Ribeiro, assistant professor in the
Department of Computer Science.
“This same failure has another
seemingly unrelated side.
“Suppose you have an idea for a
new gadget; an idea so good that you
are willing to put your life on hold to
pursue it. But first, you want to make
sure the idea is novel. What would
you do? Ask your friends? Google a
few keywords that you believe
describe the idea? What if you don’t
use all the right key words and spend
precious time and resources invent-
ing a gadget that already exists?”
The problem in both scenarios
is the overreliance on keywords to
deliver information, Ribeiro says. In
the shopping scenario, the word
“socks” triggered a tornado of socks
ads; in the invention scenario, the
catch-22 is that we need to know the
right keywords to find information
about an unfamiliar topic.
“Part of my research harnesses
the collective power of a billion users
navigating the internet every day to
find better ways to deliver information — from what a specific user
wants to buy after he or she buys
socks to finding information without
keywords,” he says.
Ribeiro is seeking a mathematical formula to better understand user
navigation behavior that can help
improve both a user’s experience as
he or she searches for information
on the web and the relevance of the
advertiser offerings he or she sees.
He explains that a system or
model that understands the entire
user trajectory on the web, where
they start, where they want to go, and
when they are looking for something
else is possible and would soften the
“force” of these ads that pop up for
textbooks … or socks. It’s all about
that user experience.
“If we know the right way to get
you to what you want, that will be
really helpful. And the same mathematical model can be applied to
detecting account hijackings and
helping startups find new customers
through social networks,” Ribeiro
It’s happened to millions of Americans: You purchase
some socks on Amazon. Suddenly, there are ads for more
socks on every other website you visit. You can’t escape.
Images and deals for socks are every where. This happens
whether the item is a staple or a one-time purchase.
“Within the algorithmic failure to recognize that you
are not a sock aficionado lies a deeper issue in how
information is delivered today on the web,” says Bruno
SHOP AND NEVER DROP