Berkeleyan

Avideh Zakhor, UC Berkeley professor of electrical engineering, holds forth at Nefeli Café. (BAP photo)

A mind is a wonderful thing to feed
Electrical engineering's Professor Avideh Zakhor, over lunch, talks about what it's like being a science nerd in Iran, why Larry Summers made her mad, and just what the heck a 4-D model is anyway

By Bonnie Azab Powell, Public Affairs | 27 April 2005

In this new interview series, we'll lunch with not-yet-famous professors, learning about their fields and the paths that led them there. With luck - and the right sandwich - we hope to provide a glimpse into what makes these Berkeley minds tick. E-mail your suggestions for likely candidates to bpowell@berkeley.edu

Avideh Zakhor has been interested in engineering since she was a little girl growing up in Iran. "My father started the first button-manufacturing factory in Tehran," she says in between quick bites of her Athenian sandwich. Since the Zakhors lived in the same building as the factory, "I was always downstairs watching how the workers made the buttons: they'd pull the resin in - the raw material was a byproduct of the Iranian petroleum industry - and a machine would cut it into discs. Another one would carve it, another put the holes in it, and then the coloring process. It fascinated me."

We're sitting outside at Nefeli's Café on Euclid Avenue, around the corner from Zakhor's office in Cory Hall, where she is a professor of electrical engineering. Zakhor eats at the Greek café so often that she has an account. It's easy to see why. My Alpino sandwich is a tasty mix of savory and sweet: a crunchy french roll filled with prosciutto, fontina cheese, Anjou pears, roasted walnuts, and arugula. Zakhor's sandwich has roasted eggplant and red peppers, feta, onions, kalamata olive spread, and parsley.

Contrary to stereotypes about women in the Middle East, Zakhor felt no stigma as a mechanically minded Persian girl. "Not many people know this, but half of all college graduates in Iran are female," she says. "It wasn't always like that, but it's not that unusual." At 15, she won a prestigious scholarship from the Iranian government to study science and math at an international boarding school in Wales. Shortly after she got there, the Shah of Iran and his family were forced into exile, and Zakhor's family, like many of their countrymen, fled Iran and ended up in Los Angeles.

"A week after the government fell, the check for two years at school arrived," she laughs. Subsequent scholarship checks were also supposed to cover her college and graduate school costs, too; but those, of course, didn't materialize. After finishing high school in Wales, Zakhor chose the California Institute of Technology, in order to be closer to her family. The male-to-female ratio in 1980, only eight years after Caltech began admitting women, was around 7 to 1.

"The joke was that you could have a different boyfriend for each day of the week," she recalls. Not that she had time. After just three years, Zakhor graduated first in Caltech's engineering class of 1983. She then proceeded to speed through her Ph.D. in electrical engineering and computer science at the Massachusetts Institute of Technology in just four years.

"What was the rush?" I ask.

"I was so cold! I wanted to finish quick so I could come back to California," she says, shivering at the memory of chilly Boston. "I used to go to my lab at 2 in the afternoon and stay there until 6 in the morning because it was one gigantic winter, always cloudy and raining."

Tenure clock, meet biological clock

'If the men are working 80 hours a week, you work 85 hours. You have to continually prove that you're better.' -Avideh Zakhor

By 24, Zakhor was an assistant professor of electrical engineering at Berkeley. She got tenure six years later, in 1994, a few months before her son was born; she made full professor in 1998. (Zakhor is married to fellow Berkeley professor of electrical engineering Seth Sanders, who she says followed her west from MIT hoping for a date.)

Her accelerated schedule made it easier to balance the pressures of an academic career with having a family, but Zakhor knows she's an exception. "I think the cards are somewhat stacked against you as a woman in academia. There's the tenure clock and the biological clock, and it's hard for the males who run departments to understand the latter," she says. "Most people get out of graduate school at 28 or 30 and won't get tenure until they're 35. The university supposedly stops the tenure clock when you're having children, but I know personally that having children earlier would have severely hampered the progress of my tenure case. So most women have to postpone childbirth until they're 35 or 36, which just complicates their life unnecessarily."

