PeopleofSuperintelligence

Researcher

Miika Aittala

Miika Aittala is a senior research scientist at NVIDIA who works on neural generative modeling, from GANs to diffusion models. He earned his M.Sc. and Ph.D. at Aalto University under Jaakko Lehtinen, and his thesis on photorealistic surface appearance won the Finnish Society for Computer Science dissertation award in 2017. After a postdoc at MIT CSAIL with Frédo Durand, he joined NVIDIA Research at Helsinki in November 2019 and coauthored GAN papers on StyleGAN image quality, limited-data training and alias-free synthesis. His paper Elucidating the Design Space of Diffusion-Based Generative Models won a NeurIPS Outstanding Paper award in 2022, and Guiding a Diffusion Model with a Bad Version of Itself was a NeurIPS 2024 best paper runner-up.

Now at
NVIDIA
Works
52
Citations
13,413
h-index
27
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2016–2017

    An award-winning Aalto dissertation

    He earns his M.Sc. and Ph.D. at Aalto University, advised by Jaakko Lehtinen. His dissertation, Computational methods for capture and reproduction of photorealistic surface appearance, is among the three best of 2016 in Aalto's School of Science. The Finnish Society for Computer Science presents him its dissertation award in 2017, and NVIDIA licenses two of the thesis algorithms.

  2. 2017–2019

    Postdoc at MIT CSAIL

    After graduating, he joins MIT CSAIL as a postdoctoral researcher and works with Frédo Durand on computer graphics, computer vision and machine learning. In 2018, he coauthors Noise2Noise at ICML, which shows that models can learn to restore images by looking only at corrupted examples.

  3. November 2019

    Joining NVIDIA Research Helsinki

    In November 2019, he leaves MIT and joins NVIDIA Research at Helsinki. His previous stops include an internship at NVIDIA Research Helsinki and work at or visits to Inria GraphDeco and VTT Technical Research Centre of Finland.

  4. 2020–2021

    Improved StyleGAN and alias-free GANs

    In 2020, he coauthors Analyzing and Improving the Image Quality of StyleGAN at CVPR, which redesigns generator normalization and adds a path length regularizer. He also coauthors Training Generative Adversarial Networks with Limited Data at NeurIPS 2020, whose adaptive discriminator augmentation often matches StyleGAN2 results with an order of magnitude fewer images. In 2021, he coauthors Alias-Free Generative Adversarial Networks at NeurIPS, whose networks are fully equivariant to translation and rotation even at subpixel scales.

  5. 2022

    A NeurIPS Outstanding Paper

    In 2022, he coauthors Elucidating the Design Space of Diffusion-Based Generative Models with Tero Karras, Timo Aila and Samuli Laine, which receives a NeurIPS Outstanding Paper award. The paper separates the design choices of diffusion models and reaches a class-conditional CIFAR-10 FID of 1.79 with 35 network evaluations per image.

  6. 2023–2024

    Rethinking diffusion model training

    In December 2023, he coauthors Analyzing and Improving the Training Dynamics of Diffusion Models, which improves the record ImageNet-512 FID from 2.41 to 1.81. His NVIDIA Technical Blog posts from December 2023 and March 2024 explain diffusion model sampling, parameterization and training.

  7. 2024

    NeurIPS 2024 best paper runner-up

    Guiding a Diffusion Model with a Bad Version of Itself, coauthored with Tero Karras, Tuomas Kynkäänniemi, Jaakko Lehtinen, Timo Aila and Samuli Laine, is a NeurIPS 2024 best paper runner-up. It guides generation with a smaller, less-trained version of the model itself and sets record ImageNet FIDs of 1.01 at 64x64 and 1.25 at 512x512.

  8. As of October 2026

    Senior research scientist at NVIDIA

    He is a senior research scientist at NVIDIA, where he works on neural generative modeling and computer graphics, with recent research on the fundamentals of diffusion models and GANs and their applications to imaging. In 2025, he coauthors Elucidated Rolling Diffusion Models for Probabilistic Weather Forecasting, published at NeurIPS 2025.

Career

  1. Since 2019NVIDIAIntern
  2. 2018–2019Massachusetts Institute of Technology
  3. 2017–2019MIT CSAILPostdoctoral researcher
  4. 2009–2010Ministry of Economic Affairs and Employment

Education

  1. 2016Aalto UniversityM.Sc.

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

Most-cited works

  1. Analyzing and Improving the Image Quality of StyleGAN2020 · IEEE Computer Society Conference on Computer Vision and Pattern Recognition · 6,312 citationsWith Samuli Laine and Tero Karras
  2. Alias-Free Generative Adversarial Networks2021 · 1,169 citationsWith Samuli Laine and Tero Karras
  3. Training Generative Adversarial Networks with Limited Data2020 · 1,159 citationsWith Samuli Laine and Tero Karras
  4. Noise2Noise: Learning Image Restoration without Clean Data2018 · 944 citationsWith Samuli Laine and Tero Karras
  5. Elucidating the Design Space of Diffusion-Based Generative Models2022 · neural information processing systems · 684 citationsWith Samuli Laine and Tero Karras

From OpenAlex, as of 21 October 2026.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Miika Aittala. 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 NVIDIA. 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. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. 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.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

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 Miika Aittala yet. Send us a free-licensed photo.

  1. MIT postdoc Miika Aittala receives dissertation award | Aalto University aalto.fi. Checked 20 October 2026.
  2. Miika Aittala people.csail.mit.edu. Checked 20 October 2026.
  3. Noise2Noise: Learning Image Restoration without Clean Data arxiv.org. Checked 20 October 2026.
  4. Miika Aittala | Research research.nvidia.com. Checked 20 October 2026.
  5. Analyzing and Improving the Image Quality of StyleGAN arxiv.org. Checked 20 October 2026.
  6. Training Generative Adversarial Networks with Limited Data arxiv.org. Checked 20 October 2026.
  7. Alias-Free Generative Adversarial Networks arxiv.org. Checked 20 October 2026.
  8. NeurIPS 2022 Awards neurips.cc. Checked 20 October 2026.
  9. Elucidating the Design Space of Diffusion-Based Generative Models arxiv.org. Checked 20 October 2026.
  10. Analyzing and Improving the Training Dynamics of Diffusion Models arxiv.org. Checked 20 October 2026.
  11. Author: Miika Aittala | NVIDIA Technical Blog developer.nvidia.com. Checked 20 October 2026.
  12. NeurIPS 2024 Awards neurips.cc. Checked 20 October 2026.
  13. Guiding a Diffusion Model with a Bad Version of Itself arxiv.org. Checked 20 October 2026.