PeopleofSuperintelligence

Researcher

Roger Grosse

Roger Grosse is a machine-learning researcher who studies how neural networks train and now applies that understanding to AI alignment. After a computer science PhD at MIT in 2014 under Bill Freeman and Josh Tenenbaum, he worked as a postdoctoral researcher at the University of Toronto from 2014 to 2016. In 2015, he coauthored K-FAC with James Martens and the importance weighted autoencoder with Yuri Burda and Ruslan Salakhutdinov, and in 2017 he became a Canada CIFAR AI Chair at the Vector Institute. He received a Sloan Research Fellowship in 2021, scaled influence functions to language models with up to 52 billion parameters in 2023, and won an ICML Best Paper award in 2024 with three coauthors for twisted sequential Monte Carlo inference in language models. Today, he is an associate professor of computer science at the University of Toronto and a Member of Technical Staff on Anthropic's Alignment Science Team, where his work focuses on training data attribution.

Timeline

  1. 2008–2014

    Stanford degrees and an MIT PhD

    He earns a BS in symbolic systems from Stanford in 2008, an MS in computer science there in 2009, and a PhD in computer science from MIT in 2014, studying under Bill Freeman and Josh Tenenbaum. He wins Best Student Paper at the Conference on Uncertainty in AI in 2012.

  2. 2014–2016

    Postdoc with Salakhutdinov

    He is a postdoctoral researcher at the University of Toronto, working with Ruslan Salakhutdinov, and receives an NSERC Banting Postdoctoral Fellowship in 2015. With Colorado Reed, he creates Metacademy, a website that uses a dependency graph of concepts to build personalized learning plans for machine learning.

  3. 2015

    K-FAC and importance weighted autoencoders

    In March 2015, he and James Martens post Kronecker-Factored Approximate Curvature (K-FAC), a method for approximating natural gradient descent in neural networks. In September 2015, he, Yuri Burda and Ruslan Salakhutdinov post the importance weighted autoencoder, a generative model that uses a tighter log-likelihood bound derived from importance weighting.

  4. Since 2017

    Canada CIFAR AI Chair

    In 2017, he is appointed a Canada CIFAR AI Chair at the Vector Institute, and the chair is renewed in 2023. He is a founding member of the Vector Institute and a Canada Research Chair in Probabilistic Inference and Deep Learning. He also receives the Connaught New Researcher Award in 2017 and the Ontario Early Researcher Award in 2018.

  5. 2021

    A Sloan Research Fellowship

    In 2021, he receives a Sloan Research Fellowship as an assistant professor in the Department of Computer Science at the University of Toronto, one of five Toronto researchers awarded the fellowship for 2021. His research focuses on neural network training dynamics, with the aim of improving training speed, generalization, uncertainty estimation and automatic hyperparameter tuning.

  6. 2023

    Influence functions for language models

    In August 2023, he and coauthors including Juhan Bae and Cem Anil use the EK-FAC approximation to scale influence functions, which trace a model's behavior back to the training examples that contribute most, to language models with up to 52 billion parameters.

  7. 2024

    ICML Best Paper and AI2050

    With Stephen Zhao, Rob Brekelmans and Alireza Makhzani, he receives a Best Paper award at ICML 2024 for Probabilistic Inference in Language Models via Twisted Sequential Monte Carlo, which uses learned twist functions to focus inference-time computation on promising partial sequences. Schmidt Sciences also names him a Senior AI2050 Fellow in 2024.

  8. As of October 2026

    Toronto professor, Anthropic researcher

    He is an Associate Professor of Computer Science at the University of Toronto, Schwartz Reisman Chair in Technology and Society, and a Member of Technical Staff on the Alignment Science Team at Anthropic, where his work focuses on training data attribution. His research now applies an understanding of deep learning to AI alignment.

Career

  1. Since 2017Vector InstituteCanada CIFAR AI Chair
  2. Since 2014University of TorontoPostdoctoral Researcher
  3. NowAnthropicMember of Technical Staff

Education

  1. 2014Massachusetts Institute of TechnologyPhD
  2. 2009Stanford UniversityMS
  3. 2008Stanford UniversityBS

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 Roger Grosse. 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. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes Anthropic. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.
  3. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Anthropic. 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.
  4. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Anthropic. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.

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 Roger Grosse yet. Send us a free-licensed photo.

  1. SRI Seminar Series: Roger Grosse, “On the origin of rogue AI” — Schwartz Reisman Institute srinstitute.utoronto.ca. Checked 22 October 2026.
  2. Roger Grosse – CIFAR cifar.ca. Checked 22 October 2026.
  3. Optimizing Neural Networks with Kronecker-factored Approximate Curvature arxiv.org. Checked 22 October 2026.
  4. Importance Weighted Autoencoders arxiv.org. Checked 22 October 2026.
  5. Roger Grosse - Vector Institute for Artificial Intelligence vectorinstitute.ai. Checked 22 October 2026.
  6. Two Department of Computer Science researchers receive 2021 Sloan Research Fellowships web.cs.toronto.edu. Checked 22 October 2026.
  7. Roger Grosse cs.toronto.edu. Checked 22 October 2026.
  8. Studying Large Language Model Generalization with Influence Functions arxiv.org. Checked 22 October 2026.
  9. ICML 2024 Awards icml.cc. Checked 22 October 2026.
  10. Roger Grosse - AI2050 ai2050.schmidtsciences.org. Checked 22 October 2026.