Designing and Analyzing Societal Networks

TANZEEM CHOUDHURY
Cornell University

SCOTT KLEMMER
University of California, San Diego

Computing is increasingly woven into the fabric of everyday life. Many remarkable societal changes emerge as social technologies are adopted at massive scale—in social networks, smart mobile devices, digital health tools, online education. Socially and physically embedded computing yields new possibilities for increased sensing and data mining and personalized information. The speakers in this session focus on the opportunities and challenges of this quickly growing scale.

One dramatic change over the past decade is the number of people who carry Internet-connected sensing devices. Smart mobile devices make it possible to monitor health, capture rich media, and access vast information repositories at the touch of a button. Perhaps more than any other technology, these devices are ushering in new techniques for massive data science by correlating sensing and behavior—and also raising concerns about privacy in a “transparent” society.

Another important change is massive, socially connected online media. Social networks open up new avenues for communication and civic engagement. Online health platforms enable people to share health information. Online education has enrolled millions of students in just the past year and is causing many universities to rethink their long-term strategy.

All of us—as citizens, people, and educators—are affected by these changes. How should citizens, families, and universities think about massive online social interaction? How does it change the services we provide, the science we conduct, even our very conversations? This panel explored these issues.

Tony Jebara (Columbia University) led off the session with a presentation on modeling large-scale networks based on mobility data. Although most network growth models are based on incremental link analysis, he explored how to draw on users’ data profiles alone—without any connectivity information—to infer their

connectivity with others. For example, in a class of incoming college freshmen with no known friendship connections, is it possible to predict which pairs will become friends at the end of the year using only information such as their dorm or relationship status? Similarly, based only on the location history of a population of mobile phone users, can an observer predict which pairs of users are likely to communicate with each other?

Rob Miller (Massachusetts Institute of Technology) followed with remarks on the use of crowd computing to harness the power of people for tasks that are hard for individual users or computers to do alone.¹ He described prototype crowd-computing systems that he and his colleagues have built: a Word plugin that crowdsources text editing tasks, an app that helps blind people see using a crowd’s eyes, and a system for code reviewing by a crowd of programmers. Crowd computing raises new challenges at the intersection of computer systems and human-computer interaction, to improve quality of work, minimize latency, and provide the right incentives to the crowd.

In the third talk, Kate Starbird (University of Washington) examined the crowdsourcing phenomenon during natural disasters and other crisis events. Armed with mobile devices and connected through social media platforms, people at the site of a disaster event are newly enabled to share information about unfolding events. This real-time information could be a vital resource for affected people and responders, but it remains difficult to transmit the right information to the right person at the right time. She described various ways that the crowd works to process data during disaster events and suggested future directions for leveraging “crowd work” to improve response efforts.

In the session’s final presentation, Duncan Watts (Microsoft) reviewed exciting progress and challenges in the new field of “computational social science.” He cited three obstacles to the widespread use of this resource at the convergence of the social sciences and the computer sciences. First, social scientific problems are almost always more difficult than they seem. Second, the data required to address many problems of interest to social scientists remain hard to assemble. And third, thorough exploration of complex social problems often requires the complementary application of diverse research traditions. He described some ideas for addressing these challenges.

___________________

¹ Dr. Miller’s presentation is not included in this volume.