PeopleofSuperintelligence

Researcher

Mathieu Aubry

Mathieu Aubry is a computer vision researcher at École des Ponts ParisTech who builds neural networks that find patterns in image collections without prior annotation, with uses in historical data and Earth imagery. He earned his PhD at ENS in 2015, co-advised by Josef Sivic and Daniel Cremers, and spent a year as a postdoc with Alexei Efros at UC Berkeley. His papers include The Wave Kernel Signature in 2011, AtlasNet on 3D shape surfaces in 2018 and Deep Transformation-Invariant Clustering in 2020, and The Learnable Typewriter won an IAPR best paper award at ICDAR 2024. In 2022 he received an ERC Starting Grant for the DISCOVER project. He is a tenured researcher in the Imagine group at the LIGM lab and an ELLIS Scholar attached to the Paris Unit.

Timeline

  1. 2011

    The Wave Kernel Signature

    In 2011, he coauthors The Wave Kernel Signature, a quantum mechanical approach to shape analysis, presented at the ICCV workshop on Dynamic Shape Capture and Analysis.

  2. 2015

    A PhD and a Berkeley postdoc

    He earns his PhD at ENS in 2015, co-advised by Josef Sivic of INRIA and Daniel Cremers of TUM. In 2015, he also spends a year as a postdoc with Alexei Efros at UC Berkeley.

  3. 2018

    AtlasNet for 3D shape surfaces

    In 2018, he coauthors AtlasNet with Thibault Groueix, Matthew Fisher, Vladimir G. Kim and Bryan C. Russell, presented at CVPR 2018. The method represents a 3D shape as a collection of parametric surface elements to generate its surface.

  4. 2018–2026

    Computer vision for historians

    He has a leading role in the ANR EnHerit project on enhancing heritage image databases, which runs from 2018 to 2022. Computer vision tools for historians continue through the ANR VHS project on historical scientific illustrations from 2021 to 2025 and the ANR EIDA project on historical astronomical diagrams from 2022 to 2026.

  5. 2020

    Transformation-invariant image clustering

    In 2020, he coauthors Deep Transformation-Invariant Clustering with Tom Monnier and Thibault Groueix, presented at NeurIPS 2020. The method clusters images directly in image space with frameworks that jointly learn prototypes and transformations.

  6. 2022

    An ERC Starting Grant

    In 2022, he receives an ERC Starting Grant for the DISCOVER project, Discovering and Analyzing Visual Structures, which runs from 2023 to 2028. His team develops neural networks that extract information from images without experts annotating documents beforehand, with applications to historical data and Earth imagery.

  7. 2024

    A best paper award at ICDAR

    In 2024, his paper The Learnable Typewriter, a generative approach to text line analysis, receives the IAPR best paper award at ICDAR 2024.

  8. As of October 2026

    Researcher at École des Ponts

    He is a tenured computer vision researcher at École des Ponts ParisTech, working in the Imagine group of the A3SI team at the LIGM lab, and an ELLIS Scholar attached to the Paris Unit. In 2026, he serves as area chair for CVPR and ECCV.

Career

  1. NowÉcole des Ponts ParisTechTenured computer vision researcher
  2. NowEuropean Laboratory for Learning and Intelligent SystemsELLIS Scholar
  3. From 2015University of California, BerkeleyPostdoc

Education

  1. 2015École Normale SupérieurePhD

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Mathieu Aubry. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. Google has the highest combined h-index among AI organizationsIncludes European Laboratory for Learning and Intelligent Systems. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 44% of AI career moves go from one university to anotherIncludes European Laboratory for Learning and Intelligent Systems. We estimate how 26,439 AI career moves link universities, companies, labs, nonprofits and government research centers, with decade counts, net flows and the top paths.
  4. 5 of the 25 people with the most-cited recent papers also lead on career citationsIncludes European Laboratory for Learning and Intelligent Systems. We compare citations to 2023–2025 papers with career citations and h-index, after correcting reviewed publication-date errors: who leads, when they started and where they work.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

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

  1. Mathieu Aubry mathieuaubry.github.io. Checked 21 October 2026.
  2. Mathieu Aubry - Sony CSL - Paris csl.sony.fr. Checked 21 October 2026.
  3. Alexei A. Efros: Curriculum Vitae people.eecs.berkeley.edu. Checked 21 October 2026.
  4. [1802.05384] AtlasNet: A Papier-Mâché Approach to Learning 3D Surface Generation arxiv.org. Checked 21 October 2026.
  5. [2006.11132] Deep Transformation-Invariant Clustering arxiv.org. Checked 21 October 2026.
  6. Highlights – Fédération de Recherche d'Animation Bézout bezout.math.cnrs.fr. Checked 21 October 2026.
  7. When artificial intelligence helps tell the story of old manuscripts | Institut Polytechnique de Paris ip-paris.fr. Checked 21 October 2026.
  8. Mathieu Aubry | European Laboratory for Learning and Intelligent Systems ellis.eu. Checked 21 October 2026.
  9. Mathieu Aubry: “Computer Vision Applications to Historical Data” — TU Wien Informatics informatics.tuwien.ac.at. Checked 21 October 2026.