IBM’s supercomputer Watson, named after founder Thomas Watson, will take on champions of the quiz show Jeopardy starting tonight. Nearly 15 years after an IBM computer beat world chess champion Garry Kasparov, Watson has a good shot at upsetting the reigning humans. Here’s a video to show just how impressive the computer is under pressure:
Fun and games? Literally, but serious business, as well. The powerful analytic tools that IBM has put into play also have implications for system design and business analytics. The ability to find answers to questions is a critical one for any corporation, suggesting that those with a business intelligence degree should expand their view to look at advanced programming and artificial intelligence.
Business intelligence is the IT discipline of getting relevant information to decision makers in a timely fashion. Ultimately, that really means getting answers to questions. That’s what IBM engineers wanted to do with Watson: natural language processing. As Craig Rhinehart, a manager in IBM’s software group, writes about the computer:
Natural language processing (NLP) is a field of computer science and linguistics concerned with the interactions between computers and human (natural) languages. It describes a set of linguistic, statistical, and machine learning techniques that allow text to be analyzed and key information extracted for other uses such as Question Answering or Content Analytics.
If a system cannot understand the question, it cannot possibly find the answer. Often, people with information technology degrees focus on either back-end data processing systems to better crunch information or user interfaces to make software easier to use. The Watson Jeopardy challenge demonstrates the critical need to combine understanding of freestyle English with the ability to quickly manage overwhelming volumes of data — the equivalent of 200 million pages of information — to find the one right answer to a question.
The process of translating spoken language into text, understanding the question, finding the answer, and then speaking takes about three seconds. A desktop computer would take two hours to perform the same amount of work.
IBM did not spend four years building the technology behind Watson solely to answer trivia questions on television. The company plans to work with multiple universities to expand the question answering capabilities. The technical areas involved include the answering architecture and methodology, text search and information retrieval, a visualization component, machine learning, conversational agents, and an interactive question and answer system.