IBM announced late Wednesday that it's making its artificially intelligent computer system, Watson, available to researchers as a cloud service.
Scientists from universities, pharmaceutical companies and commercial research centers have been using Watson, which was built to understand human language, to analyze and test hypotheses in their data, along with data held in millions of scientific papers available in public databases.
Early adopters have been trying out the cloud service, but it's officially available today, according to Rob Merkel, vice president of IBM's Watson Healthcare Group.
Merkel declined to talk about the cost of the service.
Watson gained mainstream fame early in 2011 when the supercomputer went up against Jeopardy champions in a special episode of the question-and-answer game show.
In the man-vs-machine dustup, Watson trounced its human opponents. The machine may have faltered in a few categories, but was faster to the buzzer and more knowledgeable than its challengers, who had won many games against knowledgeable opponents in regular matchups.
At the time, Watson was touted by some analysts as one of the biggest computing advancements in the past several decades.
What makes it stand apart from other supercomputers is not just its ability to make calculations. Watson was designed to essentially converse with humans, answering verbal questions and even beginning to understand colloquialisms and jokes.
Merkel said that natural language ability puts Watson in a good position for scientific research. For instance, a scientist could have Watson digitally ingest as much information - say, research papers, proprietary information and licensed information -- about a topic as possible.
Then the scientist could ask the super computer to find all the drugs that had been repurposed for a particular use in the past five years. Or the scientist could ask Watson to go through that information and find all of the known drugs with certain characteristics.
"It's about understanding human language, scientific language and images," said Merkel. "It could be used anywhere where huge bodies of information need to analyzed."
Patrick Moorhead, an analyst with Moor Insights & Strategy, said it's a good idea for universities or commercial research houses to use Watson as a cloud service.
"Going with software-as-a-service lets you try it out, lowering the risk of bringing infrastructure in-house before you test it," he explained. "This is a good way for universities and pharmaceutical companies to kick the tires and see if Watson does bring value and if it works at scale. Meaning if they would do a lot of work on it, possibly they would bring it in house."
That, according to Moorhouse, would be a great thing for IBM, which would like to sell more Watson-like systems.
Dan Olds, an analyst with The Gabriel Consulting Group, said using Watson as part of a cloud service actually is a great idea for IBM.
"Delivering Watson as a service is a much better way to get potential clients to give the technology a try," he said. "It's an easy first step that will allow customers to see if Watson is for them, without having to shell out a whole bunch of money for what is essentially a supercomputer."
This story, "IBM Uses Watson as Part of New Cloud Service" was originally published by Computerworld.