DeepRacer, which is about the size of a shoebox and designed by Amazon's cloud computing division, Amazon Web Services (AWS), can be preordered for US$249 and will sell for $399.

The designers of DeepRacer have equipped the autonomous car with an artificial intelligence (AI) technique called reinforcement learning, which was inspired by how animals learn from responses to their behavior.

Reinforcement learning is one of the technologies that has made self-driving cars a reality, and can "train" software to react to changing conditions.

The small race car has an HD camera, a dual-core Intel processor and other hardware that will allow it to pilot itself using reinforcement learning.

"This is a technology that has been almost completely out of reach to all but the most well-funded and -motivated organizations," the AWS AI programs leader Matt Wood said on Nov. 28, when the DeepRacer was launched during the AWS re:Invent 2018 in Las Vegas, the United States.

AWS DeepRacer and Amazon autonomous racing league. Amazon tapping into the autonomous developers community on full speed. Nice move and nice development HW! #AWS #DeepRacer #AutonomousVehicle pic.twitter.com/XJaOupvDHh

— Dimitar (mitko) Vasilev‏ (@iotcoi) November 28, 2018

The AWS DeepRacer also includes a fully configured cloud environment that is utilized by the reinforcement learning installed in each race car. The cloud environment is built on Amazon SageMaker, a service that was launched last year and designed to make it cheaper for everyday developers to build, train and deploy machine learning "into a production-ready hosted environment".

Amazon also launched the DeepRacer League, under which it plans to hold 20 races for DeepRacer owners at AWS summits, where they can win credits and win tickets to the championship cup at next year's AWS re:Invent conference.

Trials for the competition already began on Wednesday on a track at the MGM Grand, with the top three finishers to compete the following day during a keynote address.

Amazon plans to use reinforcement learning for more serious tasks in the future. The company said that the technique could be used in industrial scenarios, such as optimizing wind turbine operations under changing weather or power demands, or prioritizing the arrivals of container ships at seaports.

AWS CEO Andy Jassy said that building an AI-based race car would not be a focus for Amazon, and that DeepRacer was just one of multiple ways AI could be applied in everyday life.

"This is still not our focus, although people want us to make more of these things. We just want people to know how these things could be applied," Jassy told a press briefing on Wednesday after delivering his three-hour keynote address.

AWS also showcased the increasingly influential role of AI in traditional car races during the conference.

Formula 1 managing director Ross Brawn said the use of AI had been key to improving driver performance and to make racing more appealing to fans.

"With so much data that we can collect, we can use it to predict results in a race, [such as] whether or not a racer could overtake others," said Brawn.

Earlier in his keynote address, Jassy also unveiled more than a dozen new products and services that could cement AWS' place as the leader in cloud computing.

AWS now controls 51.8 percent of the global market in cloud computing, beating its rival Microsoft, which controls a little over 13 percent, and Google in a distant fourth with 3.30 percent. In third is Chinese internet giant Alibaba, which has 4.6 percent of the market share.

Some of AWS' newly launched services include a blockchain service, Amazon Managed Blockchain. which can manage transactions between organizations without a central party that is responsible for overseeing the activity. Another is Amazon Quantum Ledger Database, a variant of the company's database technology used to track changes within its massive cloud services.

AWS also unveiled a breakthrough in building its own computer chip to handle certain AI processing jobs, which marks an attempt from Amazon to rely less on chip giant Intel.

It also launched Amazon Textract, a service that can automatically extract text and data from scanned documents.

"This will allow the processing of millions of document pages in a few hours," Amazon said in a statement.

Popular

Kadin urges stability after Constitutional Court ruling

Jokowi says transition to new government to begi

BCA loan disbursement surges 17.1 percent yoy in first quarter