Traditional Machine Learning and Deep Learning Models for Argumentation Mining in Russian Texts

Abstract

Argumentation mining is a field of computational linguistics that is devoted to extracting from texts and classifying arguments and relations between them, as well as constructing an argumentative structure. A significant obstacle to research in this area for the Russian language is the lack of annotated Russian-language text corpora. This article explores the possibility of improving the quality of argumentation mining using the extension of the Russian-language version of the Argumentative Microtext Corpus (ArgMicro) based on the machine translation of the Persuasive Essays Corpus (PersEssays). To make it possible to use these two corpora combined, we propose a Joint Argument Annotation Scheme based on the schemes used in ArgMicro and PersEssays. We solve the problem of classifying argumentative discourse units (ADUs) into two classes -"pro" ("for") and "opp" ("against") using traditional machine learning techniques (SVM, Bagging and XGBoost) and a deep neural network (BERT model). An ensemble of XGBoost and BERT models was proposed, which showed the highest performance of ADUs classification for both corpora.

RuArg-2022: Argument Mining Evaluation

“…(Fishcheva and Kotelnikov, 2019) translated into Russian and researched the English language Argumentative Microtext Corpus (ArgMicro) (Peldszus and Stede, 2015;Skeppstedt et al, 2018). In (Fishcheva et al, 2021) this corpus was expanded with machine translation of the Persuasive Essays Corpus (PersEssays) (Stab and Gurevych, 2014). XGBoost and BERT were applied to classify "for"/"against" premises.…”

“…The stance or premise label was chosen as the one where the corresponding input example had the maximum softmax output. sevastyanm (vyatsu) This participant utilized pre-trained ruRoberta-large language model 12 which was trained on additional data obtained from "PersEssays Russian" and "ArgMicro Russian" datasets (Fishcheva et al, 2021) with similar annotation schemes. Both datasets were united by argumentative discourse units and used to train model to solve 4-class classification problem.…”

Argumentative Text Generation in Economic Domain

“…The first annotated corpus for the Russian language (Fishcheva and Kotelnikov, 2019) was created based on the translation of the English language Argumentative Microtext Corpus (ArgMicro) (Peldszus and Stede, 2015;. It was then expanded with machine translation of the Persuasive Essays Corpus (PersEssays) (Stab and Gurevych, 2014) and a Joint Argument Annotation Scheme was proposed (Fishcheva et al, 2021). By using XGBoost and BERT, the authors were able to improve the results of automatic classification of "for" / "against" premises.…”

“…In our work, to the best of our knowledge, the problem of generating argumentative texts in Russian is being investigated for the first time. We are expanding the Russian-language corpus from (Fishcheva et al, 2021) by translating the UKP Sentential Argument Mining Corpus (UKP Sentential) (Stab et al, 2018). Based on the extended corpus, we train RuBERT model, which we then use to annotate sentences of the economic news corpus.…”

“…There are several publicly available pre-trained language models for the Russian language, including RuBERT (Kuratov and Arkhipov, 2019), SBERT (SBERT, 2020), andruGPT-3 (ruGPT-3, 2020). Recently, several papers have appeared on the argument mining in Russian (Fishcheva and Kotelnikov, 2019;Fishcheva et al, 2021;Salomatina et al, 2021;Ilina et al, 2021).…”