Roughly five years ago, computer science professor Judy Goldsmith sent an email to the Cognitive Science mailing list wondering if anyone would be interested in working on a grant proposal around the topic of decision making. The level of interest—and where it was coming from—surprised her.
“I expected to get responses from psychology and business, and I did; what I didn’t expect was a large response from biology—people who look at animal decision makers and even plant decision makers at the population level. The evolutionary biologists were very interested. I had no idea. I hadn’t thought about biology, public health or medical decision making,” she recalls.
Thus, the interdisciplinary field of “Comparative Decision Making” was born. In the five years since that email, Goldsmith has contributed to the burgeoning area of research by focusing on computational decision making, which she describes as, “computers making decisions and computers helping people make decisions.”
In addition to theoretical work involving algorithms and computational complexity pertaining to elections and tournaments, Goldsmith is interested in developing decision support systems and decision aids. Currently, she is working on a decision aid for rheumatoid arthritis patients.
“The medications for rheumatoid arthritis can be very scary, and the disease is very scary. But patients sometimes fixate on the small probability side effects and we want to help them think more rationally—to weigh the actual probabilities and not the looming image of horrible side effects,” she explains. “Ultimately, that goes toward two areas of research. One is helping people think more probabilistically and the other is building game-like decision aids, using game technology to help people imagine possible futures.”
One of Goldsmith’s past projects that investigated decision making and planning under uncertainty led her to interview and study case managers responsible for deciding whether someone was able to work or should receive welfare. The project fascinated Goldsmith, who built mathematical models of the process case managers use every day to make decision that affect people’s livelihoods; however, communication—linking her complex mathematical models to the case managers’ daily work—was a challenge. Nonetheless, the project was informative and inspiring.
“I have deep respect for what those case managers do,” she says. “I learned a lot about how incredibly smart and resourceful many of them are.”
Another area of recent focus had to do with planning under uncertainty with regard to academic advising.
“When we give advice to students, there is uncertainty about how well the student will do, so our long-term planning is directed at increasing the probability that a student will graduate with a good GPA and in a reasonable number of years guided, I hope, by the student’s preferences regarding classes, professors and his or her long-term goals.”
One approach to advising is to use a collaborative filtering approach that some have characterized as a “Netflix approach.” Undergraduate student Josiah Hanna, who earlier this year won the inaugural Duncan E. Clarke Memorial Innovation Award, tried this approach a few years ago. More recently, he spent the summer of 2012 in Paris, working with top researchers on multi-criteria optimization approaches to long-term planning under uncertainty, a more mathematical approach. Meanwhile, former students Thomas Dodson and Nicholas Mattei worked on explanation-generating systems that would offer students explanations for the mathematically generated suggestions, such as “Students who had similar grades to you in these courses did well in this course.”
Goldsmith recently wrote commentary for a chapter in the book Comparative Decision Making, edited by Philip H. Crowley and Thomas R. Zentall. The book was published by Oxford University Press in February, confirming that multi-disciplinary interest continues to grow. As a result, Goldsmith expects that new questions and initiatives will make remaining in this nascent field an easy decision.