Sotabase

Career

· Director, AI Safety and Alignment, Google DeepMind2023–
· Director of AI Safety and Alignment, Google DeepMind2023–
· Consultant, Waymo2017–
· Senior Research Scientist (Part-Time), Waymo2017–2023
· Associate Professor, EECS Department, University of California, Berkeley2015–
· PhD, Robotics, Carnegie Mellon University2009–2015
· B.S., Computer Science, Jacobs University Bremen2006–2009

Publications (221)

Gemma 2: Improving Open Language Models at a Practical Size
arXiv.org · 2024
1,634
cited
Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback
Trans. Mach. Learn. Res. · 2023
738
cited
Legibility and predictability of robot motion
IEEE/ACM International Conference on Human-Robot Interaction · 2013
735
cited
Cooperative Inverse Reinforcement Learning
Neural Information Processing Systems · 2016
704
cited
Planning for Autonomous Cars that Leverage Effects on Human Actions
Robotics: Science and Systems · 2016
533
cited
On the Utility of Learning about Humans for Human-AI Coordination
Neural Information Processing Systems · 2019
488
cited
Inverse Reward Design
Neural Information Processing Systems · 2017
456
cited
Active Preference-Based Learning of Reward Functions
Robotics: Science and Systems · 2017
435
cited
A policy-blending formalism for shared control
Int. J. Robotics Res. · 2013
393
cited
Managing extreme AI risks amid rapid progress
Science · 2023
339
cited
Effects of Robot Motion on Human-Robot Collaboration
IEEE/ACM International Conference on Human-Robot Interaction · 2015
333
cited
SQIL: Imitation Learning via Reinforcement Learning with Sparse Rewards
International Conference on Learning Representations · 2019
294
cited
Toward seamless human-robot handovers
J. Hum. Robot Interact. · 2013
293
cited
DART: Noise Injection for Robust Imitation Learning
Conference on Robot Learning · 2017
289
cited
Hierarchical Game-Theoretic Planning for Autonomous Vehicles
IEEE International Conference on Robotics and Automation · 2018
254
cited
Automatically Auditing Large Language Models via Discrete Optimization
International Conference on Machine Learning · 2023
215
cited
Shared Autonomy via Deep Reinforcement Learning
Robotics: Science and Systems · 2018
201
cited
Information gathering actions over human internal state
IEEE/RJS International Conference on Intelligent RObots and Systems · 2016
196
cited
The Social Cost of Strategic Classification
FAT · 2018
195
cited
Reward-rational (implicit) choice: A unifying formalism for reward learning
Neural Information Processing Systems · 2020
194
cited
Sotabase