It was the graduate life for 17 Duke undergraduate students who spent their summer conducting research at the Department of Computer Science.
"It's a full-time research experience for most of them," said Professor Susan Rodger, who helped coordinate the summer research program with Computer Science Associate Chairman Richard Lucic. "They get to experience what it's like to go to graduate school. They're doing research; they're in an office; they're working on a team with other people; and they get exposed to other research areas."
Students worked nine to ten weeks during the summer semester, and met each Wednesday for department-provided lunches. Each lunch featured a different faculty member or graduate student as speaker. This year's series included Professor John Reif, who spoke about DNA computing, and Assistant Professor Ashwin Machanavajjhala, who talked about privacy in the big-data world.
"As part of the speaker series, we encourage the speakers to talk about the graduate school experience and to talk about why they went to graduate school. This lunch series is to encourage the undergraduates to think about going to graduate school," Rodger said. "Most of the undergraduates do not realize that they do not have to pay for graduate school. They think they need to get a job right away."
Rising senior Andrew Shim is one of the students who participated this summer as a requirement of the Computer Science Undergraduate Research Fellows Program, which also seeks to interest undergraduates in graduate school by involving them in research. Shim worked with Associate Professor Jun Yang on a project in computational journalism to examine the validity of claims by politicians and others in the public eye. "In essence, we want to automate what the people at factcheck.org are doing," Shim said of the reverse engineering queries. As an example, he noted former New York Mayor Rudy Giuliani once claimed the number of adoptions in New York increased by about 65 to 70 percent across his term. The claim is verified when looking at particular time frames in Giuliani's incumbency. "But then we ask ourselves, 'Why are we looking at these particular time periods?'" Shim said. "It turns out that towards the end of Giuliani's term, the number of adoptions actually decreased after having been on the rise even before Giuliani came into office. So Giuliani's data staff chose particular time windows that would mask the trend of decrease and highlight that of increase."
Rachel Harris, a rising senior who also participated as part of the C-SURF program, worked this summer on a project in pattern recognition with Professor Carlo Tomasi. Given a handwritten document and a transcription of that document, her goal was to find additional features that could best distinguish one character from another to aid in document synchronization. "What we're trying to do is match each character in the transcription to the image of that character in the manuscript," she said. Features describing sections of each character could include the percentage of black pixels vertically above or below a baseline, indicating an ascending or descending letter; the percentage of writing in a slice of an image; or the number of times a black line is crossed when traveling vertically from the bottom of an image (two times in the letters "D" and "O," and once in the letter "T"). The research could lead to more in-depth search capabilities within handwritten documents and to automatic transcription of such documents.
Prior to the summer program, Harris thought she knew what graduate school would entail -- more classes, research, a thesis and dissertation. "You can know that intellectually, but it's very different when you're actually sitting here," she said. "Day to day, I'm setting my own schedule. There's no assigned weekly reading. It's a lot more responsibility, but I think it's better as a system. You do more and you learn more."
Harris is considering graduate school and noted she learned a lot from the summer program, including how to manage her time and interact with other researchers. "It's a very short time period, so it feels very intense," she said. "There's a steep learning curve at the start."