YOU ARE HERE: LAT HomeCollections

The Next Logical Step in Dressing


Call it Compu-Coif. The Digital Dresser, perhaps? Microchip Clothier? Or how about just . . . Soft-wear?

Jean Chen, a 26-year-old USC graduate student in engineering, uses a computer program to get dressed. It chooses everything--skirts, blouses, accessories--each choice geared to the weather, season, event, even her sartorial whimsy.

When you stop to think about it, Chen's program, developed as a project for a class in Neural Fuzzy Logic, offers the frightening potential of relieving humanity from the responsibility of the casual decision.

Imagine. Computer programs to solve such trifling questions as: Where do you wanna eat? What movie do you wanna see? Should I punch out my therapist?

"I guess you could design a program for any kind of decision," said Chen, who apparently has not considered the horrific ramifications of her invention. "When I was designing it, I was just trying to make something for me to use."

Of course, programs of this ilk could have been devised years ago. What is new--indeed, revolutionary--about Chen's "Perfect Fashion Design" is her use of so-called "fuzzy logic," a relatively new concept that allows the computer to deal in probabilities rather than absolutes; gray instead of black and white. Just like humans (with the possible exception of Oliver North) do.

In other words, in lieu of a plain "yes or no" offered by a non-fuzzy program, fuzzy programs might advise a user to lean toward Chinese food for dinner, or to probably steer clear of that stupid "Gump" movie, or merely irritate one's therapist.

More to the point, "Perfect Fashion Design's" database contains information on Chen's entire expansive wardrobe. Rather than sitting stricken with indecision in front of her crowded Hacienda Heights closet--as she often did--Chen now can feed the following variables into her program: season, weather, type of occasion and degree of formality.

Without fuzzy logic, her computer would be restricted to simple suggestions such as, in the question of weather: "hot," "cold," "cool," "warm." With fuzzy logic, Chen can slide her cursor anywhere between cool and cold, or between warm and warmer--lukewarm, you could say.

On the issue of formality, she might slide the cursor barely toward formal, or just over the line toward casual, or tentatively in the direction of . . . flashy. Such possibilities as quasi-formal are available, or somewhat casual, or just a tad flamboyant, or a wee-bit chilly, or summery, or "casual day" at work, or rather funereal, or knock 'em dead.

And . . . voila! The computer "spits out," to use Chen's picturesque term, five choices of ensemble--in order of recommendation, and in full color. In the rare instance in which the master's candidate doesn't like any of the options, she simply orders the computer to cough up five more. Yes, she still bears the burden of final choice--but, as she will tell you, it beats the hell out of trying on a dozen outfits before going to work an hour late.

And lest you think that something as technically wondrous as fuzzy logic is demeaned by such mundane application, consider this: In Japan, fuzzy circuits control subways, elevators, bread-makers, rice cookers, washing machines and various other everyday appliances.

"My program is for everyone who doesn't really have time to look into their wardrobe every day to figure out what to wear," said Chen. "You can make it adapt to your needs. If you program the same request two days in a row, it will not repeat the preceding top five choices. If you want different clothes every day, you can do that. Or you can train the computer to learn your taste."

That might be a dubious achievement in the case of Chen's fiance, Robert Lee, a self-described "nerdy engineer" who laughingly summed up his fashion sensibilities with the phrase "whatever's on the top of the pile." Lee is nonetheless helping Chen improve the program for possible marketing. (They're looking for investors.)

"We're in the process now," said Lee, "of putting adaptive strategy into the software, so what the program will do is try to derive your tastes in clothing. In other words, after you use the program for awhile, the software is smart enough to know your color taste, etc., just based on experience."

Chen brainstormed "Perfect Fashion Design," or PFD, for a class taught by USC professor Dr. Bart Kosco, who for the past six years has held annual contests in his popular Neural Fuzzy Logic course. Past winners have come up with fuzzy orange sorters, a fuzzy pet door that opens only for recognizable meows, a star-movement tracker to facilitate automatic photography of the night sky, and grander concepts (including a fuzzy sewer system). Chen was actually runner-up to a student who designed a fuzzy pacemaker.

Los Angeles Times Articles