Program to detect fake papers

Researchers in the IU School of Informatics have developed a program to detect fake technical papers and articles. The program, called the Inauthentic Paper Detector, is designed to analyze technical papers and determine whether they were created by a machine or by human hands. \nThe program analyzes the papers, and compares them to a database comprised of real papers, computer generated papers and fake papers derived of authentic ones. Senior informatics major Wyatt Clark says the system is relatively simple compared to other similar programs, because it uses such a small database.\n"Our predictor is simple compared to that of a Web page like turnitin.com, which is actually designed to detect plagiarism, and searches a database of billions of Web pages and previously written papers," Clark said in an e-mail. "Relative to the size of our database and the simplicity of our method our page does an alright job detecting 'fake papers'."\nClark went on to say that the program works differently from other programs like turnitin.com, because it "is independent of the vocabulary used in the paper" and only looks "at the repetitiveness or flow of the paper."\nClark was joined in his efforts by assistant professors Mehmet Dalkilic and Predrag Radivojac, as well as informatics doctoral student James Costello. The team began working last May, and completed their work by presenting their results at the Society of Industrial and Applied Mathematical Conference between April 20 and April 22. \nThe program has been under development since last year. All the work was done within the School of Informatics, and the team worked for approximately 10 hours a week on the project. \nClark said he was inspired to help undertake the project because he was interested in it and it helped him to gain experience and notoriety in his field of study. He also said he believed the program could be of use to the scientific community as a whole, not just specifically IU.\nClark said he knows that no program is perfect and that the IPD has its flaws. But he said he believed that the project was a success and the program was positive and ultimately beneficial. \n"I do believe that the page is a good predictor of the amount of information that a piece of text contains," Clark said in an e-mail. "I'm sure there are people out there that are writing scripts as we speak to mimic the repetitiveness of a paper in order to directly fool our predictor"