Prof. Jon Kleinberg ’93, chair of the department of information science, described his research on how information spreads through the internet at the spring Phi Beta Kappa Invitational Lecture on Wednesday evening.
The internet experienced a significant shift around 2004, according to Kleinberg. In its early years, he said the internet was similar to a library — functioning primarily as a storehouse of information — but with the advent of social media, the internet became able connected people easily and directly.
“If in 1993, the way I interacted with you on the internet was that I created a page and my page linked to your page, now suddenly we were interacting much more directly,” Kleinberg said.
Kleinberg focused his lecture on two projects he conducted — one analyzing common attributes of viral content and the other studying structural elements of online social networks.
In the first study, Kleinberg described his research on the news cycle during a presidential election campaign. He said the first step was to determine what information they could feed to an algorithm that a computer would understand.
“Computers are very good at tracking discrete objects, like photos … things that have a relatively stable identity,” he explained.
After reasoning that quotes — even when passing through different sources — tend to remain relatively unchanged, Kleinberg and his students looked into which quotations dominated the news cycle. He said that they attempted to determine whether a quotation’s popularity was a function of context or if it was due to the content of the quotation itself.
“Why were these the particular quotes that were succeeding?” Kleinberg questioned. “Was it sort of accidental or was there something special about these particular quotes?”
Kleinberg compared ethologist and evolutionary biologist, Richard Dawkins’ mechanisms of biological evolution and the propagation of textual memes.
“Bits of culture or bits of text behave like genes,” Kleinberg said, referencing Dawkins. “Some sort of die out and some — through selective pressure, through mutation — evolve and become part of the culture.”
Kleinberg then panned out from discussing the nature of the content to discussing the networks through which this content flows. Again, Kleinberg emphasized the significance of social media’s contributions to the analysis.
“The fact that [social media websites] have assembled structures like this is both a reflection of the richness of our offline social lives, but is also something, I would argue, genuinely new,” Kleinberg said. “These systems are acting as sort of prosthetics … allowing people to sustain ties that would otherwise fall apart.”
These studies were examples of “how we can take sources of data in our everyday life … and try to find computational problems there,” Kleinberg said, adding that there is immense promise to use this mundane information to improve the people’s social, intellectual and professional lives.
However, he cautioned that as researchers mine the rich data sources now available online, they must also assess the costs of stockpiling personal data.
“As we think about the implications for our own lives, for academic scholarship, for the new kinds of systems that we build, for law and policy … I think there’s really an opportunity for computing to integrate with many different disciplines around an institution like Cornell to really inform the next steps,” he said.