A Window on Research

Tufts engineering grad students host Morehouse College undergrads who are tracking Twitter trends and using brain imaging technologies to explore human cognition
Brockton Chase Starling and Jordan Crouser
Morehouse College undergraduate Brockton Chase Starling and School of Engineering graduate student Jordan Crouser compare notes at the new Interdisciplinary Laboratory for Computation. Photo: Matthew Modoono
August 23, 2011

Share

When Brockton Chase Starling came to Tufts this summer to do research, his topic was as wide as the world: to create a visual representation of trending topics on Twitter, and how those issues and ideas travel the globe.

“We’re looking at how Twitter and other social networks are starting to play a big role in society and decided to look at how trends originate and spread,” says Starling.

He was one of three undergraduates who came to Tufts School of Engineering’s computer science department from Morehouse College in Atlanta to take part in the Leadership Alliance, a nationwide program that provides research experience for undergraduates. The students worked side-by-side with Tufts graduate students, investigating issues in computer science such as human-computer interaction and data visualization. They were joined by other Leadership Alliance students working in research labs and with faculty across the university.

“The Leadership Alliance program gives students a window into the research world,” says Yvette Dalton-McCoy, the associate director of graduate diversity programs who administers the Leadership Alliance program at Tufts. “Research experiences provided by our Tufts faculty are invaluable for keeping the pipeline flowing from undergrad to graduate school for underrepresented students in engineering and the sciences.”

The Morehouse students worked in the Interdisciplinary Laboratory for Computation with Remco Chang, an assistant professor of computer science in the School of Engineering. “It’s great having these students around during the summer,” says Chang. “It makes my lab more dynamic, and the graduate students here learn something about their own mentoring style, too.”

We’re All Connected

Starling, a computer science major at Morehouse, worked with doctoral student R. Jordan Crouser, E13, on the Twitter trend project. “The project entails working with the Twitter API, which is basically the source code. It’s coding and parsing Twitter data,” says Starling.

From that data, Starling and Crouser wanted to create a visual representation from the first mention of a particular trending topic—such as raising the U.S. debt ceiling—to its propagation through social networks.

Using a data visualization computer program, they plotted movement of trending topics on a global map, “so you can see where they start and how they’re connected,” says Starling.

“The project was started because we were interested in how Twitter specifically, and social media platforms in general, supported some of the uprisings that we recently saw in Egypt and other places,” says Crouser. “With Twitter, you can access it from a mobile phone, so even if the government shuts down the Internet, there’s still this connectivity—there’s still some way to communicate.”

It’s All in Your Mind

Two other Morehouse students, Austin Tucker and Jamal Thorne, worked on research projects that use a brain imaging technology called fNIRS, or functional near-infrared spectroscopy. The fNIRS tracks blood flow increases to the parts of the brain that are being actively used.

In the lab, a test subject is fitted with a headband-like device that sends near-infrared light through the forehead at a relatively shallow depth—two to three centimeters—to interact with the brain’s frontal lobe. Light usually passes through the body’s tissues, except when it encounters oxygenated or deoxygenated hemoglobin in the blood. Light waves are absorbed by the active, blood-filled areas of the brain, and any remaining light is captured by the fNIRS’ detectors.

Tucker worked with computer science graduate student Megan Strait, E15, who conducts fNIRS research in the Human-Robotics Interaction Laboratory. Strait uses fNIRS to decipher what our brains look like when we ask robots to interpret a morally ambiguous situation, such as choosing who to save first from a burning building.

The fNIRS device could help interpret the mental workload—how much we use our brains—when we instruct robots to make that kind of ethical decision.

“The hypothesis is that asking a robot to do a task creates less workload for you than if you were in the situation yourself—that you consider the situation less because you're a layer removed,” says Strait. 

“Given that robots are going to become ubiquitous in society—that there are going to be significantly more autonomous robots—how much autonomy do we want to give robots in this kind of ethically questionable situation?” says Tucker. He notes that providing autonomy can get further complicated when robots must make real-time decisions in a warfare situation.

“The robot only has a certain limited amount of time to process this information, to act on what is morally ‘more correct,’ and decide which action would have fewer bad consequences,” says Tucker, who helped design automated data collection from the fNIRS device, providing a platform for Strait’s team to gather data on a larger scale.

You May Also Like

Morehouse student Jamal Thorne worked with doctoral student Evan Peck, E13, to create a brain-driven recommendation system using fNIRS data.

Current automated recommendation systems—like those used by Netflix and the music service Pandora—ask users to rate how much they like or dislike a particular movie or song. The recommendation system then returns suggestions based on preferences across all users.

“Say you rate Pirates of the Caribbean as 5,” says Thorne. “Then the system is going to search for somebody else who rated that movie as a 5 and send you a suggestion from that other person’s list of other movies that they rates as 5s. Hopefully what they liked you would like as well.”

The reason movie recommendation systems work well, Peck says, is because people are frequently buying and rating movies. But the system falls apart when purchasers buy a product only sporadically, such as a new car.

Peck is working on an fNIRS recommendation system that would take active user rating out of the equation, and would simply “read” brain activity as an indicator of preference of just about anything: movies, cars or music.

“We’re basically trying to see if brain information affects recommendation engines,” said Thorne.

The proposed futuristic fNIRS system would passively collect data from the “brain information” from all users and create suggestions seamlessly based on only your thoughts.

“You can imagine a Pandora system where you just listen and the music gradually adjusts to what you like, and you never have to click thumbs up or thumbs down for any song,” says Peck

The three Morehouse undergraduates presented posters on their research at the Leadership Alliance National Symposium in Old Greenwich, Conn., at the end of July.

Even though 10 weeks was a short amount of time to make a lot of headway on such complex projects, the research “has potential impact on a really large scope,” says Chang, whose graduate students will continue to work on these and other complex computer science questions. “This will change the way people behave—this will change the way that society works.”

Julia Keller, communications manager for the School of Engineering, can be reached at j.keller@tufts.edu.