What's Going On: Lerman's Living Data Lab

August 26, 2016

In late July, Kristina Lerman gave the latest in ISI's "What's Going On" research breakfast series: a dive into social media via her Living Data Lab, which conducts computational analysis of behavioral data. This Friday, August 26, Aram Galstyan will continue the series on forecasting research.

Among Lerman's cheerily presented, if somewhat unsettling, conclusions: Human behavior online is surprisingly predictable and easily nudged in particular directions. Lerman also pointed out that even small changes, like redesigning user interfaces, can noticeably impact how people act online.

The series features research group leaders presenting their work in conversational, TED-talk-like ways that answer the question, "What's going on?" The roughly one-hour sessions aim to introduce ISIers to work taking place outside their immediate groups, foster an Institute-wide sense of community and encourage fresh, cross-disciplinary conversation and collaboration.

Those were certainly the case at Lerman's talk, where lively questions and challenges among the roughly 40 attendees had to be cut short for time reasons. Other speakers to date have been Matthew French on reconfigurable computing and Federico Spedalieri on quantum computing. The series began in May this year and is scheduled to run through March 2017. Still to come: Galsytan, Greg Ver Steeg, Pedro Szekely, Jose Luis Ambite and Craig Knoblock of Intelligent Systems; Jelena Mirkovic and John Heidemann of Internet and Networked Systems; and Andrew Schmidt of Computational Sciences and Technology. Gully Burns is coordinating the series.

Lerman launched her talk with a look at the current battle for civilians' hearts and minds, from the lure of radical groups to the potential for counteracting social manipulation, and the opportunity presented by big data analytics. She went on to cover key concepts in cognitive heuristics, such as position and social influence biases, and the intriguing ability of crowds to outperform experts on certain tasks.

Her case studies included analysis of eight billion comments posted on Facebook, by nearly 30 million people, in a single day. Lerman's team was able to forecast emailing behavior, test performance and viral outbreaks in social media, among other outcomes. Contrary to intuition, the team also discovered that as people's cognitive loads increase, their behavior becomes more predictable, not less.

"What's Going On" talks will be listed in ISI's home page calendar, and ISIers are encouraged to attend.