Darshil Shah took a lot away from his first time participating in Humber Polytechnic's annual AWS Deepracer competition.
Shah – Artificial Intelligence, a graduate of cloud computing and recently with machine learning programs, felt that signing up for AWS Deepracer was a great learning opportunity for him, taking into account his field of research. Shah said much of what he learned in the class was used to help his team excel in the competition.
AWS Deepracer is a completely autonomous racing league with 1/18 scale race cars driven by machine learning (ML) algorithms. Students are tasked with programming the vehicle how to drive on a cloud-based 3D racing simulator. You can test the effectiveness of your programming, adjust it as needed, and then see what its impact on knee time.
If the competition is ready, the model car will be brought from a virtual truck to a real truck and compete to determine the best time to declare the winner.
Other skills were utilized in addition to the programming skills required to navigate the car through the truck. Initially, some models were off the truck. Therefore, Shah and his team have used their collective leadership skills to ensure that everyone remains motivated and tasked, leveraging the problem-solving skills needed to solve problems.
“I'm happy you're joining AWS Deepracer,” Shah said. “It was a great experience and a valuable experience for students taking artificial intelligence with machine learning. We really enjoyed it and were late at work, so we had to tell them to go home one night.”
Mihai Albu is a professor in the Faculty of Applied Sciences & Technology and a researcher at Humber. He oversees the AWS Deep Lacer competition for Humber, which was expanded this year to include teams from Cumbria, Conestoga, Seneca and George Brown College for the first time.
In total, over 65 students from 15 teams were on the North Campus, where they were there for the competition.
Albu said the competition is valuable to learners as they develop programming skills while gaining practical experience.
It also teaches students about the concept of reinforcement learning. This is an ML technique to train your software to make decisions to achieve the most optimal results. Students can learn when experimenting with different parameters in programming and see the outcome of decisions in real time based on how the model does it in lap.
Albu said the competition is a good way for students to experiment and learn AI and ML, indicating the practical application of technology.
“What I find great about AWS Deepracer is that students learn and enjoy themselves during the race,” Albu said, adding that as AI becomes more common in everyday life, it is worthwhile for students to gain technology experience regardless of what they are studying.
A computer science student, Farass Tabataba Ray was part of the Seneca College team. He has been interested in robotics and coding since an early age, and discovered that participation is illuminated. He learned more about the Python programming language and gained a better understanding of AI.
“It feels so good to see the cords come to life in the truck,” Tabatabaei said.
The Cambrian College shark team finished first, with George Brown in second place. Humber's Techtitans and Team Jarvis came in third and fourth place, respectively.
This AWS Deepracer competition is open to students from all faculties and programs across Humber. Interested students are invited to send an email to Albu at Mihai.albu@humber.ca.
