In the Information Age, the future belongs to algorithms—those little sequences of instructions that help computers sift through data and solve problems. Two Stanford Ph.D. candidates devise an algorithm that uses the relationships between Web sites to rank search results, for example, and presto: Google is born. But as sophisticated as Google is, it can’t really think for itself. And that’s where scientists like Daphne Koller come in.
A mathematical theoretician at Stanford, Koller is, in the words of one colleague, “on the bleeding edge of the leading edge” in the world of artificial intelligence. As the ever-increasing amount of information we collect threatens to overwhelm us, Koller, 41, is busy devising computational tools for computers that reveal telling connections between tiny data points and “teach them how to use logic and statistics to make reasonable assumptions on their own.” The potential applications for this kind of algorithm are endless. Koller’s work already has helped scientists improve machine vision, understand the spread of cancer, and predict traffic jams. Next up: programs that scour patient medical records for the largely invisible indicators that typically presage a complication and speed the development of new drugs by scanning the human genome for signs of sickness and health.
For most of us, though, Koller’s influence will be felt most directly in the next generation of Internet search: software systems that “read” Web pages, understand unstructured text and identify new images by observing the relationship of elements within the frame. “The world is noisy and messy,” Koller recently told The New York Times. “You need to deal with the noise and uncertainty.” Google beware.