Supreme Court Justice Potter Stewart once famously observed when commenting about pornography, “I’ll know it when I see it.” Unwittingly, Stewart was elucidating a problem that has plagued computer scientists for years. Computers, for all of their incredible power and capabilities, are nowhere as good as human at discerning faces, judging beauty, and, if need be, telling someone what may or may not be pornographic.
Now researchers are marrying the power of computers to the pattern-recognizing skills of humans in order to design better products. The simplest way to think of this is to consider a nice photo of yourself but one that is perhaps slightly out of focus and suffers from poor lighting. Photoshop and a number of other software packages provide the tools to improve the photo, but they require that the user be familiar with hues, gamma corrections, and other assorted tools. All the average person knows is that she wants the photo to look better.
With “hunch” software, a series of mutated photos would appear on the screen with different adjustments in focus and lightening, and the user can then select the photo that appears best. The process continues until the photo is refined to the person’s preference.
The beauty of this approach is that the users don’t have to know anything about photo software, nor do they even have to describe what precisely they are looking for. Like Justice Stewart, they will just know the best photo when they see it.
This technology is still in its infancy, but such software is already being used by researchers at pharmaceutical companies, who are using their “hunches” to select certain molecules that might make promising drug candidates. (The computer then recombines the molecule into a series of other mutated molecules in the hopes that it will bring them one step closer to an effective drug.) Home decorating stores are also using the technology to assist customers select color patterns and room layouts that are more to their liking by allowing them to continually refine the colors until they find the ones most suitable to their tastes.
The technology, however, goes well beyond these limited applications. For example, few people can describe precisely what they want in a car, a Web site, or clothing, but they will know what they like when they see it. To this end, Amazon.com now offers a tool to help businesses do this. It is called the Mechanical Turk, and the company calls it “artificial artificial intelligence,” because it uses human intelligence to improve computer intelligence.
Exponential Insight
If you have a hunch that you should be employing more technology but are not yet convinced that machines do a lot of things as well as a person, you’re correct. But new tools do now exist that can help increase the performance of both man and machine.
Related Posts
Jump the Curve Strategy #8: Run the Numbers
Jump the Curve Strategy #7: Reorganize Your Data Storage Closet
Jump the Curve Strategy #6: Let the Computer Do It
Jump the Curve Strategy #5: Bet On It
Jump the Curve Strategy #4: Just “Wiki” It
Jump the Curve Strategy #3: Look to the Kindness of Strangers
Jump the Curve Strategy #2: Take a Bird’s Eye View of the World
Jump the Curve Strategy #1: Learn to Spell Zenzizenzizenzic
Introduction to Chapter Two: The Power of Zenzizenzizenzic
Welcome to the Exponential Economy: Prepare to Jump the Curve
Introduction to Jump the Curve
