Hoping to attract commercial IT shops to its Watson cognitive computing platform, IBM snapped up a small company...
that could supply the technology pieces to make that breakthrough.
Perhaps more importantly, AlchemyAPI brings along some 40,000 corporate and third-party developers -- compared to IBM's 6,000 Watson developers -- that have written applications for its platform. Many of those applications span a much broader range of markets than IBM has cracked with Watson thus far.
"This deal gives [Watson] a wider set of multilingual text analytics APIs that have been field-tested by a large number of cloud-based users," said Dave Schubmehl, research director for IDC's Content Analytics, Discovery and Cognitive Systems. "The other interesting advantage is the whole unsupervised deep learning capability that pulls out Web resources or unstructured information that can be used in a knowledge graph including relationship extractions."
The addition of unsupervised learning capabilities should help speed up the process of teaching Watson applications about all the specific capabilities of a particular app. So far, Watson has been built exclusively around supervised learning.
"This gives IBM some of the capabilities Google and Microsoft have around their knowledge graph work," Schubmehl said.
Some IBM users are hopeful the acquisition will result in delivery of practical applications that fit their needs, but they don't expect that anytime soon.
"I think there is some interesting technology at work there, but I don't see the relevant application of it for us yet," said one purchasing agent with a Bay Area shipping container company. "It could be something we are interested in down the road."
Together with IBM, AlchemyAPI plans to pursue opportunities in broader commercial markets such as finance, insurance, retail and travel, said Elliot Turner, the company's founder and CEO.
"We have some missionary work to do to educate people, but these applications can change how these industries process and understand data so they can make better decisions," Turner said.
Alchemy API injects intelligence into Watson
With IBM's resources behind it, AlchemyAPI will focus on deep learning technologies, which, as Turner explains, is the ability to leverage algorithms known as deep neural networks that make it possible for systems to train and adapt themselves directly from the data. This minimizes the typical amount of human involvement needed for systems to adapt to new niche industries and domains.
"We can then build out core capabilities with a knowledge graph, which gives us the ability to read documents like a human and to see the deeper meaning of what those words are saying, as well as the interconnections represented in those documents," Turner said.
The acquisition should bolster IBM's efforts to make Watson more attractive to companies building their own technologies on top of the Watson platform, according to Tom Austin, a research fellow at Gartner, Inc. in Stamford, Conn. So far, IBM has attracted only users and business partners at the high end of the food chain, or those wanting to create enterprise-class applications that Watson can consume, Austin said.
"This acquisition should help change that around to the extent that IBM motivates AlchemyAPI corporate developers to exploit Watson's cloud services," Austin said.
The AlchemyAPI purchase follows a couple of other Watson-related moves IBM made at its InterConnect 2015 conference late last month. The first was the unveiling of the Watson Zone, a resource center that works with BlueMix and Watson APIs, sample code and examples to help guide users in building hybrid cloud applications laced with cognitive computing features.
The second was the Watson Personality Insights service which allows corporate users to integrate new capabilities to enable analyzing trends and patterns in complex social media data streams.
About the author:
As Senior Executive Editor, Ed Scannell is responsible for overseeing the day-to-day operation of several TechTarget sites in the Data Center and Virtualization media group, as well as editing and writing news, features and other technical stories.