IBM‘s Watson, an artificial intelligence computer system which famously defeated two all-time Jeopardy! quiz show champions, may soon go from playing games to answering questions in the workplace and call centers as IBM looks to modify the technology for business.
Watson was powered by distributed computing technologies such as Apache Hadoop, which was used for pre-processing, to answer the game’s queries on diverse subjects from science to popular culture by searching through 200 million pages of structured and unstructured information taking up 4 terabytes of disk storage. It had to do so faster than the three second time limit, and had to understand the show’s host, Alex Trebek. In other words, Watson was capable of understanding complex spoken queries and finding results rapidly from massive databases.
The practical application of such technology is diverse and vast. One of the first fields IBM set to apply Watson was medicine, where eight universities are working on turning the system into an automated physician’s assistant, hearing a patient’s symptoms and then firing back questions to narrow down possible diagnoses. Another proposed use has been as a replacement for “phone trees” at call centers. Currently, automated call centers mean that customers have to choose from limited, numbered options then wait for a new set of choices until they reach their specific query, which is imprecise, frustrating, and time consuming. Watson-like technology would allow them to list what’s wrong as if they were speaking to a person while the computers search through documentation and past problems. Another proposed application is allowing marketing departments rather than statisticians, programmers, and data whizzes work with business intelligence and analytics, as Watson would allow them to phrase their queries in natural language and then dive into the Big Data for them.
Yet these are only the few possible applications of such technology currently in the media. By solving several common problems, from speech and language recognition to fast Big Data analysis, Watson could be applied to almost any field, such as intelligence, helping analysts quickly test hypotheses. or law, sifting through past cases for precedent faster than any paralegal. With data everywhere growing constantly bigger and more complex, technology like Watson has the potential to revolutionize how we ask and answer questions.
- Calling for tech support? IBM’s Watson might be on the other end (engadget.com)
- IBM’s Watson may soon be answering your customer service calls (digitaltrends.com)
- IBM researcher talks up Watson’s potential (venturebeat.com)
To clarify, this post was written by Alex Olesker. The connectors from my blog to govloop re-post these things under my name, but Alex writes all the good stuff. See more from him at: http://ctovision.com/author/alexolesker/
Today, the burden of sifting through the tremendous amount of information to answer questions and make decisions rests with us. Sure we have search engines but they bring back even more information, faster than ever. We need to change the paradigm so that the computer (Watson) helps provide us with likely answers and provides the “evidence” so we can decide if we want to trust the proposed answers.
I provided some additional thoughts on our Govloop Analytics Forum
Something tells me this will be an expensive technology if it’s sold the way of software. If I were IBM I’d turn Watson into a free service (easy to say when you aren’t writing the checks) to compete as a spoken search engine.
I’ve seen this movie before. Wasn’t it called “2001:A Space Odessy?”