Machine-learning is already starting to reinvent modern business, and the technology is poised for massive growth and development over the next few decades. E-commerce giant Amazon highlighted that trend on November 30 at the company's developer conference AWS re:Invent, when Amazon Web Services CEO Andy Jassy announced Amazon would open three of its AI services to developers everywhere.
These hosted services – an image analysis tool called Amazon Rekognition; a text-to-speech tool known as Amazon Polly; and Amazon Lex, which helps developers build chatbots – mean that developers don't need a background in machine-learning to use Amazon's systems when building new applications. Consider it AI as a service: Amazon will handle the machine-learning, so developers can focus on building new, more powerful software. [See Related Story: AI Comes to Work: How Artificial Intelligence Will Transform Business]
"Thousands of machine-learning and deep-learning experts across Amazon have been developing AI technologies for years," Raju Gulabani, VP of databases, analytics, and AI for Amazon Web Services, said in a statement. "Now, we are making the technology underlying innovations [like Alexa] available to any developer in the form of three fully-managed Amazon AI services that are easy to use, powerful and cost-effective."
Let's take a deeper look at each service and what Amazon purports that they offer.
Amazon Rekognition allows developers to build applications that analyze and recognize images. Using deep-learning, Rekognition can identify human faces, inanimate objects and entire scenes. It also provides a confidence score along with each identification. When it comes to faces, Rekognition can determine expressions and individual attributes, as well as compare and search specific faces. Amazon offered an example of an application developers can build with Rekognition by imagining a system that measures the likelihood that a face in two separate images belongs to the same person. Developers pay for each image they analyze and the facial feature vectors they ultimately store. [See Related Story: 6 Incredible Ways Businesses are Using Artificial Intelligence Today]
Amazon Polly is able to translate text to speech in a natural way, even recognizing homophones and identifying the appropriate term to use based on the context of a sentence. The system includes 47 male and female voices and supports 24 languages, according to Amazon, and Polly can also use a variety of accents. Amazon claims Polly speaks in a fluid, conversational manner, just as a human would if they were to read the converted text. Developers pay based on the text they end up converting to speech, but can archive any converted text for later use at no cost.
Amazon Lex is based on the same voice recognition technology behind Amazon's personal AI assistant Alexa. The company envisions Amazon Lex as a tool for developers whom are looking to build and test conversational bots, which could be used in a variety of applications, from booking a flight to serving retail customers. Lex could also be harnessed to retrieve and structure data through systems like Salesforce, Microsoft Dynamics, Marketo, Zendesk, QuickBooks, and HubSpot. The system is also integrated with communication applications like Facebook Messenger and Slack. The company claims Amazon Lex scales along with traffic and that developers only pay for the calls made to the Amazon Lex API.
Jassy signaled during the November 30 keynote address that Amazon intends to add more services to its public AI platform in the future.