How IBM Is Building A Business Around Watson
In 2004, Charles Lickel was eating in a dinner with some colleagues when he noticed that all of the patrons were rushing to the bar. Curious, he followed them to see what all the commotion was about. As it turned out, they were going to see Ken Jennings’ historic six-month run on the game show, Jeopardy!
He was transfixed. Paul Horn, then director of IBM Research, had been bugging Lickel to come up with an idea for the company’s next “grand challenge,” Big Blue’s tradition of tackling incredibly tough problems just to see if they can be solved. The last one drew wide attention when the firm’s Deep Blue computer beat Garry Kasparov at chess in 1996.
The rest, as they say, is history. Seven years later, in 2011, IBM’s Watson beat Jennings and another Jeopardy! champion, Brad Rutter. Today, Watson has become much more than a clever parlor trick, but a potentially huge line of business for IBM. CEO Ginni Rometty expects it to become the heart of Big Blue’s future plans. Yet there are still challenges ahead.
A Brief History of Artificial Intelligence
The field of artificial intelligence got its start at a conference at Dartmouth in 1956. Optimism ran high and it was believed that machines would be able to do the work of humans within 20 years. Alas, it was not to be. By the 1970’s, funding dried up and the technology entered the period known as the AI winter.
Slowly, however, progress was made and by 1992, interest in artificial intelligence revived somewhat. The US government began hosting a series of conferences which posed challenges for question answering or “QA” systems. IBM took part in these conferences and began making advances in a range of techniques.
In the beginning, the researchers experimented with rule-based systems, similar to Doug Lenat’s Cyc project that would answer questions based on information provided by human experts, almost the way an encyclopedia works. However, they soon found that those types of systems don’t scale beyond a certain point.
So they began exploring different techniques that more closely resembled how the human brain takes in information, processes it and makes decisions. For example, deep parser techniques break down sentences into component parts of speech, while support vector machines can mine large amounts of data, learn and begin to draw some conclusions.
Up to that point though, these were isolated projects worked on by separate teams. What Lickel saw that night in the bar was an opportunity to weave it all into a coherent system. “The Jeopardy Grand Challenge problem allowed us to pool all of this disparate work we were doing and focus our energies to see if we could solve a really big problem,” says Eric Brown, who worked on the project and is currently a Director within the Watson division.
Working Through the Jeopardy Grand Challenge
Jeopardy! presents a unique challenge for an artificial intelligence system. First, it covers an impossibly wide variety of topics, so you can’t just train the system to operate within a single domain. The clues are also in complex language, and contain puns and cultural references, which often obscure the meaning of what is actually being asked.
“Hard times,” indeed! A giant quake struck New Madrid, Mo., on Feb. 7, 1812, the day this author struck England.
According to C.S. Lewis, it was bordered on the east by the Eastern Ocean and on the north by the River Shribble.
To answer the first clue correctly, “Who is Charles Dickens?” you would have to realize that “struck England” refers to a birth date and that “Hard times” refers to one of Dickens’ books. To answer the second one, “What is Narnia” you would have to realize that it is a fictional, not an actual geography that is being referred to.
There are other aspects of the game that increase the difficulty even further. For example, you get penalized for wrong answers, so you have to not only come up with a viable response, but also gauge what confidence you have that you have the right answer. There are also time constraints, so you need to be able to respond in just a few seconds at most.
Starting in 2007 with a staff of just 13 researchers, although it eventually grew to more than 25, the Watson team got to work, designing and building an architecture that could analyze the data quickly enough to compete. Over the next four years, they had to not only solve complex technical issues, but also change the way they worked with each other.
“We had to adopt agile techniques in order to develop Watson,” Brown told me, “which as research scientists was new to us. We weren’t just building one system, but had to develop hundreds of algorithms, each an expert on different domains and each bringing a different approach to problems. Then we had to build other systems to weigh the disparate opinions that developed within the system.”
Despite the difficulties, Watson not only won the Jeopardy! match, it trounced the human players over three rounds. For his final response, Jennings wrote, “I, for one, welcome our new computer overlords.”
Building A Business Model Around Watson
After its triumphant victory in Jeopardy!, IBM took Watson to market. One of the first commercial applications was working with Memorial Sloan Kettering Cancer Center and Wellpoint to design an advisory system for its medical staff. Since then, the system has been deployed to a number of top medical institutions, like the Cleveland Clinic and MD Anderson, through Watson Health.
But where the company really sees great opportunity is by offering Watson as a service other companies and developers can access through API’s in order to develop their own applications. “We see Watson as an intelligent engine for our partners to build solutions that will better serve their customers,” Jonas Nwuke, IBM Watson Platform Manager, told me.
To date, the program has attracted over 550 partners, including Satisfi, which is using Watson to help customers navigate retail spaces, Fluid, an online shopping assistant and, Wayblazer, an intelligent travel guide. The developers access the Watson API’s on a metered model, so they only pay for the services they use, which makes the service very startup friendly.
The difference between these apps and conventional recommendation engines is twofold. First, they can analyze unstructured data, like product descriptions and customer comments. Second, they are able to learn user preferences. So, for example, if a hotel gets some bad ratings because people complain about screaming children, but I’m often looking for a place I can take my screaming child, Watson can learn that about me.
Growing Up Watson
Like any precocious young prodigy, Watson now needs to find its place in the world. It cannot, as many suspect, replace the role of human professionals. There are certain things that machines will probably never be able to do, like show the genuine empathy required to understand, interact and build effective working relationships with people.
Still, the potential is undeniable. Think about how much more effective an ordinary doctor can be with Watson as an assistant. First, even before the patient enters the room, it can analyze their personal medical history, which often runs to hundreds of pages. Then, it can compare the case history with the 700,000 academic papers published every year as well as potentially millions of other patient records.
All of this is, of course, beyond the capabilities of human doctors, who typically only get a few minutes to prepare for each examination. So being able to consult with Watson will be enormously helpful. At the same time, as doctors provide feedback as to which of Watson’s recommendations are helpful, the system continues to learn as it now also does with travelers, shoppers and others.
So while robots are unlikely to become our overlords, they do have the potential to become immensely valuable collaborators. “One of the ongoing goals of the system is to overcome human biases,” Brown told me. “So Watson is not only giving answers it is also, in some cases, posing questions to human conventional wisdom.”
The potential for this collaboration is enormous. A doctor that can consult with a system that, for all intents and purposes, can make all of human medical knowledge available in an instant, will be able to spend more time actually caring for patients. They can become healers once again, rather than merely technicians.
I, for one, welcome our new robot collaborators.