Sergey Levine

The engines agree that Sergey Levine is a UC Berkeley EECS professor working on reinforcement learning and robotics, with relevance to AI safety through safe RL, offline data and inverse reinforcement learning. Gemini adds that he co-founded Physical Intelligence. Copilot adds vision-language-action models such as RT-2 and OpenVLA. Perplexity adds his lab RAIL and a 2017 project with Anca Dragan on objective misspecification. The most cited pages are people.eecs.berkeley.edu, youtube.com and arxiv.org. All five engines name him.

• ChatGPTNot named

  ChatGPT says he is a UC Berkeley EECS professor researching algorithms for autonomous agents to learn complex behaviors through reinforcement learning, inverse reinforcement learning and robotic learning.

  Known for

  • Deep reinforcement learning
  • Inverse reinforcement learning
  • Robotic learning
• GeminiNot named

  Gemini describes a UC Berkeley professor and co-founder of Physical Intelligence whose work covers reinforcement learning, robotics, safe RL, offline data and Lyapunov density models.

  Known for

  • Safe reinforcement learning
  • Offline reinforcement learning
  • Lyapunov density models
• Google AI ModeNot named

  Google AI Mode describes a UC Berkeley EECS professor in machine learning, reinforcement learning and robotics, whose work addresses safe offline reinforcement learning and robustness under distribution shift.

  Known for

  • Offline reinforcement learning
  • Safe exploration for robots
  • Decision-making under uncertainty
• PerplexityNot named

  Perplexity says he is a UC Berkeley EECS associate professor leading the RAIL lab on robot learning and reinforcement learning, with an embodied view of alignment.

  Known for

  • Robot learning
  • Reinforcement learning
  • Empirical embodied alignment
• Bing AI ModeNot named

  No description yet.

• CopilotNot named

  Copilot says his work covers safe reinforcement learning with offline data, scalable inverse reinforcement learning and vision-language-action models for robotics such as RT-2 and OpenVLA.

  Known for

  • Safe reinforcement learning
  • Scalable inverse reinforcement learning
  • Vision-language-action models

• #7Roboticsof 38
  Named in 39 answers