Professor Evan Peck, computer science, hopes that computers will someday know exactly how and when to interact with humans when information or assistance is needed.
"Humans collaborate and connect with each other on projects. The idea is to replicate that type of partnership between humans and computers, but there is a disconnect with the computer part," says Peck, whose research in human-computer interaction involves personalized interfaces, brain-computer interfaces and information visualization. "When people trade information, there are social cues such as tone of voice or posture that tell us when conversations are going well or poorly. Computers don't know when to ideally present information, and we are often interrupted at inappropriate times."
In his doctoral research, Peck and a team of computer scientists and biomedical engineers used a brain-scanning headband that measured the mental exertion of the wearer. Sensors detected information overload or boredom based on brain activity, and the computer adjusted tasks to achieve a balance between the two extremes.
Machines can take advantage of psychological studies that track body and brain signals, says Peck, who hopes to translate those signals into a code the computer will understand. "I'm interested in how changes in heart rate or eye movements indicate when a computer causes a person stress. If people are interrupted in ways that are wrong, then their anxiety increases and performance goes down." He notes that studies have shown that when computer programmers are interrupted, they don't resume working for another 10 to 15 minutes – a significant loss of work time.
If a computer "knows" when or how information should be delivered, Peck says that machines might someday become true collaborators in solving problems, or brainstorming partners in creative applications. That would be particularly useful if the end user was an air traffic controller faced with an emergency, for example. "There are a million ways for a computer to respond that don't match our experience, emotions or working environment. I am trying to deeply personalize our engagement with computers by understanding how they can respond in ways that make the most sense to us," he says.
Posted Sept. 29, 2014