Abstract

From autonomous robots to digital assistants, the future of AI is inherently multiagent—where systems must learn, adapt, and strategize in dynamic environments. How do robots collaborate and compete to outmaneuver opponents in soccer and push the limits in quadrotor racing? How can AI help millions of users find the best routes every day? This talk explores the role of reinforcement learning in such settings, where cooperation and competition shape intelligent behavior. We will dive into robotics applications like bipedal robot soccer and quadrotor racing, alongside large-scale digital systems such as Google Maps. By highlighting key challenges and breakthroughs, we will uncover how multiagent systems are shaping the next generation of autonomy.

Bio: Markus Wulfmeier is a researcher in machine learning and robotics at Google DeepMind with a focus on fundamental and applied research on reinforcement, imitation, and transfer learning. His work aims at efficiently scalable algorithms across a variety of real-world applications including robotics, navigation, and language modelling. Markus was a postdoctoral research scientist at the Oxford Robotics Institute and a member of Oxford University’s New College where he completed his PhD. Over the years, he has held visiting scholar positions with UC Berkeley, MIT , and ETH . His work received best paper awards including IROS and GVSETS and was covered in the press including 60 Minutes, The Verge, MIT News, Wired, BBC News, New Scientist, and Popular Science.