David Waltz

waltz@research.nj.nec.com
Vice President, Computer Science Research, NEC Research Institute
~What is your particular speciality in Artificial Intelligence?
I have worked on computer vision (my Ph.D. thesis), natural language processing, connectionist/neural net models, and "memory-based reasoning," a method of classification that uses nearest neighbors in a database to aid in making decisions.

~Why did you choose a career in AI research / development?
I was excited by a professor while in graduate school (at MIT). I had thought that I wanted to work on radio astronomy or control theory before discovering AI.

~What is the AI system you are researching/developing designed for?
The last systems I worked on were used for 1) automatically assigning keywords to news articles, 2) classifying people in the US Census to occupation and industry categories, based on their brief answers to some questions, and 3) predicting the 3-D structure of protein sequences.

~What approaches are you using in your research / development ?
Most recently MBR-- memory-based reasoning -- as described above. I (with a paost-doctoral employee) am also using statistics to build lexicons for natural language processing out of text databases and on-line dictionaries.


~What do you see as some fundemental ways that AI in general will impact
people's lives in the future?
Systems will be smarter -- or perhaps just less stupid. Many Web applications will use AI to tailor system behavior to match your patterns and tastes; houses, cars, appliances, etc. will be smarter, saving energy, adapting their behavior to your needs and the current situation; automatic accident avoidance for cars will be followed by self-driving cars; household robots are possible in 15 years, likely in 30; education will become much more geared toward teaching students to find and use Web resources, and less toward memorizing anything. Work as we now know it may become unnecessary, and the overall productivity and wealth of societies can become vastly greater.

~What Do you think AI tecnology will be like in 10 years? in 20? In 50?
See comment on robots and cars above. I think that technology will move toward processors and memories on the same chip, leading to intelligent memory. An intelligent memory could search/compare each action you do with all the items it storesMatched items can be used to suggest shortcuts, remind you of things you've done or need to do, etc. Computers will be much more proactive, though they can become unobtrusuve if requested. People will have continuous portable Web access, and will depend heavily on it for work, entertainment, communication, education, etc.

~Do you think Computers will ever be able to think and talk like humans?
"Ever" is a long time... Talking can already be pretty convincingly simulated. Really thinking is much more challenging. People think as they (we) do because of innate genetic endowment, shaped in development by contact with culture. But unless a computer were constructed with desires and innate properties like those of people, it probably wouldn't think exactly like us. Most likely they will be superior in some ways -- they could have very large short-term memory capacities, while humans can remember only 7 + or - 2 items at a time -- but not as good at other things. Except to show that we really understand the organization of the brain, and to show that modeling the brain is sufficient to build a smart system, it's hard to see why we'd want to build a system exactly like a person in every way. But if you want to know whether I think it would possible in principle, the answer is yes. But even at the fastest, this probably won't happen in our lifetimes (unless we figure out ways to dramatically extent human lifetimes).

~What is most exciting part of AI that encourages you stay in the field?
Understanding intelligence is one of the greatest -- perhaps the greatest -- scientific challenge. It's fascinating, not only because of the beauty, subtlety, and complexity of the brain/mind, but also because there may be consequences (for education, therapy, philosphy, etc.) because of what we find. And the technology has the potential to change the productivity of all societies, the nature of work, etc.

~What subjects would you encourage high-school level students to take, who
are interested in AI?

Math, first and foremost. Biology and physics. Computer science if it's offered. And you should read widely -- inspiration for solutions and for understanding the nature and properties of human intelligence and society require as wide a background as possible. Read sci-fi, biographies, history, science, great novels.

David Waltz

David L. Waltz Vice President, Computer Science Research NEC Research Institute, Inc.
4 Independence Way Princeton, NJ 08540 waltz@research.nj.nec.com http://www.neci.nj.nec.com/homepages/waltz.html 609-951-2700 fax: 609-951-2483
[ HOME ]
[ Basics | History | Approaches | Applications | Interactive | People | Links | Messages | Programs]