PeopleofSuperintelligence

Researcher

Stig Petersen

Stig Petersen co-authored DeepMind papers on deep reinforcement learning, on game-based research environments and on protein-structure prediction. He co-authored a 2015 Nature paper on a deep Q-network that learns from sensory inputs, and a 2015 preprint on a massively distributed architecture for deep reinforcement learning. In December 2016, he was one of four authors of the post that open-sourced DeepMind Lab, and in August 2017 he co-wrote the paper that introduced SC2LE. He also co-wrote Spectral Inference Networks, which ICLR presented as a 2019 poster. His most recent confirmed work was the 2021 Nature paper on AlphaFold.

Works
14
Citations
90,435
h-index
11

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2015

    The deep Q-network paper

    He co-authors the Nature paper on a deep Q-network in February 2015, an artificial agent that learns from high-dimensional sensory inputs.

  2. 2015

    Massively parallel methods for deep RL

    A preprint from July 2015 that he co-authors presents a massively distributed architecture for deep reinforcement learning.

  3. 2016

    Open-sourcing DeepMind Lab

    In December 2016, he is one of four authors of the DeepMind post that open-sources DeepMind Lab, a platform for agent-based AI research.

  4. 2017

    The StarCraft II learning environment

    A paper from August 2017 introduces SC2LE, a reinforcement learning environment built on the StarCraft II game, and includes him among its authors.

  5. 2018–2019

    Spectral Inference Networks

    In 2018, he co-writes Spectral Inference Networks, a framework for learning eigenfunctions of linear operators by stochastic optimization, and ICLR 2019 presents it as a poster.

  6. 2021

    AlphaFold in Nature

    The 2021 Nature paper on AlphaFold, a system for protein-structure prediction, includes him among its authors.

Sources and credits

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

  1. Human-level control through deep reinforcement learning nature.com. Checked 14 October 2026.
  2. Massively Parallel Methods for Deep Reinforcement Learning arxiv.org. Checked 14 October 2026.
  3. Open-sourcing DeepMind Lab deepmind.google. Checked 14 October 2026.
  4. DeepMind Lab arxiv.org. Checked 14 October 2026.
  5. StarCraft II: A New Challenge for Reinforcement Learning arxiv.org. Checked 14 October 2026.
  6. Spectral Inference Networks: Unifying Deep and Spectral Learning iclr.cc. Checked 14 October 2026.
  7. Spectral Inference Networks: Unifying Deep and Spectral Learning mlanthology.org. Checked 14 October 2026.
  8. Spectral Inference Networks: Unifying Deep and Spectral Learning arxiv.org. Checked 14 October 2026.
  9. Highly accurate protein structure prediction with AlphaFold nature.com. Checked 14 October 2026.