Automatic Detection of Ethnic Hate Speech in Russian-Language Blogs (EthnoHate)
Project Leader: Olessia Koltsova
Participants from SCILA: Ekaterina Pronoza, Polina Panicheva, Tatyana Yefimova, Maxim Terpilowski
International partners: Paolo Rosso
This research was a continuation of the 2015-2017 project "Development of a Concept and Methodology for Multilevel Monitoring of Interethnic Relations Based on Social Network Data," grant RFBR No. 15-18-00091. It aimed to improve the system for monitoring ethnosocial processes, focusing on automatic detection of ethno-relevant texts, automated sentiment analysis of such texts, and the development of an online monitoring system for ethnocultural and political processes. The "EthnoHate" study was focused exclusively on identifying ethnic hate speech in Russian-language blog texts. The research tasks included preparing a training corpus to recognize attitudes toward ethnic groups and developing a model for automatic detection of ethnic hate speech using traditional classifiers (Naïve Bayes, SVM, Logistic Regression, etc.) and neural networks (LSTM, BERT). Specifically, the project extensively tested various fine-tuning methods for the pre-trained BERT network on Russian-language informal texts for the task of identifying ethnic hate speech.
Publications:
Articles of the previous stage:
Participants from SCILA: Ekaterina Pronoza, Polina Panicheva, Tatyana Yefimova, Maxim Terpilowski
International partners: Paolo Rosso
This research was a continuation of the 2015-2017 project "Development of a Concept and Methodology for Multilevel Monitoring of Interethnic Relations Based on Social Network Data," grant RFBR No. 15-18-00091. It aimed to improve the system for monitoring ethnosocial processes, focusing on automatic detection of ethno-relevant texts, automated sentiment analysis of such texts, and the development of an online monitoring system for ethnocultural and political processes. The "EthnoHate" study was focused exclusively on identifying ethnic hate speech in Russian-language blog texts. The research tasks included preparing a training corpus to recognize attitudes toward ethnic groups and developing a model for automatic detection of ethnic hate speech using traditional classifiers (Naïve Bayes, SVM, Logistic Regression, etc.) and neural networks (LSTM, BERT). Specifically, the project extensively tested various fine-tuning methods for the pre-trained BERT network on Russian-language informal texts for the task of identifying ethnic hate speech.
Publications:
- Pronoza, E., Panicheva, P., Koltsova, O., & Rosso, P. (2021). Detecting ethnicity-targeted hate speech in Russian social media texts. Information Processing & Management, 58(6), 102674. Download (PDF, 3.53 Мб)
Articles of the previous stage:
- Koltsova, O., Nikolenko, S., Alexeeva, S., Nagornyy, O., Koltcov, S. (2017) Detecting Interethnic Relations with the Data from Social Media // Digital Transformation and Global Society: Second International Conference, DTGS 2017, St. Petersburg, Russia, June 21–23, 2017, Revised Selected Papers, pp.16-30.
- Bodrunova, S. S., Koltsova, O., Koltcov, S., & Nikolenko, S. (2017). Who’s Bad? Attitudes Toward Resettlers From the Post-Soviet South Versus Other Nations in the Russian Blogosphere. International Journal of Communication, 11, 3242–3264. http://ijoc.org/index.php/ijoc/article/view/6408
- Koltsova, O. Y., Alexeeva, S. V., Nikolenko, S. I., & Koltsov, M. (2017). Measuring Prejudice and Ethnic Tensions in User-Generated Content. Annual Review of CyberTherapy and Telemedicine, 15, 76–81. http://www.arctt.info/volume-15-summer-2017
- Apishev, M., Koltcov, S., Koltsova, O., Nikolenko, S., & Vorontsov, K. (2016). Mining Ethnic Content Online with Additively Regularized Topic Models. Computación y Sistemas, 20(3), 387–403. https://doi.org/10.13053/cys-20-3-2473
- Nikolenko, S. I., Koltcov, S., & Koltsova, O. (2017). Topic modelling for qualitative studies. Journal of Information Science, 43(1), 88–102. https://doi.org/10.1177/0165551515617393
Have you spotted a typo?
Highlight it, click Ctrl+Enter and send us a message. Thank you for your help!
To be used only for spelling or punctuation mistakes.