The AWS DeepRacer League gives you the opportunity to race for prizes and glory, while also having fun & learning about reinforcement learning. You can use the AWS DeepRacer 3D racing simulator to build, train, and evaluate your models. You can review the results and improve your models in order to ensure that they are race-ready.
Winning a FORMULA 1 (F1) race requires a technologically sophisticated car, a top-notch driver, an outstanding crew, and (believe it or not) a healthy dose of machine learning. For the past couple of seasons AWS has been working with the Formula 1 team to find ways to use machine learning to make cars that are faster and more fuel-efficient than ever before (read The Fastest Cars Deserve the Fastest Cloud and Formula 1 Works with AWS to Develop Next Generation Race Car to learn more).
Each month the AWS DeepRacer League runs a new Virtual Race in the AWS DeepRacer console and this month is a special one: the Formula 1 DeepRacer ProAm Special Event. During the month of May you can compete for the opportunity to race against models built and tuned by Formula drivers and their crews. Here’s the lineup:
Rob Smedley – Director of Data Systems for F1 and AWS Technical Ambassador.
Daniel Ricciardo – F1 driver for Renault, with 7 Grand Prix wins and 29 podium appearances.
Tatiana Calderon – Test driver for the Alfa Romeo F1 team and 2019 F2 driver.
Each pro will be partnered with a member of the AWS Pit Crew tasked with teaching them new skills and taking them on a learning journey. Here’s the week-by-week plan for the pros:
Week 1 – Learn the basics of reinforcement learning and submit models using a standard, single-camera vehicle configuration.
Week 2 – Add stereo cameras to vehicles and learn how to configure reward functions to dodge debris on the track.
Week 3 – Add LIDAR to vehicles and use the rest of the month to prepare for the head-to-head qualifier.
At the end of the month the top AWS DeepRacer amateurs will face off against the professionals, in an exciting head to head elimination race, scheduled for the week of June 1.
The teams will be documenting their learning journey and you’ll be able to follow along as they apply real-life racing strategies and data science to the world of autonomous racing.
Bottom line: You have the opportunity to build & train a model, and then race it against one from Rob, Daniel, or Tatiana. How cool is that?
Start Your Engines
And now it is your turn. Read Get Started with AWS DeepRacer, build your model, join the Formula 1 DeepRacer ProAm Special Event, train it on the Circuit de Barcelona-Catalunya track, and don’t give up until you are at the top of the chart.
Training and evaluation using the DeepRacer Console are available at no charge for the duration of the event (Terms and Conditions apply), making this a great opportunity for you to have fun while learning a useful new skill.
Good luck, and see you at the finish line!
Originally posted on AWS News Blog
Author: Jeff Barr