Plagiarism is a widespread problem in assessment tasks; in computing courses, students often plagiarise source code. For all but the smallest classes, manual detection of such plagiarism is impractical, and, while automated tools are available, none has been applied to detect inter-lingual plagiarism, where source code is copied from one language to another. In this work, we propose a novel approach, XPlag, to detect plagiarism involving multiple languages using intermediate program code produced by a compiler suite. We describe experiments to evaluate XPlag, and show that we can detect inter-lingual plagiarism with reasonably good precision.

Plagiarism and copyright infringement are major problems in academic and corporate environments. Existing solutions for detecting infringements in structured text such as source code are restricted to textual similarity comparisons of two pieces of work. In this paper, we examine authorship attribution as a means for tackling plagiarism detection. Given several samples of work from several authors, we attempt to correctly identify the author of work presented as a query. On a collection of 1 640 documents written by 100 authors, we show that we can attribute authorship in up to 67 % of cases. This work can be a valuable additional indicator for the more difficult plagiarism investigations.

Abstract: When students have little to lose and a great deal to gain by cheating, some of them will do so. Students have even less to lose if caught cheating in remote online courses than in face-toface courses, and so are more likely to cheat in online courses. Th is paper describes an electronic ‘watermarking ’ system that we have used to detect a particular form of cheating in remote online exams. Findings from the use of the system raise some disturbing questions about the practice of distance education and education in general.

Students are often asked to submit electronic copies of their program code as part of assessment in computer science courses. To counter code plagiarism, educational institutions use tools to detect similarity between submissions. Previous research has identified that using a modified text search engine to identify similar code within large code collections is both efficient and effective. The similarity functions used internally by such search engines have historically been devised manually by experts in the field; in this work, we investigate the practicality of using evolutionary computing techniques to evolve similarity functions. We use particle swarm optimisation to find optimal values of variables in human constructed similarity functions, and use genetic programming to generate new similarity functions specifically for this task. We show empirically that our optimised similarity functions perform better than standard Okapi BM25 across a range of collections. Our results indicate that evolutionary computing techniques have promise in evolving code similarity functions, but overtraining was evident for our genetically programming-evolved