Mathematics has the potential to provide inroads to almost any scientific discipline. No one understands that better than Ben Goertzel, who is currently chief scientific officer of BioMind LLC in Bethesda, Maryland, which applies artificial intelligence to the analysis of microarrays. In college and graduate school, his interests ranged from theoretical physics to psychology. "I couldn't make up my mind what I wanted to do, but I found math to be very easy, so I studied it," he recalls.
In graduate school at Temple University, he began with a theoretical physics project, but it bogged down and he switched to numerical analysis. "I gravitated toward the principle of minimum effort, so I chose a thesis topic that would [progress] much faster." It was a fortuitous decision. Numerical analysis turned out to be in demand in an otherwise dismal job market, and in 1989 Goertzel took a post in the mathematics department of the University of Nevada, Las Vegas (UNLV), which had been looking for someone with a specialty in numerical analysis.
Once there, "I immediately stopped doing work in numerical analysis," he says.
Mathematics: The ultimate high and holy pursuit
Instead, he turned his attention to theoretical psychology and mathematical theory of the mind, with a goal of understanding how the human mind works and developing a theory of artificial intelligence. He published a number of books and journal articles on those topics, but eventually grew bored with the academic environment. He was interested in applied mathematics, and the academic environment didn't seem a good fit for it. "By and large [academic mathematicians] believe that the pursuit of mathematics is the ultimate high and holy pursuit. Even physicists are a little inferior. Everyone else is profoundly inferior."
In 1994, he spent a year in New Zealand as a lecturer in computer science and then spent 3 years as a fellow at the University of Western Australia, researching cognitive science. When he arrived in New York to take a position at the College of Staten Island in 1997, he remained interested in pursuing artificial intelligence (AI), and he realized that it would be easier to assemble the resources in industry than in academia. And it was the late 1990s. "At that time people were raising money selling underwear online or any ridiculous thing. We thought we ought to be able to get money to create a thinking machine," Goertzel recalls.
It turned out to be harder than anticipated, but 18 months of work finally netted the start-up funding they needed, and WebMind opened its doors in 1997. The company suffered the same fate as most of its contemporaries: Goertzel and the rest of the management team were inexperienced and the company had little focus or revenue, and it went out of business in 2001. Goertzel soon decided that biology would be his next foray. While at UNLV, he had written a book about the parallels between biological evolution and the growth and change of ideas in the human mind ( The Evolving Mind ), and he continued to follow the biological literature with interest.
He read about the massive amount of data being generated--gene expression, protein-protein interaction, and more recently RNA interference--"which allow you to get a picture of the dynamics of what happens inside cells. I thought this was an outstanding application for applied mathematics and AI ideas. I decided to start a company applying that technique to the analysis of biological data, beginning with microarray data," Goertzel says. The transition required some boning up. "I could read research articles, but I didn't understand to the point where I could write them."
With a far more practical focus than its predecessor, BioMind is generating revenues, selling AI software and services to clients such as the Centers for Disease Control and Prevention (CDC), the University of Virginia Medical School, and Northrop Grumman. For CDC, the company's software is analyzing microarray data obtained from chronic fatigue syndrome patients in an attempt to identify gene expression profiles capable of diagnosing CFS.
Business, management, artificial intelligence, and transhumanism
Goertzel currently spends most of his time combining business development and management responsibilities. He is also the company's chief executive officer, but he would like to replace himself so that he can focus more on technical problems.
He still practices mathematics, "because even though you can simulate everything, you have to do a lot of pencil-and-paper calculations to give you an idea of how your algorithms will work. You couldn't publish it in a mathematics journal, but you couldn't do it if you didn't have rigorous mathematical experience," he says.
For all BioMind's improved business focus, Goertzel remains a generalist at heart. The experience with WebMind taught him the importance of a sound business plan, but he remains interested in pure AI and even transhumanism--a hypothetical future when the human mind evolves and intermingles with a machine mind. Some of WebMind's revenues get funneled into that research, as well as antiaging projects focusing on the dynamics of apoptosis.
Having one foot in the practical present and one in the inscrutable future seems to suit Goertzel fine, and mathematics has been key to performing that balancing act. "The nice thing about it is it lets you be at the center of what's going on. Mathematics is just as important to biology now as it was to the quantum physics era of the 1920s."
Read the companion article The Mathematical Biology Job Market, also part of this Next Wave feature.