Not written your list to Santa yet? Good job because here’s a toy you might want to add to the list – make way for Amazon DeepRacer, a fully autonomous, 1/18th scale radio-controlled car, designed to help you learn and understand machine learning (it’s priced at $399 so let’s hope Santa’s feeling generous)!
Two years ago, Alphabet researchers made computing history when their artificial intelligence software AlphaGo defeated a world champion at the complex Chinese board game ‘Go’. Amazon now hopes to democratise the AI technique behind that milestone with a pint-size self-driving car.
Launched at AWS Re:invent last week, the car comes with a HD camera, a dual-core Intel processor, and other hardware it needs to pilot itself—but a blank slate where its driving skills should be. Programmers must help it learn those, using new Amazon tools to support reinforcement-learning projects.
Reinforcement Learning is one of the technologies that are used to make self-driving cars a reality. It works by enabling algorithms to pick up skills through repeated trial and error. They are guided by feedback from a “reward function” that provides a kind of simulated motivation—for example, by telling the software it must try to maximise its score or lift objects without dropping them. Over many attempts to win a virtual sumo bout or use a robot gripper, the software can gradually improve at achieving the goal it was set. The fact it can train software to react appropriately to changing conditions makes it a good fit for industrial scenarios, such as optimising wind turbine operations under changing weather or power demands, or prioritising ship and container scheduling in ports.
However, it can take millions of failures for a reinforcement-learning system to become proficient, so most projects using the technology depend on simulations to speed up the laborious process. Therefore, programmers who want to play with Amazon’s DeepRacer must first train their code in a virtual world created by Amazon for the project, in which a digital double of the car can drive—and crash—over and over.
Developers first get started using a virtual car and tracks in a cloud-based 3D racing simulator, powered by AWS RoboMaker. Here, they can train an autonomous driving model against a collection of predefined race tracks included with the simulator, then evaluate them virtually or choose to download them to the real-world AWS DeepRacer car.
The first AWS DeepRacer League took place at the re:Invent conference, but if you weren’t one of the lucky few to take part, don’t fret, the league will continue with a series of live racing events starting in 2019 at AWS Global Summits worldwide. Virtual tournaments will also be hosted throughout the year with the goal of winning the AWS DeepRacer 2019 Championship Cup at re:invent 2019.
Amazon say they have launched this car because machine learning’s biggest problem is that most organisations either don’t know what is possible, or don’t trust AI because they don’t understand how it works. DeepRacer aims to solve both of these problems by enabling organisations to get hands-on experience with reinforcement learning and encourage them to apply it to weightier business challenges.
Only time will tell to see how much of an impact DeepRacer will have, but it is great to see AWS exploring different ways to help organisations understand and embrace ML in a fun environment which can then be applied to help solve the world’s biggest challenges.