Join us and learn  about our newest project which introduces state-of-the-art methods in data-enabled control, from online adaptation to offline reinforcement learning.

Safe Learning in Robotics

Typically, the decision-making rules in automated systems, commonly known as controllers, are designed prior to operation and not adapted during operation. Could these hand-designed rules be improved by leveraging the history of the data your system has access to? In response to this problem, newer techniques such as Reinforcement learning (RL) and learning-based control theory have made significant progress in the past 5 years – in particular, when it comes to algorithms that work on real hardware, where data efficiency and safety is key.

By signing up to this project, participants will learn how to:

• Leverage historical data to improve automated processes involving a sensor and an actuator
• Incorporate reinforcement learning methodologies into your use case
• Train intelligent agents using a simulation environment 

For more details on the project’s structure and timeline, please refer to the Safe Learning in Robotics – Brochure

Facilitated by Angela Schoellig

Angela Schoellig is an Associate Professor at the University of Toronto Institute for Aerospace Studies and a Faculty Member of the Vector Institute. She’s known for leading the team that won the North-American AutoDrive Challenge for four consecutive years, and for speaking at the 2020 and 2021 NeurIPS workshops. Angela has been recognized with numerous awards, a few of which include the MIT Technology Review’s Innovators Under 35, the Robohub’s 25 Women in Robotics you Need to Know About, and the Alexander von Humboldt Professorship in 2021.

This info session is open to Vector Sponsors, Vector Researchers, and invited health partners only. Any registration that is found not to be a Vector Sponsor, Vector Researcher or invited health partner will be asked to provide verification and, if unable to do so, will not be able to attend the event. Please contact events@vectorinstitute.ai with any questions.