Abstract
Linguistic literature on irony discusses sarcasm as a form of irony characterized by its biting nature and the intention to mock a victim. This particular trait makes sarcasm apt to convey hate speech and not only humour. Previous works on abusive language stressed the need to address ironic language to lead the system to recognize correctly hate speech, especially in spontaneous texts, like tweets [13]. In this context, our main hypothesis is that information about the presence of sarcasm could help to improve the detection of hateful messages, especially when they are camouflaged as sarcastic. To corroborate this hypothesis: i) we perform analysis on HaSpeeDe20_ext, an Italian corpus of tweets about the integration of cultural minorities in Italy, ii) we carry out computational experiments injecting the knowledge of sarcasm in a system of hate speech detection, and iii) we adopt strategies of validation in terms of performance and significance of the obtained results. Results confirm our hypothesis and overcome the state of the art.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
One of the most complete definitions is provided by [25]: a content is considered hateful on the basis of its action and its target. The action is the illocutionary act of the utterance aimed to spread or justify hate, incite violence, or threat people’s freedom, dignity, and safety. The target must be a protected group or an individual belonging to such a group, attacked for his/her individual characteristics.
- 2.
Sir, everybody has the right to a dignified life, but if you put a migrant in my way, I will be Salvini. (Matthew 15, 83).
- 3.
A plate of pasta and let’s apologize for not being Muslims too. Maybe then they become our friends and won’t kill us anymore.
- 4.
- 5.
- 6.
- 7.
These three language models are trained on different genres of texts in Italian and available on the Hugging Face platform: https://huggingface.co/models.
- 8.
The schema of annotation of ironic language is inherited by [5] who annotated the IronITA corpus of tweets for the first time, to our knowledge, as ironic and sarcastic.
- 9.
In particular, they used a linear layer with a softmax on top of the CLS token, applying a novel technique of layer-wise learning rate. That is the main difference with our approach.
- 10.
So if an Italian dies in the hospital in the midst of ants it is an ‘episode’ while if a Nigerian dies of a circumcision, free healthcare is required for immigrants. Stuff from civil war and riots up to the ramparts of Orion.
- 11.
References
Basile, V., et al.: SemEval-2019 task 5: multilingual detection of hate speech against immigrants and women in Twitter. In: Proceedings of the 13th International Workshop on Semantic Evaluation, pp. 54–63 (2019)
Baxter, J.: A Bayesian/information theoretic model of learning to learn via multiple task sampling. Mach. Learn. 28(1), 7–39 (1997)
Bowes, A., Katz, A.: When sarcasm stings. Discourse Process. Multi. J. 48(4), 215–236 (2011)
Caselli, T., Basile, V., Mitrović, J., Kartoziya, I., Granitzer, M.: I feel offended, don’t be abusive! implicit/explicit messages in offensive and abusive language. In: Proceedings of the 12th Language Resources and Evaluation Conference, pp. 6193–6202 (2020)
Cignarella, A.T., Frenda, S., Basile, V., Bosco, C., Patti, V., Rosso, P.: Overview of the EVALITA 2018 task on irony detection in Italian tweets (IronITA). In: Proceedings of the Sixth Evaluation Campaign of Natural Language Processing and Speech Tools for Italian (EVALITA 2018) Co-located with the Fifth CLiC-it, vol. 2263, pp. 1–6 (2018)
Cimino, A., De Mattei, L., Dell’Orletta, F.: Multi-task learning in deep neural networks at EVALITA 2018. In: Proceedings of the Sixth Evaluation Campaign of Natural Language Processing and Speech Tools for Italian (EVALITA 2018) Co-located with the Fifth CLiC-it, vol. 2263 (2018)
Davies, C.: Jokes and Targets. Indiana University Press, Bloomington (2011)
Devlin, J., Chang, M., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT 2019), pp. 4171–4186 (2019)
Dews, S., Winner, E.: Muting the meaning a social function of irony. Metaphor Symbolic Act. 10(1), 3–19 (1995)
Douglass, S., Mirpuri, S., English, D., Yip, T.: ‘They were just making jokes’: ethnic/racial teasing and discrimination among adolescents. Cultur. Divers. Ethnic Minor. Psychol. 22(1), 69–82 (2016)
Fersini, E., Rosso, P., Anzovino, M.: Overview of the task on automatic misogyny identification at IberEval 2018. In: Proceedings of the Third Workshop on Evaluation of Human Language Technologies for Iberian Languages (IberEval 2018) co-located with 34th Conference of SEPLN, vol. 2150, pp. 214–228 (2018)
Fortuna, P., Nunes, S.: A survey on automatic detection of hate speech in text. ACM Comput. Surv. 51(4), 85:1–85:30 (2018)
Frenda, S., Patti, V., Rosso, P.: Killing me softly: creative and cognitive aspects of implicitness in abusive language online. Nat. Lang. Eng. 1–22 (2022)
Gambino, G., Pirrone, R.: CHILab @ HaSpeeDe 2: enhancing hate speech detection with part-of-speech tagging. In: Proceedings of the Seventh Evaluation Campaign of Natural Language Processing and Speech Tools for Italian (EVALITA 2020), vol. 2765 (2020)
