Detection of tortured phrases in scientific literature

Eléna Martel (1, 2) , Martin Lentschat (1, 3) , Cyril Labbé (2, 1)

1 SIGMA - Systèmes d’Information - inGénierie et Modélisation Adaptables
2 LIG - Laboratoire d'Informatique de Grenoble
3 GETALP - Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole

This paper presents various automatic detection methods to extract so called tortured phrases from scientific papers. These tortured phrases, e.g. flag to clamor instead of signal to noise, are the results of paraphrasing tools used to escape plagiarism detection. We built a dataset and evaluated several strategies to flag previously undocumented tortured phrases. The proposed and tested methods are based on language models and either on embeddings similarities or on predictions of masked token. We found that an approach using token prediction and that propagates the scores to the chunk level gives the best results. With a recall value of .87 and a precision value of .61, it could retrieve new tortured phrases to be submitted to domain experts for validation.

Domains: Artificial Intelligence [cs.AI]; Computation and Language [cs.CL]; Digital Libraries [cs.DL]; Document and Text Processing

Eléna Martel, Martin Lentschat, Cyril Labbé. Detection of tortured phrases in scientific literature. Proceedings of the 2nd Workshop on Information Extraction from Scientific Publications, Nov 2023, Bali, Indonesia. ⟨hal-04423458⟩