Abstract:

Plagiarism has become a serious problem in academic writing over the years. The rise of the internet has improved the accessibility of information and facilitated the sharing of media, this in turn has made plagiarizing academic work something trivial. The huge availability of sources of material that can be found on the internet make finding a source to plagiarism from as easy as finding a grain of salt on a desert, and conversely detecting whether a paper has been plagiarized is equivalent to finding where that grain of sand used to belong to.

However not all hope is gone. With the years not only has the availability of information improved but also computer's computational power, storage and clever algorithms such as Deep Learning which allows us to use that computational power and storage to train models capable of learning and automating task that were previously though to be almost impossible.

In this survey paper we analyze the evolution of different natural language processing algorithms being used for extrinsic and intrinsic plagiarism detection, the types of plagiarism they excel at finding and the ones they struggle with, their limitations, performance and in which direction the future of plagiarism detection seems to be heading.

Keywords:

Deep Learning, Natural Language Processing, Neural Networks, Plagiarism Detection.