By creating a cognitive computing system that could play Jeopardy, IBM charmed the world with Watson, the system that beat two top champions of the quiz show. Since that famous game in February, 2011, IBM has engaged Watson on more practical pursuits, including solutions for the healthcare and finance industries.
But we’re also using Watson to solve our own problems. In an organization as large as IBM, one of the biggest challenges is knowledge sharing. In IBM, it is generally true that for any technology question, there is at least one person in the company with the correct answer. But finding that person is too often impossible.
Over the years, there have been several solutions to this knowledge sharing problem. The proliferation of wikis and online communities is the most current attempt to provide a repository of knowledge and expertise. While these tools are immensely helpful and go a very long way toward solving the knowledge sharing problem, users still struggle to navigate this vast data source. Successful navigation requires some prior knowledge of who the experts are, and their ontology, or how they logically structure and organize their information.
So we know we have experts, data, and answers to just about every question. But we can’t find a tool to help us sift through that vast store of information.
Watson excels as culling through pedabytes of information and deriving meaning from disparate sources. Watson can associate people with areas of expertise, and can place information in a historical context. Watson, therefore, is the ideal cognitive system for IBM’ers trying to solve problems for clients.
But Watson requires care and feeding to get to that seemingly magical state of expertise. The data store is built by submitting thousands of questions and providing links to the correct answers. It also helps to give Watson a data corpus for a subject area, even something as broad as cloud computing.
Imagine, if you will, that you had the opportunity to submit questions and answers for Watson about cloud computing, What questions would you want Watson to be able to answer? There are obvious questions, like “What is cloud computing?” and there are thousands of questions related to the technical depths below the umbrella phrase of “cloud computing.” But what about less obvious questions? For example, is there agreement on who first coined the term “cloud computing?”
In the next few weeks, I’ll be contributing to the database of cloud computing questions for Watson. If you have questions you think I should include, feel free to post them here.