PeopleofSuperintelligence

Researcher

Thomas Mensink

Thomas Mensink is a research scientist at Google DeepMind Amsterdam as of October 2026; his research spans computer vision and machine learning. He earned an MSc in artificial intelligence at the University of Amsterdam and a PhD in computer science at Grenoble, then worked as a postdoctoral researcher at UvA and an assistant professor in 3D machine learning. His Fisher-kernel paper received ECCV's Koenderink Award, and his work earned best-paper awards from ACM Multimedia and ICMR. He coauthored the paper introducing Encyclopedic VQA in 2023, a pixel-space diffusion paper in 2025 and a latent-representation paper in 2026.

Works
154
Citations
7,952
h-index
30
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2010

    Improving the Fisher kernel

    In 2010, he coauthors “Improving the Fisher Kernel for Large-Scale Image Classification” with Florent Perronnin and Jorge Sánchez at ECCV.

  2. 2012

    An AFRIF PhD Thesis Prize

    In 2012, his PhD thesis, “Learning Image Classification and Retrieval Models,” wins AFRIF's PhD Thesis Prize.

  3. By 2013

    AI degrees from Amsterdam and Grenoble

    At CCVW 2013, he holds an MSc in artificial intelligence from the University of Amsterdam and a PhD in computer science from the University of Grenoble.

  4. 2013

    Postdoctoral research in Amsterdam

    At CCVW 2013, he is a postdoctoral researcher at the University of Amsterdam, with research interests in machine learning and computer vision.

  5. 2014–2020

    Three computer-vision awards

    His work receives ACM Multimedia's Best Paper Award in 2014 and ACM ICMR's Best Paper Award in 2016. In 2020, he and coauthors Florent Perronnin and Jorge Sánchez receive ECCV's Koenderink Award for their Fisher Kernel paper.

  6. 2023

    Encyclopedic VQA at ICCV

    At ICCV 2023, he coauthors the paper introducing Encyclopedic VQA, a dataset of visual questions about detailed properties of fine-grained categories and instances. It contains 221,000 unique question-and-answer pairs, each matched with up to five images.

  7. 2025–2026

    Diffusion and latent models

    At CVPR 2025, he coauthors “Simpler Diffusion,” a pixel-space diffusion model that achieves 1.5 FID on ImageNet512. In February 2026, he coauthors “Unified Latents,” a framework for learning latent representations that uses a diffusion prior and a diffusion model for decoding.

  8. As of October 2026

    Research at Google DeepMind

    He is a research scientist at Google DeepMind Amsterdam after an assistant professorship in 3D machine learning at UvA.

Career

  1. Since 2025Google DeepMindResearch Scientist
  2. 2019–2026Google
  3. 2013–2021University of AmsterdamPostdoctoral researcher
  4. 2011–2013Xerox
  5. 2008–2013Inria

Education

  1. 2013University of AmsterdamMSc in artificial intelligence
  2. 2013University of GrenoblePhD in computer science

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Thomas Mensink. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  2. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Thomas Mensink. 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.
  3. 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.
  4. 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.

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 Thomas Mensink yet. Send us a free-licensed photo.

  1. Research mensink.nu. Checked 20 October 2026.
  2. Invited Speakers - Computer Vision Center fer.unizg.hr. Checked 20 October 2026.
  3. Thomas Mensink research.google. Checked 20 October 2026.
  4. Award Calls | ECCV 2020 eccv2020.eu. Checked 20 October 2026.
  5. Encyclopedic VQA: Visual questions about detailed properties of fine-grained categories research.google. Checked 20 October 2026.
  6. Simpler Diffusion: 1.5 FID on ImageNet512 with Pixel-space Diffusion openaccess.thecvf.com. Checked 20 October 2026.
  7. [2602.17270] Unified Latents (UL): How to train your latents arxiv.org. Checked 20 October 2026.
  8. Thomas Mensink mensink.nu. Checked 20 October 2026.
  9. CV Group @ University of Amsterdam ivi.fnwi.uva.nl. Checked 20 October 2026.