Deep Blue's legacy
Two years ago, IBM's Deep Blue™ made history when it outplayed the world chess champion, Garry Kasparov. What captured many people's attention, however -- beyond the drama of the man-machine competition itself -- was the prospect of comparable achievements in other areas. Deep Blue provided the first clear demonstration that sufficiently powerful machines, together with innovative programs fine-tuned by experts, could give computers an edge in realms hitherto reserved for humans.
Success at the chessboard also showed that computers do not have to possess "artificial intelligence" to give the impression of human intelligence. What counts, aside from sheer computational power, are algorithms, or rules, that enable computers to uncover, in vast collections of data, patterns and relationships that elude direct perception. Such an approach, we are discovering, can lead to deeper insights into the state of a business, the creditworthiness of potential customers or the path to creating more effective drugs.
We call this activity Deep Computing. Today, we can point to major successes in solving real-world problems arising in businesses as diverse as airlines, pharmaceutical firms, supermarkets and insurance companies. The applications are not only changing the way businesses operate but opening up novel possibilities for individuals, from customized weather forecasting at the neighborhood level to cancer drugs optimized for one's own genetic makeup.
Deep Computing can extend the capabilities of experts. For instance, some of the most complex problems companies face every day concern scheduling, whether it is dispatching vehicles to minimize costs, staffing to meet variable demand or optimally assigning airline crews to flights. Our scientists have solved such problems by developing novel optimization algorithms tailored to the needs of individual companies. With computers taking over the most labor-intensive aspects of these tasks, scheduling experts can begin to experiment with alternative scenarios, react more quickly to unforeseen changes and gain a clearer understanding of their options. Huge efficiencies can result.
As the seas of data rise -- largely in response to the proliferation of computing devices -- and answers are sought to increasingly complex questions, the demands on computing will grow. Deep Computing will enable us to continue to extract meaning, find optimal solutions and gain novel insights. Companies will use Deep Computing as a means to differentiate themselves in a world where data is plentiful, but knowledge is hard to find. And it will provide answers to problems that will ultimately improve the health and well-being of us all.
Paul M. Horn
senior vice president and director,
