Data mining in the social-​​media ecosystem

Raymond Fu designs algorithms capable of rapidly analyzing photos and videos on Facebook. Photo by Brooks Canaday.

Ray­mond Fu, a newly appointed assis­tant pro­fessor of elec­trical and com­puter engi­neering, wants to build a better social-media ecosystem, one in which Face­book makes expert friend rec­om­men­da­tions and consumer-product suggestions.

“My goal is to bring the cur­rent social-networking system to the next level,” said Fu, a machine-learning expert who will hold joint appoint­ments in both the Col­lege of Engi­neering and the Col­lege of Com­puter and Infor­ma­tion Sci­ence. “Users,” he added, “will def­i­nitely have access to a higher-quality ser­vice in the near future.”

At the begin­ning of the year, Fu received funding from both a sci­en­tific research orga­ni­za­tion and a coali­tion of gov­ern­ment agen­cies to design algo­rithms capable of rapidly ana­lyzing con­tent on social-networking and video-sharing web­sites. In Jan­uary, for instance, he received a two-year, $360,000 post­doc­toral research fel­low­ship award from the U.S. Intel­li­gence Com­mu­nity, and in Feb­ruary he received a three-year, $483,500 grant from the Air Force Office of Sci­en­tific Research.

One algo­rithm in progress, Fu explained, will be capable of pre­dicting the demo­graphics, behav­ioral ten­den­cies and rela­tion­ships between people who appear in photos and videos on Face­book or YouTube. The other algo­rithm will be capable of pin­pointing their geo­graphic location.

“Social-media users are sharing a lot of infor­ma­tion every single day,” Fu explained. “Our goal is to develop an auto­matic method that could process this data very fast.”

Face­book adver­tisers will have the ability to pur­chase the data for the pur­poses of tar­geting ads to par­tic­ular users. But, Fu explained, “Users would have to allow this con­tent to be deliv­ered to them.” Issues of pri­vacy, he added, “would have to be resolved.”

Both social-media projects dove­tail with the university’s focus on con­ducting use-inspired research that solves global chal­lenges in cyber­se­cu­rity. As Fu put it, “This research con­siders secu­rity issues such as data com­mu­ni­ca­tion, sharing and pri­vacy in cyber­phys­ical networks.”

Prior to joining the North­eastern fac­ulty, Fu served as an assis­tant pro­fessor of com­puter sci­ence and engi­neering at State Uni­ver­sity of New York at Buf­falo and was the founding director of its Syn­er­getic Media Learning Lab.

Fu hopes to set an ambi­tious research agenda in his first year at North­eastern. He looks for­ward to col­lab­o­rating on inter­dis­ci­pli­nary research projects with sev­eral fac­ulty mem­bers, including Deniz Erdogmus, an elec­trical and com­puter engi­neering pro­fessor, and Marty Vona, an assis­tant pro­fessor of com­puter and infor­ma­tion science.

“North­eastern has a very strong engi­neering pro­gram,” Fu said, adding that five grad­uate stu­dents and three post­doc­toral researchers under his watch at SUNY Buf­falo have trans­ferred to the uni­ver­sity. “There are many col­leagues who share sim­ilar research inter­ests who I look for­ward to col­lab­o­rating with.”