Detecting online plagiarism

Mar 06, 2003, vol. 26, no. 5



Document Tools

Print This Article

E-mail This Page

Font Size
S      M      L      XL

Related Stories

Two SFU professors and a PhD graduate student have spent the past year assisting a local Vancouver company to develop a new software program capable of detecting plagiarism from online sources.

Maite Taboada, an assistant professor in linguistics, Veronica Dahl, a professor in computing science, and computing PhD graduate student Manual Zahariev, helped Vancouver Software Labs (VSL) create a new approach to plagiarism that detects not only exact copying but also adapted copying.

“Most approaches to plagiarism compare the text and figure out whether a piece of text is exactly the same as another,” explains Maite. “What we're also looking for is slight changes in wording, which is also a very common form of plagiarism.”

The SFU trio used computational linguistics to develop algorithms for finding similar text. Computational linguistics is the scientific study of language from a computational perspective. It provides models of various linguistic phenomena, which can then be used in practical applications, such as machine translation, speech recognition, web search engines or text editors. The SFU computational linguists found a way to automate text comparison techniques that are commonly used in literary analysis to determine, for example, whether Shakespeare really did author all of his works.

The new plagiarism software is a value-added feature to VSL's on-line software product, called Scriptum, which enables teachers to accept online assignments, then grade and annotate them online before returning them to students. Now, they can also check assignments for plagiarism. The recently released product is in use at about 20 post-secondary institutions across Canada. There are no SFU professors yet using the system.

Search SFU News Online