Methods for Detoxification of Texts for the Russian Language

Abstract

We introduce the first study of the automatic detoxification of Russian texts to combat offensive language. This kind of textual style transfer can be used for processing toxic content on social media or for eliminating toxicity in automatically generated texts. While much work has been done for the English language in this field, there are no works on detoxification for the Russian language. We suggest two types of models—an approach based on BERT architecture that performs local corrections and a supervised approach based on a pretrained GPT-2 language model. We compare these methods with several baselines. In addition, we provide the training datasets and describe the evaluation setup and metrics for automatic and manual evaluation. The results show that the tested approaches can be successfully used for detoxification, although there is room for improvement.

Toxic Fake News Detection and Classification for Combating COVID-19 Misinformation

“…Fake news generally has these toxic and non-toxic emotions associated with it. Toxicity refers to any undesirable or unwanted behaviour shown by the users on the internet with the intent to offend, insult, or cause harm to an individual or community [46]. These toxic emotions can have several attributes such as Toxicity, Severe Toxicity, Obscene, Identity Attack, Insult, Threat, and Sexually Explicit.…”

Towards a Russian Multimedia Politeness Corpus

“…In a broader context, there have been works on speech strategies in interactions between interlocutors in situations of persuasion and provocation (Issers, 2009), functions of imperative verb forms in a situation of request or prompt (Paducheva, 2010), and extralinguistic conditions of spontaneous speech interactions (Zemskaja et al, 1981). Most of these studies describe qualitative traits of communication while quantitative studies have been confined to the tasks of automatic detection of speech aggression and detoxification of online communication (Dementieva et al, 2021;Dementieva et al, 2022). The aim of our project is to develop the annotation framework for oral communicative interaction that on one hand takes into account approaches from linguistic politeness theories and on the other could be used for qualitative research and the NLP applications.…”

Exploring Cross-lingual Text Detoxification with Large Multilingual Language Models.

“…The task of Textual Style Transfer (Textual Style Transfer) can be viewed as a task where certain properties of text are being modified while rest retain the same 1 . In this work we focus on detoxification textual style transfer (dos Santos et al, 2018a;Dementieva et al, 2021a). It can be formulated as follows: given two text corpora D X = {x 1 , x 2 , .…”

“…Textual style transfer gained a lot of attention with a rise of deep learning-based NLP methods. Given that, Textual Style Transfer has now a lot of specific subtasks ranging from formality style transfer (Rao and Tetreault, 2018;Yao and Yu, 2021) and simplification of domain-specific texts (Devaraj et al, 2021;Maddela et al, 2021) to emotion modification (Sharma et al, 2021) and detoxification (debiasing) Dementieva et al, 2021a).…”

“…There exist a variety of Textual Style Transfer methods: from totally supervised methods (Wang et al, 2019b;Zhang et al, 2020;Dementieva et al, 2021a) which require a parallel text corpus for training to unsupervised (Shen et al, 2017;Wang et al, 2019a;Xu et al, 2021) that are designed to work without any parallel data. The latter sub-field of research is more popular nowadays due to the scarcity of parallel text data for Textual Style Transfer.…”