Researcher

Chris Watkins

Also known as Chris JCH Watkins, Christopher John Cornish Hellaby Watkins

Chris Watkins introduced Q-learning, a method for reinforcement learning that lets agents learn actions from experience. He developed it in his 1989 Cambridge PhD thesis and proved its convergence with Peter Dayan in 1992. In 2014, DeepMind applied Q-learning to Atari games; Watkins also coauthored a 2009 study of epidemic awareness. As of October 2026, he is a Professor in Computer Science at Royal Holloway and leads High Stakes AI research there.

Works
84
Citations
23,199
h-index
20
Badges

Publication numbers from OpenAlex, as of 19 October 2026.

Timeline

  1. 1989

    A thesis that introduces Q-learning

    At the University of Cambridge, he completes Learning from Delayed Rewards, a 1989 PhD thesis that introduces Q-learning as an incremental way to learn action values without a model of the environment.

  2. 1992

    Proving Q-learning converges

    In 1992, he and Peter Dayan publish the first proof that Q-learning converges.

  3. 2009

    Modeling epidemic awareness

    In 2009, he publishes a mathematical study with Sebastian Funk, Erez Gilad, and Vincent A. A. Jansen on awareness spreading alongside disease. Their model finds that locally spreading awareness can stop an outbreak when the infection rate stays below a threshold.

  4. 2014

    Bringing Q-learning to Atari

    In 2014, Google DeepMind presents a deep learning application of Q-learning that plays Atari 2600 games at expert human levels.

  5. As of October 2026

    Leading high-stakes AI research

    As of October 2026, he is a Professor in Computer Science at Royal Holloway and High Stakes AI Research lead at its Centre for AI and Skills, which unites the university's AI community. His research interests include reinforcement learning, evolutionary algorithms, kernel methods in machine learning, and LLMs. In January 2026, he coauthors an IEEE conference paper on convolutional neural networks for sound classification.

Career

  1. Since 1998Royal Holloway University of LondonProfessor in Computer Science
  2. 2018–2023CSIRO
  3. 2006–2026University of London

Education

  1. 1989University of CambridgePhD

Most-cited works

  1. Q-learning1992 · Machine Learning · 9,072 citationsWith Peter Dayan
  2. Learning from Delayed Rewards1989 · 5,627 citations
  3. Technical Note: Q-Learning1992 · Machine Learning · 3,943 citationsWith Peter Dayan
  4. The spread of awareness and its impact on epidemic outbreaks2009 · Proceedings of the National Academy of Sciences · 1,130 citations
  5. Support vector machines for multi-class pattern recognition1999 · 794 citationsWith Jason Weston

Sources and credits

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

  1. Q Learning cml.rhul.ac.uk. Checked 19 October 2026.
  2. The spread of awareness and its impact on epidemic outbreaks pmc.ncbi.nlm.nih.gov. Checked 19 October 2026.
  3. Chris Watkins - Royal Holloway Research Portal pure.royalholloway.ac.uk. Checked 19 October 2026.
  4. Chris Watkins cs.rhul.ac.uk. Checked 19 October 2026.
  5. Centre for AI and Skills royalholloway.ac.uk. Checked 19 October 2026.