Finalists of the inaugural HTX x AWS DeepRacer Challenge 2022 striking the quintessential HTX “X” pose. (Photo: HTX/ Heng Shu Yun)
Xponents, ready, get set, go!
The buzz and excitement in the air at HTX’s headquarters on Friday afternoon was palpable as an autonomous miniature race car sped around a track that had been set up for the inaugural HTX x AWS DeepRacer Challenge 2022 finale race on September 23, 2022.
The finale race was the culmination of weeks of learning and experimentation with Reinforcement Learning – an advanced Machine Learning technique – by HTX officers (or “Xponents” as we call ourselves). The DeepRacer Challenge was organised by the HTX House Challenge (HTXHC) Organising Committee, in collaboration with Amazon Web Services (AWS) Singapore. AWS DeepRacer is an autonomous 1/18th scale race car designed to test Reinforcement Learning models by racing on a physical track.
“The DeepRacer Challenge is unlike any other challenge that HTX House Challenge has organised over the last two years,” said HTXHC Organising Committee Chairman Ng Shu Herng, who is also the Director of the Protective Security & Safety Centre of Expertise at HTX. HTXHC provides all Xponents, who belong to one of five Houses (Water, Space, Fire, Earth, and Air), opportunities to build camaraderie beyond their work divisions through games and activities.
Xponents cheering on the finalists at the finale race. (Photo: HTX/ Heng Shu Yun)
Gearing up on machine learning
More than 300 Xponents from technical and non-technical backgrounds signed-up for the Challenge, which included workshops and immersion visits to AWS. Over 200 Xponents submitted their Reinforcement Learning models in a community race, consuming more than 3,000 training hours collectively on AWS Cloud in the process.
“Cloud technology, AI, and Machine Learning are key to what we do at HTX and the DeepRacer Challenge has been a great way for Xponents from technical and non-technical backgrounds to learn and explore Machine Learning in a fun and interactive way,” said HTX Chief Executive Chan Tsan. “We are an exuberant bunch, and we love a good competition.”
AWS Singapore Country Manager, Worldwide Public Sector, Elsie Tan, who joined Xponents at the finale race said the DeepRacer Challenge was about getting hands-on into how technology can be a part of everyday life and having fun learning together. “Games like this showcase the power of Cloud and how it extends to many functional uses. We are so excited to partner with HTX on your Cloud journey and grow with you.”
(From left to right) HTX House Challenge Organising Committee Chairman Ng Shu Herng, AWS Country Manager, Worldwide Public Sector, Singapore Elsie Tan, and HTX Chief Executive Chan Tsan at the HTX x AWS DeepRacer Challenge 2022 finale race at HTX headquarters (Photo: HTX/ Heng Shu Yun)
Exciting mission accomplished
The finale race saw twenty Xponents – comprising four from each House – pitting their skills and cars against one another in a closely contested race with House Air sweeping two of the three top positions.
House Air’s Koh Buck Hui, Head, Joint Wireless Comms, ICT Infra, emerged as the champion with an impressive timing of 7.78 seconds.
“It was a great learning experience,” said Buck Hui, who had no prior experience with Reinforcement Learning and did not expect to win the finale race. “I felt a sense of achievement when I was able to train the model to achieve even just a 0.01-second improvement and was able to climb the leaderboard.”
House Air’s Koh Buck Hui (right), Head, Joint Wireless Comms, ICT Infra, who placed first in the finale race, receiving his prize from HTX Chief Executive Chan Tsan. (Photo: HTX/ Heng Shu Yun)
House Space’s Marie Chan, Head, Narcotics, Corrections & Narcotics Programme Management Centre, was second with a timing of 8.11 seconds, with House Air’s Eunice Chan, Scientist, CBRNE Centre of Expertise a close third at 8.31 seconds.
Accelerating the digitalisation movement
House Fire’s Jonathan Goh, Senior Manager, Cybersecurity Audit, sharing with Xponents about his DeepRacer experience and lessons learnt. (Photo: HTX/ Heng Shu Yun)
For many Xponents, the Challenge drove them to explore and pick up Reinforcement Learning through the workshops and AWS immersion sessions that included hands-on labs on Cloud, AI, and Machine Learning.
“Most people will not touch Machine Learning during the course of their work, so it’s fascinating to catch a glimpse of what goes on in this domain,” said House Air’s Lee Jia Shing, Engineer, Protection & Special Ops Capability, Vehicle Systems, Land Systems Centre of Expertise, who works on autonomous vehicles.
“The DeepRacer challenge helped me learn the basics of Machine Learning, and getting to tweak and improve my own car is very fun and enjoyable,” said Jonathan Goh, Senior Manager, Cybersecurity Audit, who represented House Fire.
“The inaugural HTX DeepRacer Challenge has demonstrated how we can support and learn from each other, and gain knowledge on Cloud technology and Reinforcement Learning in a fun way,” said HTXHC Organising Committee Chairman Shu Herng. “We look forward to the DeepRacer Challenge becoming a regular fixature in the HTX calendar of events.”
Xponents representing the five Houses – Water, Space, Fire, Earth and Air – and AWS representatives at the HTX x AWS DeepRacer Challenge 2022 on September 23, 2022. (Photo: HTX/ Heng Shu Yun)