Ruzena Bajcsy

Ruzena Bajcsy, professor of computer and information science and director of the General Robotics and Active Sensory Perception Laboratory, had to be persuaded to pose for pictures for this story.

Having been born in Czechoslovakia under the Nazis and educated under the Communists, Bajcsy has developed a strong aversion to anything that even remotely smacks of glorifying an individual, including herself. She takes great pains to point out that the GRASP Lab's reputation as one of the country's top robotics institutes is the result of first-rate faculty and students collaborating on advanced research.

And yet this work, her administrative skills, and her professional honors -- she is a member of both the National Academy of Engineering (the seventh Penn faculty member so honored) and the Institute of Medicine -- are some of what led the NSF to select her as the new head of its Directorate of Computer and Information Science and Engineering, effective next month.

As director, Bajcsy will serve as an advocate for information-technology research in Washington while maintaining her Penn ties - she plans to return to campus regularly to do research and keep her door in D.C. open to the Penn community.

Her convictions aside, Bajcsy also felt strongly enough about the importance of computing to talk with the Current about research priorities in the field and how they will be funded.

Q. What do you think are some of the challenges that lie ahead for computing?
The challenge is to get money for the field - to convince the Congress, the public, the society that there is important, outstanding research that it needs to support. The next challenge is after we get the money, to distribute it wisely so that good things will happen.

Q. How much money right now does the federal government spend on research in computer science?
Computer science spends about $600 million a year, out of which NSF [funds] half. And we want to increase that by $200 million per year.

Q. What sort of research would be funded with this additional money?
There is a plan, and that plan is being conveyed by a presidential committee on information technologies that made some priorities ... So there are four priorities. The [first is] software development, and that is very important research, because we really don't understand these large-scale software systems.

Q. Such as?
The big banking, insurance, the big payrolls ... as the hardware is changing, how do you make sure that you don't have to write new software and operating systems, and when you add new computers, how do you make sure the software is consistent? We don't have good mathematical models for these kinds of parallel, distributed, adaptively growing software systems.
   The second thing is the scalable information infrastructure. That relates to the issues of software and communication technologies as you go from PCs to supercomputers.

Q. To enable computers of various sizes...
A. communicate, access, run software on these various different[-scale] machines.
   The third thing is what we call high-end computing, which is really supercomputing, the large-scale computations like weather forecasting, simulation of biochemical reactions, chemical reactions, nuclear reactions and astronomical modeling.
   Finally there is development of people, namely, the workforce, educating people in this technology. So those are the four priorities.

Q. On the engineering side of this post, what are some of the research or funding priorities?
Computer science and engineering is a very special field of inquiry, because in natural sciences, where you have nature to probe...the science of it is understanding nature, discovering the laws of nature.
   In computer science, the equivalent would be laws of information - how do we represent this information and discover laws of manipulating [it]?
   But the medium where this information resides is this hardware called a computer. So you have this situation where the information that the core of the science is about is not physical nature, but the reflection, the representation of this physical nature. But in addition to that, this computer can also represent abstract concepts.
   So the engineering aspect is how you design this medium where this information resides and you are manipulating this information. But the engineering and the science are very closely intertwined, because the architectures that the computer engineers will design will also determine how you can represent this information and what kind of manipulation you can do on it.

Q. How far have we come in understanding how to manipulate information across different architectures?
Not very far. That's what this push of the presidential committee is, to put some more effort to understanding the scalable software, the scalable algorithms. And that requires a much deeper understanding of representation - how to represent the information in different layers, and then how to do the manipulation.

Q. Would a typical computer user about 10 years from now notice anything different as a result of this research?
The goal is that the user should not know anything about what's happening inside. Computer usage should become easy as today's telephones. You push buttons, and you pick up the phone, and you shouldn't worry about how it is done that you just dial a few more digits and you talk to Australia or Japan or next door.

Q. Would you care to speculate how long until we reach that state?
I do not care to speculate, because the experience we have is that as we solve some problems, new problems occur, and the pervasive effect of this technology is you solve one layer and 10 layers show up that you need to solve. It's very hard to say what will happen.
   I mean, look. When automobiles came about, you had to be almost a mechanic to be able to drive. Today, you pass the driver's license which is relatively trivial, and so almost anybody can drive an automobile, right? Similarly with the telephone.
   Now, the computing technologies and the services that computers provide today are much more complicated than the car or the telephone, and yet it's spreading much faster than anything anybody anticipated. And we haven't scratched the potential.

Q. Do you have any sense where Congress stands with regard to funding scientific research in general?
I believe that the current Congress is quite favorably inclined toward funding information technology.

Q. Even with their talk about balancing the budget and cutting taxes?
Absolutely. I think the Republicans are quite sympathetic. NSF got a very reasonable increase - between 7 and 8 percent, I think; I've forgotten the number exactly. [It was 7.1 percent.] And, of course, the favorite child is the NIH, which got a substantial increase, 20 percent or so. But NSF is next. And I think it's a recognition that information technology has contributed to the current health of the United States and our competitiveness and jobs.
   Of course, that's the number-one reason. The number-two reason is that it really is acting on every aspect of our daily lives. We need to push it ahead in order to lead.

Front page for this issue | Pennsylvania Current home page

Ruzena Bajcsy