Tuomas Haarnoja
First author of Soft Actor-Critic and a bipedal robot soccer paper; joined DeepMind after a Berkeley PhD
Tuomas Haarnoja completed a PhD thesis, "Acquiring Diverse Robot Skills via Maximum Entropy Deep Reinforcement Learning", in Pieter Abbeel's group at UC Berkeley in December 2018, and the group's alumni list records Haarnoja's next role as Research Scientist at DeepMind. Haarnoja is first author of the Soft Actor-Critic paper (2018), an off-policy maximum-entropy deep reinforcement learning algorithm with a stochastic actor, written with Aurick Zhou, Pieter Abbeel and Sergey Levine. Haarnoja is also first author of "Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning" (2023), which lists Haarnoja's affiliation as Google DeepMind and trained a miniature humanoid robot to play one-versus-one soccer.123
Work
Soft Actor-Critic (SAC) (external site: arxiv.org)Published research
First author of "Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor" (2018).2
Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning (external site: arxiv.org)Published research
First author of the 2023 paper, with Google DeepMind as the listed affiliation, using deep reinforcement learning to train a low-cost miniature humanoid robot to play a simplified one-versus-one soccer game.3
Other documented roles
Roles the sources document but that are not confirmed as current.
- Research Scientist, DeepMind (Google DeepMind)13
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.