It's been the year of the AI announcement and Amazon Web Services (AWS) wasn't about to buck the trend at its re:Invent conference in Las Vegas this week. So it was that
It's been the year of the AI announcement and Amazon Web Services (AWS) wasn't about to buck the trend at its re:Invent conference in Las Vegas this week. So it was thatAndy Jassy announced a trio of new AI services, designed to help developers take advantage of the company’s machine learning expertise.
The first is image analysis service, Amazon Rekognition. This enables developers to include visual search and image classification in their applications, so that users can search and compare human faces, for example. According to the company, it’s based on the same technology developed by Amazon’s scientists to analyse billions of images daily for Prime Photos, its photo storage service.
The second is Amazon Polly, a text-to-speech service. Text submitted to Polly is returned as a voice file in a standard audio format, such as MP3, in a choice of 47 voices across 24 languages. In a call center environment, the Polly API might be used to deliver information to customers (such as service status and contact information) in a natural-sounding voice.
The third product is Amazon Lex, a service for building conversational applications built on the same automatic speech recognition (ASR) and natural language understanding (NLU) technology that powers Amazon Alexa, the company’s ‘intelligent personal assistant’ device. Developers might use Lex to build applications that perform automated tasks in response to a voice command - for example, ‘book me a flight’ or ‘check the weather in San Diego’. They can also use pre-built connectors to ask questions of enterprise applications such as Microsoft Dynamics and Marketo - for example, ‘What are my top ten accounts in Salesforce?’
Lex may well be the most important of the new services launched today. According to a blog entry posted by AWS chief evangelist Jeff Barr:
You can use Amazon Lex to build chatbots and other types of web and mobile applications that support engaging, life-like interactions. Your bots can provide information, power your application, streamline work activities, or provide a control mechanism for robots, drones and toys.
In his keynote, Jassy was keen to hammer home the point that parent company Amazon already uses machine learning extensively to run its own logistics operations and in the discovery and search functions that enable online shoppers to find products they want to buy, as well as the Alexa voice service that powers its Echo smart speaker. But it arguably hasn’t done a great job before of publicizing its smarts in this area. As Jassy said:
A lot of companies don’t realize the heritage Amazon has in the machine learning space.
That’s clearly got to change, because AWS’s closest competitors in the cloud services market are talking up a storm when it comes to AI and machine learning.
Two weeks ago, Google Cloud chief Diane Greene announced the launch of a dedicated machine learning unit and the recruitment of two high-profile AI researchers to lead it. They are Fei-Fei Li, director of Stanford University’s Artificial Intelligence Lab, and Jia Li, head of research at Snap, the parent company of messaging app Snapchat. Greene reportedly said that “one of the most thrilling things” about the new hires is that both are women, in a fast-emerging field where the gender imbalance remains particularly striking.
In the same week, Microsoft announced it had signed a partnership with OpenAI, the AI research group set up by Tesla frontman Elon Musk. This collaboration will see OpenAI use Microsoft’s Azure cloud service for its large-scale experiments and has come about, at least in part, due to Microsoft’s work on supporting AI workloads with tools such as Azure Batch, Azure Machine Learning and Microsoft Cognitive Toolkit. OpenAI’s investors, interestingly enough, include AWS, along with infamous Trump supporter Peter Thiel, LinkedIn cofounder Reid Hoffman and Infosys, among others. Collectively, they’ve backed the venture to the tune of $1 billion.
Still, today’s announcements from AWS certainly don’t represent the company’s first foray into supporting AI workloads. The company’s Amazon Machine Learning service was launched around 18 months ago, as Gavin Jackson, managing director for UK & Ireland points out:
The message that Andy was putting across today is that this [AI] is something we’ve been doing for a long time. We’ve been building our [own] applications that serve our customers with AI and machine learning and deep learning and so on for a long time and it’s now become very popular. So what we’re trying to do is to take all of that learning and build it into a platform, so that other companies can build their own applications on top of it - so you’ll see far more applications with built-in intelligence.
So Rekognition, Polly and Lex join Amazon Machine Learning as part of a wider family of products called Amazon AI. However, for now, the availability of the new products is somewhat limited. Rekognition is available from today in three regions: US East (serviced by the North Virginia data centre); US West (Oregon) and the EU (Dublin). Polly is available in four regions: US East (N. Virginia); US East (Ohio), US West (Oregon) and the EU (Dublin). Both services will be rolled out to additional regions in the coming months. Customers can sign up for the Amazon Lex preview starting today.
In other news from Reinvent, Workday co-founder and CEO Aneel Bhusri took the stage to announce that the cloud-based human capital management (HCM) vendor has selected AWS as its preferred public cloud infrastructure, not just for testing and development (where it has long used the service), but also for customer production workloads.
This is the first time Workday has offered customers the option of running its full suite of applications in the public cloud, said Bhusri:
The ability to run our applications on the AWS Cloud underscores our commitment to deliver the security, reliability and continuous innovation our customers expect when it comes to infrastructure, while continually expanding the ways they can optimize workforce productivity and business performance.