I am a Machine Learning researcher, currently pursuing a PhD at the École Polytechnique Fédérale de Lausanne (EPFL) with Prof. Pierre Vandergheynst. My main research interest is the modeling, analysis, and understanding of structured data. Data structured by networks, such as brain activity supported by neural connections, or manifolds, such as the temperature and wind fields on the Earth. To this end I am developing Deep Learning to leverage that structure, often modeled as a graph. In my research, I draw from theoretical insights and practical needs to develop principled methods—and strive for impactful applications by co-leading interdisciplinary research efforts. Besides research, I teach Data Science and Machine Learning (with graphs).
| CARVIEW |
Select Language
HTTP/2 200
content-type: text/html
etag: "2096650941-gzip"
last-modified: Mon, 20 Oct 2025 14:23:45 GMT
content-length: 15493
strict-transport-security: max-age=600; includeSubDomains
vary: Accept-Encoding
content-encoding: gzip
date: Mon, 29 Dec 2025 00:37:19 GMT
server: lighttpd/1.4.77
Michaël Defferrard (mdeff)
Michaël Defferrard
Michaël Defferrard
@mdeff
- Research on machine learning and graphs.
- Open science, open source, open data.
- Educator and mentor.
- Brass band musician.
Publications
GScholar S2 arXiv OpenReview ORCID DBLP EPFL-
Improving the Efficiency of Algorithmic Reasoning in Language Models via Reinforcement Learning,
Seungyong Moon, Michaël Defferrard, Corrado Rainone, Roland Memisevic, 2025.
[OpenReview] [ICLR-W'25] -
Grounding code understanding in step-by-step execution,
Michaël Defferrard, David W. Zhang, Corrado Rainone, Roland Memisevic, 2025.
[ICML'25 submitted] [ICLR'25 rejected] -
Exploring “dark-matter” protein folds using deep learning,
Zander Harteveld*, Alexandra Van Hall-Beauvais*, Irina Morozova, Joshua Southern, Casper Goverde, Sandrine Georgeon, Stéphane Rosset, Michaël Defferrard, Andreas Loukas, Pierre Vandergheynst, Michael M. Bronstein, Bruno E. Correia, Cell Systems, 2024.
[Cell Systems] -
Towards Bridging Classical and Neural Computation through a Read-Eval-Print Loop,
David W. Zhang, Michaël Defferrard, Corrado Rainone, Roland Memisevic, 2024.
[LLM-cognition@ICML] [DMLR@ICML] -
CodeIt: Self-Improving Language Models with Prioritized Hindsight Replay,
Natasha Butt, Blazej Manczak, Auke Wiggers, Corrado Rainone, David W. Zhang, Michaël Defferrard, Taco Cohen, ICML, 2024.
[arXiv] [ICML] [OpenReview] [AGI@ICLR] [poster] [code] -
Towards Self-Improving Language Models for Code Generation,
Michaël Defferrard, Corrado Rainone, David W. Zhang, Blazej Manczak, Natasha Butt, Taco Cohen, 2024.
[LLMAgents@ICLR] [GenAI4DM@ICLR] -
DE-HNN: An effective neural model for Circuit Netlist representation,
Zhishang Luo, Truong Son Hy, Puoya Tabaghi, Michaël Defferrard, Elahe Rezaei, Ryan M. Carey, Rhett Davis, Rajeev Jain, Yusu Wang, 2023.
[arXiv] [AISTATS] [code] -
NeuroSteiner: A Graph Transformer for Wirelength Estimation,
Sahil Manchanda, Dana Kianfar, Markus Peschl, Romain Lepert, Michaël Defferrard, 2023.
[arXiv] [DAC WIP poster] [AI2ASE@AAAI] -
Deep learning for classification of cortical rodent neuronal morphologies,
Lida Kanari, Stanislav Schmidt, Francesco Casalegno, Emilie Delattre, Jelena Banjac, Thomas Negrello, Ying Shi, Julie Meystre, Michaël Defferrard, Felix Schürmann, Henry Markram, 2022.
[bioRxiv] [EPFL] [data] [code] -
Leveraging topology, geometry, and symmetries for efficient Machine Learning,
Michaël Defferrard, PhD thesis, 2022.
[PDF] [EPFL] [slides] -
Deep sharpening of topological features for de novo protein design,
Zander Harteveld, Joshua Southern, Michaël Defferrard, Andreas Loukas, Pierre Vandergheynst, Michael M. Bronstein, Bruno E. Correia, Machine Learning for Drug Discovery Workshop at ICLR, 2022.
