Emory Report

March 27, 2000

 Volume 52, No. 26

Jaeger pits computers against human brain


We know that computers can play chess, diagnose diseases and recognize faces. Does that make them intelligent? How do computers perform these tasks differently from humans? And could computers eventually have feelings about their role in our lives?

Biology 190M, "Computer Intelligence: Reality and Fiction," is bursting with challenging questions like these. The course offers a fresh and complex approach to the college Freshman Seminar requirement.

"In this class, we are comparing the brain to the computer. They both operate electrically to process information," said Dieter Jaeger, assistant professor of biology. "Neurons are input/output devices like transistors, but much more complex. I'm trying to understand how and when neurons in the brain are activated in relationship to behavior."

To begin seminar participants first discuss what defines "intelligence" since it is not a simple concept­­the Encyclopaedia Britannica definition is five pages long and involves many components. So, what makes an animal or machine intelligent?

"It's not a question easily solved," Jaeger said. "Intelligence is many faceted and may be defined in ways that involve memory, creativity, cognition, problem solving and motivation. There's no hard and fast theory.

"Intelligence is not like the color red; you can't say, 'There it is.' This seminar is not about 'truth.' It's about finding information and discussing a topic that can't be fully resolved."

Students begin their investigation of intelligence by looking at the computer and its components: memory, input/output, programming, speed, sound and image representation. In the process, they discuss fictional computers, like HAL in Stanley Kubrick's 2001: A Space Odyssey, that have displayed human qualities.

"We're using the Internet as an important tool," said Jaeger, whose classroom is newly equipped with a computer display and Internet connection.

Class participants will take their exploration of intelligence from the computer to the human brain. In the second phase of the course, they will work with neuroscience Ph.D. students in preparing and presenting research on topics like neurons and synaptic transmissions, brain structures and their function, mammalian memory and language.

Jaeger's idea for graduate student involvement with undergraduates may prove beneficial for many reasons. Students on both levels will improve their academic skills, and the freshmen will be introduced to postgraduate opportunities as well.

After examining the construction of computers and the human brain, students will attempt to determine the interface between the two. For example, they will compare mammalian memory and computer memory by studying speed and the number of processors in each. "We'll be asking if the machine is fit to support the complexity of information processing that goes on inside the mind," Jaeger said. "Could we program that?"

Ethical implications are among the most difficult issues when it comes to the development of artificial intelligence. "First of all, it depends on what kind of computer you are putting out there," Jaeger said. "If you were shooting for intelligent robots that could move around independently, would they have rights? Would we be responsible for their actions, or would they be accountable? We would have to think about all this."

To present students with current research in the field, Jaeger has invited Paul Hasler, an assistant professor at Georgia Tech, to speak about his work. Hasler is engaged in making computer chips more brain-like by using neuromorphic engineering techniques; in other words, he's developing circuits that emulate neuron processing.

Cutting-edge discoveries aside, Jaeger's greatest interest in this class is to stimulate intellectual discussion among his students. "It's important to evaluate facts for their meaning," he said. "A consensus among college faculty is that we don't want 'absorbing facts' to be the total outcome of an Emory education."

Return to March 27, 2000 contents page