By the early 2000s, it had become clear that markets-places where people have exchanged money and goods for thousands of years-were becoming increasingly complex. Computers had become an integral part of many types of markets, such as those related to website advertising, wireless spectrum auctions, electricity grids, and more. At that time, Caltech's John Ledyard , the Allen and Lenabelle Davis Professor of Economics and Social Sciences, Emeritus, and others recognized the growing need for economists and computer scientists to share information so that they could together come up with solutions for computer-based economic problems.
"We wanted to create a dialogue between economists and computer scientists," says Ledyard, who, along with his colleagues, founded the Social and Information Sciences Lab (SISL) at Caltech for this purpose in the early 2000s. "Caltech was the perfect place to do this because it's small, and we don't have strong boundaries between the departments."
Years later, those efforts have exceeded expectations; SISL, which was recently renamed the Center for Social Information Sciences (CSIS) and is now funded by the Ronald and Maxine Linde Institute of Economic and Management Sciences , is thriving. The center enables an interdisciplinary network of professors, postdoctoral scholars, and students to perform cutting-edge research in the now-flourishing field of computational economics.
"What is unique about CSIS is that these two groups work together seamlessly without walls," says Laura Doval , an assistant professor of economics at Caltech and member of CSIS. Doval is a member of the Division of the Humanities and Social Sciences (HSS) while many of her computer-science colleagues are affiliated with the Division of Engineering and Applied Science (EAS). "The groups have different languages and approaches," she says. "While economists tend to focus on designs carefully tailored to induce correct incentives on market participants, computer scientists are more willing to sacrifice exact incentives in favor of computational accuracy, which is very important in large-scale systems. More and more, these different approaches are coming together."
One of the first projects tackled by SISL researchers in the early 2000s involved the practice of computational advertising. Consider what happens almost every time you load a new website. Within microseconds, a high-speed, automated auction takes place, where advertisers are given the chance to bid on the prospect of posting an ad to your webpage. The advertisers' computer programs have basic information about you and your likes and dislikes, and can decide how much they would be willing to pay to show you an ad.
"This is a huge computational task," says Adam Wierman , a professor of computing and mathematical sciences at Caltech and co-director of CSIS along with Federico Echenique , the Allen and Lenabelle Davis Professor of Economics. "But then it's also an economics problem because you have to figure out how to price the ads.
"Computational advertising was born in startups at Caltech," says Wierman. Indeed, two Caltech startups in this field were later bought by Yahoo and Google.
"What social scientists bring to the table is a deep understanding of incentives and how to get people to do what you want," says Jean-Laurent Rosenthal , the Rea A. and Lela G. Axline Professor of Business Economics and the Ronald and Maxine Linde Leadership Chair of HSS. "Computer scientists, on the other hand, devise algorithms for large-scale market systems. With CSIS, all of this is intertwined."
In other projects, CSIS researchers studied pollution trading rights and the development of methods to preserve online privacy as well as different types of auctions including land auctions and spectrum auctions-where cell phone companies and other businesses that wirelessly relay large volumes of data bid for portions of the radio spectrum.
Today, the work continues in areas such as power grids, kidney donor waiting lists, and cloud computing. Recently, Doval, Wierman, and Echenique completed a study of public school lotteries. As part of the study, they designed a new algorithm to match students to schools in the Pasadena Unified School District. With the algorithm, the study found, families were better matched to their top school choices, helping the district to retain families that might have previously left for private or charter schools.
"This is what we call a dynamic matching problem," says Doval. "Parents might receive public-school assignments but then they don't know if they got into private schools yet, so they hold on to the public-school assignment they received and wait. This takes a public-school seat from somebody else, and both parties could end up leaving the system-the first because they finally received their private-school assignment, and the second because they did not know a seat would be freed up. We designed the system to avoid this problem, to be more dynamic, by allowing the system to constantly update the assignments based on who has taken a seat and who has left."
Another vital component of CSIS is the training of students and postdoctoral scholars, which has benefits for the program as a whole. When postdoctoral scholars, for example, work on projects that bridge social and computer sciences, their faculty advisors are also brought together.
"We have seen more than 20 postdocs go on to become faculty members at top schools," says Wierman. "They have taken this field that didn't exist when the center was founded and carried it to new schools. We are sowing the seeds of a new field that are being spread around the U.S."
Other researchers once affiliated with CSIS are also bringing their knowledge into new arenas. Simon Wilkie, a former senior research associate at Caltech, served as both the senior economist at Microsoft and the chief economist of the Federal Communications Commission (FCC). He is now head of the Monash Business School in Australia. Former faculty member Preston McAfee served as the chief economist at Microsoft until 2018.
"The center is doing exactly what we hoped for in the beginning: each group is better for being a part of the other," says Ledyard. "Some of the greatest breakthroughs at Caltech and other places have come from the spaces between disciplines."