Misha Denil
Former Google DeepMind research scientist whose work includes data-driven robotics and learned optimizers
Misha Denil spent 10 years as a research scientist at Google DeepMind and now describes doing their own thing. From 2019 to 2024, Denil's work focused on data-driven robotics, covering learning from demonstrations, reward learning, real-world evaluation, and complete agent systems, and Denil is a co-author of RoboCat, a self-improving generalist agent for robotic manipulation. Denil is also a co-author, with Google DeepMind colleagues, of the 2016 paper Learning to learn by gradient descent by gradient descent, which casts the design of optimization algorithms as a learning problem.12345
Work
RoboCat: A Self-Improving Generalist Agent for Robotic Manipulation (external site: arxiv.org)Published research
Co-author of the 2023 RoboCat paper (Transactions on Machine Learning Research), a visual goal-conditioned decision transformer trained on experience from simulated and real robotic arms.43
Learning to learn by gradient descent by gradient descent (external site: arxiv.org)Published research
Co-author, listed with Google DeepMind, of the 2016 paper showing optimization algorithms implemented as LSTMs that learn to exploit problem structure.53
Other documented roles
Roles the sources document but that are not confirmed as current.
- Research Scientist (10 years), Google DeepMind15
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.