[OpenReview] -
ChebLieNet: Invariant spectral graph NNs turned equivariant by Riemannian geometry on Lie groups,
Hugo Aguettaz, Erik J. Bekkers, Michaël Defferrard, 2021.
[arXiv] [OpenReview] [latex] [code] -
RosettaSurf—A surface-centric computational design approach,
Andreas Scheck, Stéphane Rosset, Michaël Defferrard, Andreas Loukas, Jaume Bonet, Pierre Vandergheynst, Bruno E. Correia, PLOS Computational Biology, 2022.
[bioRxiv] [PLOS Computational Biology] [code] -
Learning to recover orientations from projections in single-particle cryo-EM,
Jelena Banjac, Laurène Donati, Michaël Defferrard, 2021.
[arXiv] [OpenReview] [latex] [website] [code] -
Simplicial Neural Networks,
Stefania Ebli, Michaël Defferrard, Gard Spreemann, Topological Data Analysis and Beyond workshop at NeurIPS, 2020.
[arXiv] [TDA@NeurIPS] [OpenReview] [latex] [poster] [data] [code] -
Connectome spectral analysis to track EEG task dynamics on a subsecond scale,
Katharina Glomb, Joan Rue Queralt, David Pascucci, Michaël Defferrard, Sebastien Tourbier, Margherita Carboni, Maria Rubega, Serge Vulliemoz, Gijs Plomp, Patric Hagmann, NeuroImage, 2020.
[bioRxiv] [NeuroImage] -
DeepSphere: a graph-based spherical CNN,
Michaël Defferrard, Martino Milani, Frédérick Gusset, Nathanaël Perraudin, ICLR, 2020.
[arXiv] [ICLR] [OpenReview] [latex] [slides] [video] [code] -
DeepSphere: towards an equivariant graph-based spherical CNN,
Michaël Defferrard, Nathanaël Perraudin, Tomasz Kacprzak, Raphael Sgier, Representation Learning on Graphs and Manifolds workshop at ICLR, 2019.
[arXiv] [RLGM@ICLR] [reviews] [latex] [poster] [code] -
DeepSphere: Efficient spherical Convolutional Neural Network with HEALPix sampling for cosmological applications,
Nathanaël Perraudin, Michaël Defferrard, Tomasz Kacprzak, Raphael Sgier, Astronomy and Computing, 2019.
[arXiv] [A&C] [reviews] [latex] [blog] [slides] [data] [code] -
Learning to Recognize Musical Genre from Audio: Challenge Overview,
Michaël Defferrard, Sharada P. Mohanty, Sean F. Carroll, Marcel Salathé, The Web Conference, 2018.
[arXiv] [TheWebConf] [latex] [track] [challenge] [slides] [data] [code] -
FMA: A Dataset For Music Analysis,
Michaël Defferrard, Kirell Benzi, Pierre Vandergheynst, Xavier Bresson, ISMIR, 2017.
[arXiv] [ISMIR] [EPFL] [reviews] [latex] [poster] [slides] [data] [code] -
Structured Sequence Modeling with Graph Convolutional Recurrent Networks,
Youngjoo Seo, Michaël Defferrard, Pierre Vandergheynst, Xavier Bresson, ICONIP, 2016.
[arXiv] [ICONIP] [EPFL] [OpenReview] [latex] [slides] [code] -
Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering,
Michaël Defferrard, Xavier Bresson, Pierre Vandergheynst, NIPS, 2016.
[arXiv] [NIPS] [EPFL] [reviews] [latex] [poster] [slides] [video] [code]
Software
GitHub-
PyGSP: Graph Signal Processing in Python.
[github] [documentation] [PyPI] [conda] [Arch Linux] [travis] [coverage] [zenodo] -
DeepSphere: learning on the sphere.
[github] -
PyUNLocBoX: a Python convex optimization toolbox using proximal splitting.
[github] [documentation] [PyPI] [conda] [travis] [coverage] [zenodo]
Talks
Zenodo YouTube ORCID-
Apprentissage et intelligence artificielle, séminaire de formation "Thèmes d'actualité pour l'auditeur", Lausanne, 2022-11-25.
[slides] -
Le métier de Scientifique (en Intelligence Artificielle), semaine technique au Gymnase du Bugnon, organisée par IngCH, Lausanne, 2022-11-18.
[slides] -
Chercheur en Intelligence Artificielle, Forum Perspectiva, University of Fribourg, 2022-01-25.
[slides] - Leveraging topology, geometry, and symmetries for efficient Machine Learning, PhD thesis defense, EPFL, 2021-12-15.
