PeopleofSuperintelligence

Researcher

Michael Hutchinson

Also known as Michael John Hutchinson

Michael Hutchinson is a machine learning researcher who builds diffusion models and Gaussian processes that respect the geometry of their data. After studying engineering at Cambridge, he began a DPhil in statistical machine learning at Oxford in 2019 under Yee Whye Teh and Max Welling. His 2022 paper Riemannian Score-Based Generative Modelling, written with Valentin De Bortoli, Emile Mathieu, James Thornton, Teh and Arnaud Doucet, extended score-based generative models to Riemannian manifolds and appeared among the NeurIPS 2022 Outstanding Papers. His 2024 Oxford thesis took diffusion models to manifolds with boundaries and to spaces of tensor fields and paths. He works at Isomorphic Labs on geometric machine learning.

Badges

Timeline

  1. 2015–2019

    Engineering at Cambridge

    He studies at Christ's College, University of Cambridge, for a BA in Engineering from 2015 to 2018 and an MEng in Information and Computer Engineering from 2018 to 2019. Dr Rich E. Turner supervises his master's work.

  2. 2019

    A DPhil at Oxford

    He begins a DPhil in Statistical Machine Learning at the University of Oxford in 2019, through the StatML course, supervised by Yee Whye Teh and Max Welling. He works in the OxCSML research group of the Department of Statistics and belongs to University College.

  3. 2020–2021

    Equivariant models and Gaussian processes

    In December 2020, he posts LieTransformer on arXiv with Charline Le Lan, Sheheryar Zaidi, Emilien Dupont, Yee Whye Teh and Hyunjik Kim; its LieSelfAttention layers are equivariant to arbitrary Lie groups. In 2021, he is first-listed author, with Alexander Terenin, Viacheslav Borovitskiy, So Takao, Teh and Marc Peter Deisenroth, of a NeurIPS paper on vector-valued Gaussian processes on Riemannian manifolds.

  4. 2022

    A NeurIPS Outstanding Paper

    In 2022, he co-writes Riemannian Score-Based Generative Modelling with Valentin De Bortoli, Emile Mathieu, James Thornton, Yee Whye Teh and Arnaud Doucet. The paper extends score-based generative models to Riemannian manifolds and demonstrates them on earth and climate science spherical data. NeurIPS lists it among its Outstanding Papers in November 2022.

  5. 2023

    Diffusion on constrained domains

    In 2023, he coauthors Diffusion Models for Constrained Domains with Nic Fishman, Leo Klarner, Valentin De Bortoli and Emile Mathieu, published in Transactions on Machine Learning Research in July 2023. He also coauthors Geometric Neural Diffusion Processes for NeurIPS 2023, which builds geometric priors into diffusion models of stochastic processes.

  6. 2024

    Thesis on Riemannian diffusion models

    His Oxford DPhil thesis, Generative modelling with diffusion models on Riemannian manifolds, is dated 2024, with Yee Whye Teh as supervisor. It extends diffusion models to Riemannian manifolds, to manifolds with boundaries, and to spaces of tensor fields and paths. For ICML 2024, he also coauthors Particle Denoising Diffusion Sampler with Angus Phillips, Hai-Dang Dau, Valentin De Bortoli, George Deligiannidis and Arnaud Doucet.

  7. As of October 2026

    Geometric ML at Isomorphic Labs

    He works at Isomorphic Labs, with geometric machine learning as his research area.

Career

  1. NowIsomorphic Labs

Education

  1. 2019University of CambridgeMEng in Information and Computer Engineering
  2. From 2019University of OxfordDPhil in Statistical Machine Learning
  3. 2018University of CambridgeBA in Engineering

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Michael Hutchinson. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  3. Google has the highest combined h-index among AI organizationsIncludes Google. 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.
  4. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

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 Michael Hutchinson yet. Send us a free-licensed photo.

  1. Michael Hutchinson mjhutchinson.info. Checked 21 October 2026.
  2. Generative modelling with diffusion models on Riemannian manifolds - ORA - Oxford University Research Archive ora.ox.ac.uk. Checked 21 October 2026.
  3. [2012.10885] LieTransformer: Equivariant self-attention for Lie Groups arxiv.org. Checked 21 October 2026.
  4. [2110.14423] Vector-valued Gaussian Processes on Riemannian Manifolds via Gauge Independent Projected Kernels arxiv.org. Checked 21 October 2026.
  5. [2202.02763] Riemannian Score-Based Generative Modelling arxiv.org. Checked 21 October 2026.
  6. Announcing the NeurIPS 2022 Awards – NeurIPS Blog blog.neurips.cc. Checked 21 October 2026.
  7. [2304.05364] Diffusion Models for Constrained Domains arxiv.org. Checked 21 October 2026.
  8. Geometric Neural Diffusion Processes proceedings.neurips.cc. Checked 21 October 2026.
  9. [2402.06320] Particle Denoising Diffusion Sampler arxiv.org. Checked 21 October 2026.
  10. Yee Whye Teh stats.ox.ac.uk. Checked 21 October 2026.
  11. Michael Hutchinson - Google Scholar scholar.google.com. Checked 21 October 2026.