IBM is one of the oldest tech companies in the world. Now, it’s banking on one system to be a major cash cow: a supercomputer called Watson. IBM hopes Watson will be as revolutionary as the personal computer or the mainframe. And they’re making a big investment in the supercomputer, creating the Watson Group in January and backing it with a billion dollar investment.
You might remember Watson from the trivia game show, “Jeopardy”. In 2011, the supercomputer beat two Jeopardy world champions on national television. It was a pivotal moment that showed the world cold listen, think and communicate – kind of like a human. “Jeopardy showcased how a cognitive computer could navigate the complexities of human speech,” says Steve Gold, Vice President of IBM’s Watson Group. “[It] could investigate large and vast amounts of information and in that case 200 million pages of data in a little under three seconds – and how it could learn.”
Launching IBM’s new Watson Group comes at a crucial time for the 103-year-old company. In May, IBM’s CEO admitted it’s been a “rocky time” for the company. Revenue growth at IBM has been slow, and new technologies are threatening some of the company’s traditional businesses in hardware and software.
Since Watson’s “Jeopardy” appearance, the supercomputer has undergone a profound transformation. It’s shrunk in size, from around 90 servers during its “Jeopardy” days, to a system today that can be accessed on an iPad.
Dr. Mark Kris is Chief of Thoracic Oncology at Memorial Sloan Kettering Hospital in New York City. He’s training Watson to check patient information against vast amounts of medical literature to suggest top treatment options for lung cancer patients.
“I’m one of Watson’s teachers,” said Dr. Kris. “I help it collect the medical literature, I help keep it up to date.”
Watson is good with big data. It can take in high volumes of information about any topic: medical literature, financial data, retail information and then help a person make highly informed decisions. But there are kinks in the system.
“Medical literature is growing exponentially, and trying to find it and get it in a timely manner is a struggle,” said Dr. Kris. It’s a lot of work, but once the data is loaded into Watson, it can make treatment recommendations based on an individual’s medical history. Watson ranks several treatment suggestions and will explain why other choices aren’t as good. Dr. Kris then teaches the supercomputer which recommendations work, and which do not – and Watson remembers for the next time.
This is an essential part of Watson’s ingenuity – even outside of the healthcare realm. But it could also be a detriment to potential customers. “Watson isn’t a plug and play system that you plug into a room ad it starts working,” said Peter Wahlstrom, Sr. Technology Analyst with Morninstar. “There’s a lot of initial interaction that needs to happen to make Watson ‘smarter’.”
Error
Sorry, your comment was not saved due to a technical problem. Please try again later or using a different browser.