Kuipers, G., Van der Ent, B.: The seriousness of ethnic jokes: ethnic humor and social change in the Netherlands, 1995–2012. Humor 29(4), 605–633 (2016)
Lavergne, E., Saini, R., Kovács, G., Murphy, K.: TheNorth @ HaSpeeDe 2: BERT-based language model fine-tuning for Italian hate speech detection. In: Proceedings of the Seventh Evaluation Campaign of Natural Language Processing and Speech Tools for Italian (EVALITA 2020), vol. 2765, pp. 142–147 (2020)
Lee, C.J., Katz, A.N.: The differential role of ridicule in sarcasm and irony. Metaphor. Symb. 13(1), 1–15 (1998)
Lemmens, J., Markov, I., Daelemans, W.: Improving hate speech type and target detection with hateful metaphor features. In: Proceedings of the Fourth Workshop on NLP for Internet Freedom: Censorship, Disinformation, and Propaganda, pp. 7–16 (2021)
MacAvaney, S., Yao, H.R., Yang, E., Russell, K., Goharian, N., Frieder, O.: Hate speech detection: challenges and solutions. PLoS ONE 14(8), 1–16 (2019)
Merlo, L.I., Chulvi, B., Ortega, R., Rosso, P.: When humour hurts: linguistic features to foster explainability. Procesamiento Leng. Nat. (SEPLN) 70, 85–98 (2023)
Nobata, C., Tetreault, J., Thomas, A., Mehdad, Y., Chang, Y.: Abusive language detection in online user content. In: Proceedings of the 25th International Conference on World Wide Web (WWW 2016), pp. 145–153 (2016)
Pexman, P.M., Olineck, K.M.: Does sarcasm always sting? Investigating the impact of ironic insults and ironic compliments. Discourse Process. 33(3), 199–217 (2002)
Poletto, F., Basile, V., Sanguinetti, M., Bosco, C., Patti, V.: Resources and benchmark corpora for hate speech detection: a systematic review. Lang. Resour. Eval. 55, 477–523 (2021)
Sanguinetti, M., et al.: Haspeede 2 @ EVALITA2020: overview of the EVALITA 2020 hate speech detection task. In: Proceedings of the Seventh Evaluation Campaign of Natural Language Processing and Speech Tools for Italian (EVALITA 2020), vol. 2765 (2020)
Sanguinetti, M., Poletto, F., Bosco, C., Patti, V., Stranisci, M.: An Italian twitter corpus of hate speech against immigrants. In: Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018) (2018)
Schmidt, A., Wiegand, M.: A survey on hate speech detection using natural language processing. In: Proceedings of the Fifth International Workshop on Natural Language Processing for Social Media, pp. 1–10 (2017)
Søgaard, A., Johannsen, A., Plank, B., Hovy, D., Martínez Alonso, H.: What’s in a p-value in NLP? In: Proceedings of the Eighteenth Conference on Computational Natural Language Learning, pp. 1–10 (2014)
Taulé, M., Ariza, A., Nofre, M., Amigó, E., Rosso, P.: Overview of DETOXIS at IberLEF 2021: DEtection of TOXicity in comments in Spanish. Procesamiento Leng. Nat. 67, 209–221 (2021)
Waseem, Z., Davidson, T., Warmsley, D., Weber, I.: Understanding abuse: a typology of abusive language detection subtasks. In: Proceedings of the First Workshop on Abusive Language Online, pp. 78–84 (2017)
Waseem, Z., Hovy, D.: Hateful symbols or hateful people? Predictive features for hate speech detection on twitter. In: Proceedings of the NAACL Student Research Workshop, pp. 88–93 (2016)
Weaver, S.: A rhetorical discourse analysis of online anti-Muslim and anti-Semitic jokes. Ethn. Racial Stud. 36(3), 483–499 (2013)
Wiegand, M., Ruppenhofer, J., Eder, E.: Implicitly abusive language - what does it actually look like and why are we not getting there? In: Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 576–587 (2021)
Wiegand, M., Ruppenhofer, J., Kleinbauer, T.: Detection of abusive language: the problem of biased datasets. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, vol. 1, pp. 602–608 (2019)
Zampieri, M., Malmasi, S., Nakov, P., Rosenthal, S., Farra, N., Kumar, R.: SemEval-2019 task 6: identifying and categorizing offensive language in social media (OffensEval). In: Proceedings of the 13th International Workshop on Semantic Evaluation, pp. 75–86 (2019)
Acknowledgments
The work of S. Frenda and V. Patti was partially funded by the Multilingual Perspective-Aware NLU Project in partnership with Amazon Alexa. The work of the P. Rosso was done in the framework of the FairTransNLP-Stereotypes research project on Fairness and Transparency for equitable NLP applications in social media: Identifying stereotypes and prejudices and developing equitable systems (Grant PID2021-124361OB-C31 funded by MCIN/AEI/10.130 39/501100011033 and by ERDF, EU A way of making Europe).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Frenda, S., Patti, V., Rosso, P. (2023). When Sarcasm Hurts: Irony-Aware Models for Abusive Language Detection. In: Arampatzis, A., et al. Experimental IR Meets Multilinguality, Multimodality, and Interaction. CLEF 2023. Lecture Notes in Computer Science, vol 14163. Springer, Cham. https://doi.org/10.1007/978-3-031-42448-9_4
Download citation
DOI: https://doi.org/10.1007/978-3-031-42448-9_4
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-42447-2
Online ISBN: 978-3-031-42448-9
eBook Packages: Computer ScienceComputer Science (R0)