[slides] - Leveraging topology, geometry, and symmetries for efficient Machine Learning, job talk, Google, 2021-12-09.
[slides] - Leveraging topology, geometry, and symmetries for efficient Machine Learning, job talk, Qualcomm, 2021-11-22.
[slides] - Deep Learning on the sphere for weather/climate applications, talk, Copernicus' European Geosciences Union (EGU) General Assembly, 2021-04-20.
[poster] [abstract] -
Learning from graphs: a spectral perspective, invited talk, Amsterdam Machine Learning Lab Seminar, 2021-02-25.
[slides] -
Chercheur en Intelligence Artificielle, Job-Info, University of Fribourg, 2021-02-08.
[slides] [video] -
DeepSphere: a graph-based spherical CNN, ICLR spotlight talk, 2020-04-30.
[slides] [video] -
Keynote, GCNN | RL workshop, ENS Lyon, 2020-03-23. Canceled due to COVID-19.
- DeepSphere and the Earth, ML & Climate workgroup, EPFL, 2020-01-23. [slides]
-
Chercheur en Intelligence Artificielle, Job-Info, University of Fribourg, 2020-01-22.
[slides] -
DeepSphere: a graph-based spherical CNN, talk, EPFL visual computing seminar, 2019-11-15.
[slides] -
DeepSphere: Efficient spherical Convolutional Neural Network with HEALPix sampling for cosmological applications, talk, AI methods in Cosmology, 2019-06-11.
[slides] -
Deep Learning on graphs: a journey from continuous manifolds to discrete networks, invited talk, KCL/UCL Junior Geometry Seminar, 2019-06-06.
[slides] -
Chercheur en Intelligence Artificielle, Job-Info, University of Fribourg, 2019-02-11.
[slides] -
Learning on Graphs, talk, AMLD AI & Networks track, 2019-01-28.
[slides] [video] -
Learning on Graphs, lecture, NTDS course, 2018-12-11.
[slides] -
Deep Learning on Graphs, lecture, GraphSiP summer school, 2018-09-12.
[slides] -
Advances in Deep Learning on Graphs, talk, GSP workshop, 2018-06-11.
[slides] -
Deep Learning on Graphs, talk, Peyresq summer school, 2018-07-02.
[slides] -
Challenge Overview: Learning to Recognize Musical Genre from Audio, talk, The Web Conference challenge track, 2018-04-27.
[slides] -
Learning on Graphs, lecture, NTDS course, 2017-12-08.
[slides] -
FMA: A Dataset For Music Analysis, talk, Data Jam days, 2017-11-24.
[slides] -
Opening a Large Audio Dataset, talk, Open Science in Practice, 2017-09-29.
[slides] -
Deep Learning on Graphs, talk, Deep Learning on Irregular Domain (DLID) workshop at BMVC, 2017-09-07.
[slides] [slides (figshare)] -
Deep Learning on Graphs, talk, GSP workshop, 2017-06-02.
[slides] -
Deep Learning on Graphs, talk, LTS2 group meeting, 2017-02-24.
[slides] -
Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering, NetSci-X, 2017-01-18.
[slides] -
Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering, Geometric Deep Learning Seminar at Tel Aviv University, 2017-01-17.
[slides] -
Deep Learning on Graphs, lecture, NTDS course, 2016-12-21.
[slides] [slides (figshare)] -
Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering, Swiss Machine Learning Day (SMLD), 2016-11-22.
[slides] -
Deep Learning on Graphs for Advanced Big Data Analysis, candidacy exam at EPFL, 2016-08-30.
[slides]
Teaching
- Mentor, London Geometry and Machine Learning workshop, 2021.
-
Learning from Graphs: From Mathematical Principles to Practical Tools, invited hands-on tutorial at The Web Conference, 2021.
[github] -
A Network Tour of Data Science, master course at EPFL, 2019.
Topics: Network Science, Spectral Graph Theory, Graph Signal Processing, Graph Neural Networks, Data Science.
Roles: helped design curriculum, gave lectures, managed the teaching team of 12 and the class of 137 students.
[github] [moodle] -
Learning and Processing over Networks, workshop at the Applied Machine Learning Days, 2019.
Co-organized with Rodrigo Pena, for about 65 participants.
[github] [slides] -
A Network Tour of Data Science, master course at EPFL, 2018.
Topics: Data Science, Network Science, Spectral Graph Theory, Graph Signal Processing, Machine Learning.
Roles: helped design curriculum, gave tutorials, gave a lecture, managed the teaching team of 8 and the class of 180 students.
