Timothy Lillicrap
Also known as Timothy P. Lillicrap
Timothy Lillicrap is a machine learning researcher whose work spans neural-network learning, continuous control and self-learning game systems. He studies cognitive science and artificial intelligence in Toronto, earns a systems neuroscience PhD at Queen’s, then works at Oxford and Google DeepMind. His papers explore random feedback in deep networks and continuous-action reinforcement learning; his later work includes AlphaGo Zero and MuZero. The 2026 ICLR Test of Time Award recognizes his continuous-control paper, and he is an ELLIS Fellow affiliated with DeepMind.
- Now at
- Google DeepMind
- Works
- 147
- Citations
- 62,876
- h-index
- 57
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
2005–2012
An education across disciplines
He earns an honours B.Sc. in cognitive science and artificial intelligence at the University of Toronto in 2005 and completes a systems neuroscience PhD at Queen’s University in 2012.
-
2012–2016
From Oxford to DeepMind
After a postdoctoral research fellowship at Oxford from 2012 to 2014, he joins Google DeepMind as a research scientist in 2014 and becomes a staff research scientist in 2016.
-
2014
Learning with random feedback
He and Daniel Cownden, Douglas Tweed and Colin Akerman present Random feedback weights support learning in deep neural networks, a method that uses random synaptic weights to assign error signals.
-
2015
Continuous actions for control
With his coauthors, he presents Continuous control with deep reinforcement learning, an actor-critic method that operates over continuous action spaces.
-
2017
Go learned from itself
He contributes to AlphaGo Zero, DeepMind’s Go-playing system.
-
2020
Planning across many games
He coauthors MuZero, which masters Go, chess, shogi and Atari without game rules.
-
2026
A Test of Time award
ICLR’s 2026 Test of Time Award recognizes Continuous control with deep reinforcement learning, coauthored by Lillicrap and seven colleagues.
-
As of October 2026
An ELLIS fellowship
He is an ELLIS Fellow at DeepMind, associated with Interactive Learning and Interventional Representations.
Sources and credits
Photo: We have no published photo of Timothy Lillicrap yet. Send us a free-licensed photo.
- Homepage of Timothy Lillicrap
- Random feedback weights support learning in deep neural networks
- Continuous control with deep reinforcement learning
- AlphaGo Zero: Starting from scratch
- MuZero: Mastering Go, chess, shogi and Atari without rules
- Announcing the Test of Time Awards from ICLR 2016
- Timothy P. Lillicrap | European Laboratory for Learning and Intelligent Systems