Accurate Detection of Automatically Spun Content via Stylometric Analysis

Shahid, Usman; Farooqi, Shehroze; Ahmad, Raza; Shafiq, Zubair; Srinivasan, Padmini; Zaffar, Fareed

doi:10.1109/icdm.2017.52

Cited by 11 publications

(14 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Word matching software such as Turnitin® (n.d.) has proven valuable in identifying replication of text from other sources. However, the very purpose of paraphrasing tools is to deceive software developed to detect plagiarism, and it is apparent that to date this strategy has been successful (Lancaster and Clarke 2009;Rogerson and McCarthy 2017;Shahid et al 2017). Consequently, the burden of detection remains with the human reader who has to become increasing adept at spotting stylistic variations and any other flags relating to mechanisms that have been used to avoid detection (Gillam et al 2010).…”

Section: Discussionmentioning

confidence: 99%

“…To create the spintax, a bank of potentially alternative terms is held in a synonym dictionary, which may be local to the paraphrasing tool, or held in cloud storage (Shahid et al 2017;Zhang et al 2014). In their study, Zhang et al (2014) were able to access this dictionary and reverse engineer two paraphrasing tools (Plagiarisma and The Best Spinner) to establish which words are subject to synonym substitution, referred to as 'mutables' , and which words do not appear in the synonym dictionary and thus would not be included in the spintax, referred to as 'immutables'.…”

Section: Paraphrasing Toolsmentioning

confidence: 99%

“…The paraphrasing tools that require a fee-based subscription provide a large number of parameters to manipulate the output, including the contents of the dictionary, the maximum number of synonyms used and replacement frequency, and the replacement of both single words and short phrases (Shahid et al 2017). In this study we assumed that the students were accessing the fee free version of online paraphrasing tools and as a result the output of spinning was less subject to control resulting in more words treated as mutables and thus less discretionary synonym substitution.…”

Section: Paraphrasing Toolsmentioning

confidence: 99%

See 2 more Smart Citations

Paraphrasing tools, language translation tools and plagiarism: an exploratory study

Prentice

Kinden

2018

Int J Educ Integr

View full text Add to dashboard Cite

In a recent unit of study in an undergraduate Health Sciences pathway course, we identified a set of essays which exhibited similarity of content but demonstrated the use of bizarre and unidiomatic language. One of the distinct features of the essays was the inclusion of unusual synonyms in place of expected standard medical terminology. We suspected the use of online paraphrasing tools, but were also interested in investigating the possibility of the use of online language translation tools. In order to test the outputs of these tools, we used as a seed document a corpus of text which had been provided to the students as prompt for the essay. This document was put through six free online paraphrasing tools and six separate iterative language translations through the online Google Translate™ tool. The results demonstrated that free online paraphrasing tools did not identify medical terminology as standardised or accepted nomenclature and substituted synonyms, whereas Google Translate™ largely preserved medical terminology. We believe that textual indicators such as the absence of standard discipline-based terminology may be of assistance in the identification of machine paraphrased text.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Paraphrasing Toolsmentioning

confidence: 99%

Section: Paraphrasing Toolsmentioning

confidence: 99%

See 1 more Smart Citation

Paraphrasing tools, language translation tools and plagiarism: an exploratory study

Prentice

Kinden

2018

Int J Educ Integr

View full text Add to dashboard Cite

show abstract

“…AlSallal et al [9] used common words and content words with SVM to detect plagiarism. Shahid et al [42] used syntactic and lexical features with SVM to detect "spun" content.…”

Section: Machine Learning Methods For Plagiarism Detectionmentioning

confidence: 99%

A Girl Has No Name: Automated Authorship Obfuscation using Mutant-X

Mahmood

Ahmad

Shafiq

et al. 2019

Proceedings on Privacy Enhancing Technologies

Self Cite

View full text Add to dashboard Cite

Authorship attribution aims to identify the author of a text based on the stylometric analysis. Authorship obfuscation, on the other hand, aims to protect against authorship attribution by modifying a text's style. In this paper, we evaluate the stealthiness of state-of-the-art authorship obfuscation methods under an adversarial threat model. An obfuscator is stealthy to the extent an adversary finds it challenging to detect whether or not a text modified by the obfuscator is obfuscated -a decision that is key to the adversary interested in authorship attribution. We show that the existing authorship obfuscation methods are not stealthy as their obfuscated texts can be identified with an average F1 score of 0.87. The reason for the lack of stealthiness is that these obfuscators degrade text smoothness, as ascertained by neural language models, in a detectable manner. Our results highlight the need to develop stealthy authorship obfuscation methods that can better protect the identity of an author seeking anonymity.

show abstract

“…In deception detection, author verification has been used to identify fake reviews by providing evidence that reviews published under different aliases are actually written by the same author (Layton, Watters, & Ureche, ). This technology can also be used to enhance recommender systems (Vaz, Martins de Matos, & Martins, ), opinion mining (Panicheva, Cardiff, & Rosso, ), personalized spam e‐mail detection (Shams & Mercer, ), and spun content detection (Shahid et al, ).…”

Section: Introductionmentioning

confidence: 99%

Improving author verification based on topic modeling

Potha

Stamatatos

2019

Asso for Info Science & Tech

View full text Add to dashboard Cite

Authorship analysis attempts to reveal information about authors of digital documents enabling applications in digital humanities, text forensics, and cyber‐security. Author verification is a fundamental task where, given a set of texts written by a certain author, we should decide whether another text is also by that author. In this article we systematically study the usefulness of topic modeling in author verification. We examine several author verification methods that cover the main paradigms, namely, intrinsic (attempt to solve a one‐class classification task) and extrinsic (attempt to solve a binary classification task) methods as well as profile‐based (all documents of known authorship are treated cumulatively) and instance‐based (each document of known authorship is treated separately) approaches combined with well‐known topic modeling methods such as Latent Semantic Indexing (LSI) and Latent Dirichlet Allocation (LDA). We use benchmark data sets and demonstrate that LDA is better combined with extrinsic methods, while the most effective intrinsic method is based on LSI. Moreover, topic modeling seems to be particularly effective for profile‐based approaches and the performance is enhanced when latent topics are extracted by an enriched set of documents. The comparison to state‐of‐the‐art methods demonstrates the great potential of the approaches presented in this study. It is also demonstrates that even when genre‐agnostic external documents are used, the proposed extrinsic models are very competitive.

show abstract

Accurate Detection of Automatically Spun Content via Stylometric Analysis

Cited by 11 publications

References 19 publications

Paraphrasing tools, language translation tools and plagiarism: an exploratory study

Paraphrasing tools, language translation tools and plagiarism: an exploratory study

A Girl Has No Name: Automated Authorship Obfuscation using Mutant-X

Improving author verification based on topic modeling

Contact Info

Product

Resources

About