[github] [moodle] -
Graph Signal Processing with the PyGSP, tutorial at the GraphSIP summer school, 2018.
Co-organized with Nicolas Tremblay, for about 25 participants.
[github] -
A Network Tour of Data Science, master course at EPFL, 2017.
Topics: Data Science, Network Science, Spectral Graph Theory, Graph Signal Processing, Machine Learning.
Roles: helped design curriculum, gave tutorials, gave a lecture, managed the class of 110 students and their projects.
[github] [moodle] -
A Network Tour of Data Science, master course at EPFL, 2016.
Topics: Data Science, Neural Networks, Graph Theory.
Roles: helped design curriculum, taught computational tools, managed the class of 50 students and their projects.
[github] [moodle] -
Data Science, Certificate of Advanced Studies at EPFL, 2016.
Topics: Data Science, Neural Networks, Scientific Python.
Roles: taught computational tools and the Python scientific stack to about 25 participants.
[github] -
Signal and Systems, bachelor course at EPFL, 2016.
Topic: Signal Processing.
Roles: assisted the students during exercise sessions, designed exams.
Supervision
-
Towards data-driven probabilistic weather predictions.
Yann Yasser Haddad, semester project (EPFL), 2020.
Co-supervised with Gionata Ghiggi (PhD student at EPFL LTE).
[slides] [code] -
Spherical geometry in data-driven weather prediction.
Wentao Feng, semester project (EPFL), 2020.
Co-supervised with Gionata Ghiggi (PhD student at EPFL LTE).
[slides] [code] -
Anisotropic graph neural networks.
Hugo Aguettaz, MSc thesis (UvA AMLab), 2021.
Co-supervised with Erik Bekkers (assistant professor at UvA AMLab).
[code] -
Geometric deep learning for medium-range weather prediction.
Natalie Bolón Brun, research intern (EPFL LTS2), 2020.
Co-supervised with Gionata Ghiggi (PhD student at EPFL LTE).
[proposal] [code] -
Geometric deep learning for medium-range weather prediction.
Icíar Lloréns Jover, MSc thesis (EPFL LTS2), 2020.
Co-supervised with Gionata Ghiggi (PhD student at EPFL LTE).
[proposal] [report] [slides] [code] -
Transferability of Spectral Graph Neural Networks.
Axel Nilsson, MSc thesis (NTU), 2020.
Co-supervised with Xavier Bresson (associate professor at NTU).
[report] [slides] [code] -
Implement DeepSphere in pytorch as a python package.
Laure Vancauwenberghe and Michael Allemann, internships (Arcanite Solutions), 2019.
Co-supervised with Johan Paratte and Lionel Martin (co-founders of Arcanite Solutions).
[code] -
Learning on manifolds.
Guohao Dou, PhD semester project (EPFL LTS2), 2019.
Co-supervised with Martino Milani (intern at EPFL LTS2) and Nathanaël Perraudin (research scientist at SDSC).
[proposal] [code] -
Using the FEM Laplacian in DeepSphere.
Arnaud Gaudard, bachelor semester project (EPFL LTS2), 2019.
Co-supervised with Martino Milani (intern at EPFL LTS2).
[proposal] [slides] [code] -
Protein reconstruction from multiple images with Deep Learning.
Jelena Banjac, master semester project (EPFL BIG), 2019.
Co-supervised with Laurène Donati (PhD student at EPFL BIG).
[proposal] [report] [slides] [code] -
Bilateral Trade Modelling with Graph Neural Networks.
Kobby Panford-Quainoo, MSc thesis (African Master in Machine Intelligence), 2019.
[paper] [slides] [code] -
Geometric Deep Learning for Volumetric Computational Fluid Dynamics.
Luca Zampieri, MSc thesis (Neural Concept), 2019.
Co-supervised with Pierre Baqué (CEO of Neural Concept), François Fleuret (professor at IDIAP), and Pascal Fua (professor at EPFL).
[proposal] [report] [slides] [poster] -
Discrete Laplace-Beltrami for rotation equivariant NNs.
Martino Milani, MSc thesis & research intern (EPFL LTS2), 2019.
Co-supervised with Nathanaël Perraudin (research scientist at SDSC).
[proposal] [report] [slides] [poster] [code] -
Empirical evaluation of Spherical CNNs.
Frédérick Matthieu Gusset, MSc thesis & research intern (EPFL LTS2), 2019.
Co-supervised with Nathanaël Perraudin (research scientist at SDSC).
[proposal] [report] [slides] [code] -
Exploiting symmetries (in images) with graph neural networks.
