Dr. Patrick Juola Research Project

Who Wrote That?

A police officer examines a ransom note. An airline receives an anonymous email bomb threat. A student turns in a suspiciously unoriginal essay. So who really wrote them? How can you tell?

For the past 36 months, Dr. Patrick Juola has been leading an authorship attribution study in hopes of creating software that can help authorities determine who wrote what. “It’s not just a legal issue,” said Juola, associate professor of computer science at Duquesne University. “It’s also important for historians. Our research goal is to enable a computer to look at a piece of text and say, ‘Yes, this play was written by Shakespeare’ or ‘Yes, this ransom note was written by a man in his early 40s.’”

To date, the National Science Foundation has awarded Juola with over $300,000 to support the project and he anticipates receiving a third wave of funding in the near future. Juola’s team includes James Overly and Darren Vescovi, both graduate students in the university’s Computational Mathematics program, and computer science undergraduates Peter Rutenbar and Sara Ali.

According to Juola, there are millions of tiny idiosyncrasies in our speech and text that offer telltale clues about who we are. For instance, you’re at a fancy dinner party with a formal setting of silverware in front of you. Is the salad fork “at” the left, “on” the left or on the “left-hand side”? “How you answer that question says something about who you are versus somebody else,” said Juola. “Authorship attribution also looks at spelling variations, grammar variations, and even what prepositions people like to use.”

Juola and his team plan to use their next round of funding to commercialize the software program they’re developing. “We’ve created a startup company called J Computing that will license the technology from Duquesne University, then identify a market and design a product to meet that specific need,” said Juola. “One example might be plagiarism detection in an academic context. So if you’re a high school senior thinking about plagiarizing your final paper for English class, don’t do it. We’ll catch you.”

J Computing expects to launch its first product in late 2011 in either the education or legal market. Juola’s also got his eye on the publishing industry to help validate authorship in journals.

In August, Dr. Juola presented the JGAAP authorship attribution software at ESSLLI 2010, the European Summer School in Logic, Language and Information in Copenhagen, Denmark where students and researchers from around the world gathered to attend classes and workshops. “This two-week long program attracts some of the next generation of world-class researchers,” said Juola. “So I’m delighted to bring this project to their attention, and I’ve already received some very valuable suggestions and feedback for further work.”

A professor at Duquesne University since 1998, Juola was recently inducted into the school’s Research Hall of Fame in recognition of his research and funding successes. To read more about Dr. Juola’s Authorship Attribution Project or download a demo of the software, visit the project web page at www.jgaap.com.