Zakhor is particularly irritated when asked about the recent firestorm surrounding comments by Harvard President Larry Summers, in which he suggested that the reason so few women become scientists or engineers might be due to an innate difference from men in their abilities. "Women have a hard enough time competing in academia, not just in science and engineering, and comments like this just compound the problem," she says. "Deep down, I wonder how many people think the same thing as Summers but never say it."

She taps her water glass, thinking. "I know a lot of women who deliberately don't choose academic careers in science and engineering, because they see that what it takes to succeed conflicts with their personal goals of raising children and having a family," she continues. "The ones who choose it anyway sacrifice even more than men in my opinion, because they have to fight against this inherent bias that's ingrained in the system. If the men are working 80 hours a week, you work 85 hours. You have to continually prove that you're better."

Life begins at 4-D

While Zakhor should have nothing left to prove, she's still in constant motion. She heads the Video and Image Processing Lab (www-video.eecs.berkeley.edu) at Berkeley as well as Visualization MURI (www-video.eecs.berkeley.edu/vismuri/murindex.html), a Multidisciplinary University Research Initiative in which researchers from five universities are developing four-dimensional dynamic models of the environment. On sabbatical from teaching this semester to finish a book about multimedia networking, she's also involved with TruVideo (www.truvideo.com), the second company she's co-founded, which makes it possible for Verizon, Sprint, and AllTel subscribers to watch video content on their phones.

A portion of MURI's 3D Berkeley model: With the right software, viewers can look at this street from multiple angles, even seeing the other side of the theater marquee.

As dissimilar as these projects sound, they all have something in common. "They're all about signal processing - compact ways of representing the world," she explains. A signal can be two-dimensional, like a digital image; processing it could entail compressing that signal into an MP3 or a JPG and figuring out how to put it back together. Add the element of depth, and you get a 3-D image of an object that can be rotated and examined from all angles.

So far, so good. "But I'm afraid I can't wrap my brain around the concept of four-dimensional," I say.

"Four dimensions is just a three-dimensional object moving in time," she answers, speaking slowly. "Imagine recording the performance of a ballet dancer, so that when you watch it at a distant location you can view it from any angle and from any point in time."

"Like that shot in The Matrix where the motion freezes and the cameras swing around Keanu Reeves?" I demonstrate by rotating my napkin around my fork.

"Basically yes, except to get that shot they used lots and lots of cameras, which is extremely expensive and inefficient," Zakhor says. "We want to find a cheap way of generating that same amount of information." To that end, she and her postdoc, Christian Früh, have created a 3-D model of downtown Berkeley, using a laser scanner mounted in the back of a truck. The scanner sent out infrared beams at five-centimeter intervals as they drove down the street, and a computer crunched the data as it came in, assembling an extremely detailed map of the topography of the building facades. When combined with a similar map that a group they hired made from a plane, the result is a virtual Berkeley - viewers can pan, zoom, and turn the corners.

But don't confuse this with "virtual reality," computer-simulated renderings. "I prefer to think of it as 'virtualizing' reality," says Zakhor.

As one might guess, the most immediate interest in a 4-D city model is coming from the U.S. military. The Visualization MURI project indeed has been funded by the Army Research Office; the Defense Advanced Research Projects Agency (or DARPA, the federal agency that funded the creation of the Internet) will likely step in when the ARO grant ends in July. The project is not classified, and Zakhor doesn't see the application as limited by its potential military application. She's more excited by applications such as telesurgery, where doctors in San Francisco could use 4-D video and a joystick to operate on patients in Alaska, say. (This is already happening using standard teleconferencing video.)

Mulling over what she's said, something occurs to me: "What do you watch 4-D video with? A hologram projector like they used in Star Wars?"

"It's true, 4-D viewers aren't really here yet," acknowledges Zakhor. "But they will be. Maybe even someday on your cell phone. Last year the number of cell phones exceeded the number of PCs - 1.5 billion to 1.3 billion. Who wants to carry around their PC? Your cell will be your iPod, your DVD player, your camera, your e-mail device, all of your entertainment and communication needs in one gadget."

And since someone will have to figure out how to process and transmit all those signals efficiently, chances are good that Zakhor won't be slowing her pace anytime soon. On our way out, Zakhor orders a double espresso - to go, of course.