Charles Gallay, master semester project (EPFL LTS2), 2019.
Co-supervised with Nathanaël Perraudin (research scientist at SDSC).
[proposal] [blog] [slides] [code] -
Learning Representations of Source Code from Structure and Context.
Dylan Bourgeois, MSc thesis (Stanford SNAP), 2018.
Co-supervised with Michele Catasta (postdoc at Stanford SNAP).
[report] [slides] [code] [data (raw)] [data (processed)] [code (data preprocessing)] -
Deep learning on graph for semantic segmentation of point clouds.
Alexandre Cherqui, MSc thesis (Picterra), 2018.
Co-supervised with Frank de Morsier (CTO of Picterra).
[proposal] [report] [slides] [poster] -
Music Information Retrieval on the Free Music Archive.
Chibueze Ukachi, semester project (EPFL LTS2), 2017.
[proposal] [code] -
Convolutional Neural Networks on Graphs and Stationarity.
Thomas Grivaz, semester project (EPFL LTS2), 2016.
Co-supervised with Nathanaël Perraudin (PhD student at EPFL LTS2).
[proposal] [report] -
Musical Score Generation with Recurrent Neural Networks.
Yoann Ponti, semester project (EPFL LTS2), 2016.
Co-supervised with Nathanaël Perraudin (PhD student at EPFL LTS2).
[proposal] [blog] [code]
Organization
-
Machine Learning for the Earth, Seminar, 2019–present.
Co-organized with Gionata Ghiggi, Eniko Székely, Jussi Leinonen, Raphaël de Fondeville.
[event] -
Musical Genre Recognition Challenge, The Web Conference (WWW), 2018.
Co-organized with Sharada P. Mohanty, Sean F. Carroll, Marcel Salathé.
[event] [conference] [code] [data] [paper] [slides] -
Open Science in Practice, Summer School, EPFL, 2017.
Co-organized with Marjan Biocanin, Luc Henry, Marc Lauffs, Dasaraden Mauree, Robin Scheibler, Valeria Superti.
[event] [announcement] -
Deep Learning on Irregular Domains, Workshop, BMVC, 2017.
Co-organized with Xianghua Xie, Michael Edwards, Pierre Vandergheynst.
[event] [conference] -
Graph Signal Processing Summer School, CIRM, 2016. Funding denied.
Co-organized with Nathanaël Perraudin, Yann Schoenenberger, Dorina Thanou, Pascal Frossard, Mauro Maggioni, Bruno Torrésani.
[organization] [proposal]
Projects
-
Structured Auto-Encoder with application to Music Genre Recognition, Michaël Defferrard, 2015.
MSc thesis. Advized by Xavier Bresson, Johan Paratte, Pierre Vandergheynst.
[report] [code] [slides] [results] [report src] -
GIIN: Graph-based Image Inpainting, Michaël Defferrard, 2014.
MSc semester project. Advized by Nathanaël Perraudin, Johan Paratte.
[report] [code] -
LiveMesh, a tool for real-time rendering of neuronal cells from morphologies, Michaël Defferrard, 2014.
MSc minor project. Advized by Jafet Villafranca Diaz, Stefan Eilemann, Felix Schürmann.
[report] [slides] -
Super resolution for mass spectrometry, Michaël Defferrard, 2014.
Internship. Advized by Nathanaël Perraudin, Yury Tsybin.
[code] [slides] -
Terrain: an OpenGL application that generates and renders a 3D scene,
Michaël Defferrard, Pierre Fechting, Vu Hiep Doan, 2014. MSc course project.
[code] [report 1] [report 2] [report 3] -
ATCsim: an Air Traffic Control simulation and a C++11 learning experience,
Michaël Defferrard, Enguerrand Granoux, 2013. MSc course project.
[code] -
Sensor module studies for LHC's ATLAS detector, Michaël Defferrard, 2012.
BASc thesis. Advized by Sergio Díez Cornell, Carl H. Haber.
[report] [slides] [poster] -
Eurobot: Absolute Local Positioning System, Michaël Defferrard, 2012.
BASc semester project. Advized by Wolfram Luithardt, Michael Ansorge.
[report] [slides]
Miscellaneous
Besides scientific research, I play brass band music at the Brass Band Fribourg and my town's band. Some recordings can be found on YouTube. I enjoy to read, some of which is on Goodreads, and to travel. I'm certainly too enthusiastic about computers.
I post some thoughts on Twitter, and some of my life on Facebook. I sometimes answer questions on Stack Overflow and Quora, and participate in Hacker News and reddit.