Amazon Web Services trains its spotlight on artificial intelligence
By Edwin Yapp December 2, 2016
- AI services to bolster, extend AWS’ cloud dominance despite it being the leader
- Still early days for AI; AWS, Microsoft, Google, IBM clamouring for leadership
CLOUD vendor Amazon Web Services (AWS) Inc has unveiled a slew of new cloud-based services, the most eye-catching of which is a series of new artificial intelligence (AI) services that is specifically tailored for developers to build context-aware and voice-interactive applications, in a bid to extend its dominance in the cloud computing game.
Speaking at its ongoing cloud conference re:Invent 2016 here in Las Vegas, chief executive officer (CEO) Andy Jassy (pic, above) said the three new AI-based services – Amazon Rekognition; Amazon Polly; and Amazon Lex – are designed to let developers build apps that can understand natural language, turn text into life-like speech, have conversations using voice or text, analyse images, and recognise faces, objects and scenes.
Jassy, who was promoted from senior vice president to helm AWS in April, said the company has long been involved in AI and machine learning services, and is now primed to offer its technology to developers and customers.
“What we found is that many don’t realise the heritage we have in the matching learning space just because we don’t talk about it very much,” he said in his keynote address. “But if you think about it, AWS has deep heritage in this space and we have a 1,000 people dedicated to AI.”
Jassy said one of the first features AWS’ parent Amazon.com Inc had in its e-commerce site is its recommendation engine, where the company suggests to its customers to buy things that other people had bought. This is essentially driven by AI, Jassy noted.
According to AWS, very few developers have been able to build, deploy, and broadly scale apps with AI capabilities because doing so required access to vast amounts of data and specialised expertise in machine learning and neural networks.
These services will eliminate all of this heavy lifting, making AI broadly accessible to all app developers, the company claimed. They will gain access to AWS’ deep learning algorithms and technologies as fully managed services, which any developer can access through an API call or a few clicks in the AWS Management Console, it said.
The Amazon AI services make the full power of Amazon’s natural language understanding, speech recognition, text-to-speech, and image analysis technologies available at any scale, for any app, on any device, anywhere, the company added.
How it works
The three AI services introduced on Nov 30 are:
- Amazon Rekognition: This service makes it easy to add image analysis to applications, using powerful deep learning-based image and face recognition;
- Amazon Polly: This service transforms text into life-like speech, enabling apps to talk with 47 lifelike voices in 24 languages; and
- Amazon Lex: This service is the same one that powers Amazon Alexa, which is found in Amazon’s personal assistance Amazon Echo, enables any developer to build rich, conversational user experiences for web, mobile, and connected device apps.
Rekognition and Polly services are available from Nov 30 while Lex is still in preview. All three services are currently limited to United States and some EU-based countries for now.
Demonstrating these services on stage during the keynote, Jassy said Rekognition could differentiate objects within an image, and can “pick out an image, say, of a woman, a car, a steering wheel and from that, [a user] is able to search for images of a women driving a car.”
Amazon Polly takes words provided in text form and translates them to voice editing them along the way, Jassy explained.
“If you typed a question about the temperature in ‘WA’ it will know that you are referring to the state of Washington,” said Jassy. The service will produce an MP3 audio file converting what you’ve typed in to voice, thereby intelligibly interpreting ‘WA’ as Washington. Polly is able to operate in 47 voice options speaking and in 24 languages.
Lex meanwhile is an automatic speech recognition service, which users can use to build conversational, intuitive interfaces to your applications and business data, said Matt Wood, general manager of product strategy for AWS.
The technology behind Lex is what powers Amazon Alexa, the company's virtual assistant found in its Amazon Echo and Amazon Echo Dot products. These products has surprisingly taken world by storm, having sold about 3 million units since it was introduced in 2014.
Demonstrating this on stage at the keynote, Wood said this would make it easier for one to build a travel booking application by speaking into it instead of relying on typed queries. Based on past interactions, the Lex service would know things about you, such as your preferred airline or whether you would like a window or aisle seat, Wood explained.
Next page: Clash of the Titans