This summer school offers tutorials and lectures on state-of-the-art machine learning methods for training the next generation of learning robots. This summer school is an extension supported by the many robotics groups around Montreal.
The tutorials and interaction with real robot hardware will take place at Mila, the Quebec AI Institute, the largest academic organization for deep learning researchers worldwide. Attendees will have the opportunity to learn about current deep reinforcement learning methods, how to apply them to real hardware, sim2real transfer techniques, and experience with legged robotics.
Challenge: On the last day of the summer school, teams will use their new skills to compete in designing the best vision-based navigation controller on a quadruped robot to navigate an obstacle course.
What material should you bring?
Bring your brains. We will provide robots and computers to control the hardware. If you have a laptop with a GPU, you will be able to run the code for the summer school yourself. However, we will provide access to computers that you can use for training. The days will consist of talks in the mornings to introduce the material, and the afternoons are hands-on, where you will apply those concepts to the robot hardware.
For the summer school, please bring a PC laptop (Mac won’t be working for this challenge)
During the summer school, you will attend various tutorials and have access to a real quadruped robot. In order to prepare for the summer school, you should install the following software:
