in January 2020, BIFOLD researchers achieved a wide array of advancements in the domains of Machine Learning and Big Data Management as well as in a variety of application areas by developing new Systems and creating impactfull publications. The following summary provides an overview of recent research activities and successes.