Abstract
Counterfactually augmented data has recently been proposed as a successful solution for socially situated NLP tasks such as hate speech detection. The chief component within the existing counterfactual data augmentation pipeline, however, involves manually flipping labels and making minimal content edits to training data. In a hate speech context, these forms of editing have been shown to still retain offensive hate speech content. Inspired by the recent success of large language models (LLMs), especially the development of ChatGPT, which have demonstrated improved language comprehension abilities, we propose an inclusivity-oriented approach to automatically generate counterfactually augmented data using LLMs. We show that hate speech detection models trained with LLM-produced counterfactually augmented data can outperform both state-of-the-art and human-based methods.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
- 2.
Note that on account of being a special-purpose, manually curated dataset for the task of hate speech detection there are higher than normal percentages of hate speech texts.
- 3.
- 4.
We used the lexicons from https://github.com/peterkwells/uk-attitudes-to-offensive-language-and-gestures-data/.
- 5.
We used the lexicons from https://github.com/surge-ai/profanity.
- 6.
With generations = 5, population_size = 40.
- 7.
With max_epochs = 5, batch_size = 32 (except for the third Experiment, we use 10), learning_rate = 1e−5.
- 8.
For BERT, each fold’s original test set was divided into 50%-50% validation and test set.
- 9.
A qualitative analysis of the data revealed coverage of a vast range of issues from gays in Islam to Republicans to Catholicism. In fact, the dataset diversity is highest for tweets belonging to domain “Religion”.
References
Atanasova, P., Simonsen, J.G., Lioma, C., Augenstein, I.: Fact checking with insufficient evidence. Trans. Assoc. Comput. Linguist. 10, 746–763 (2022)
Badjatiya, P., Gupta, S., Gupta, M., Varma, V.: Deep learning for hate speech detection in tweets. In: Proceedings of the 26th International Conference on World Wide Web Companion, pp. 759–760 (2017)
Bail, C.A.: Can generative AI improve social science? (2023)
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)
Belinkov, Y., Poliak, A., Shieber, S.M., Van Durme, B., Rush, A.M.: Don’t take the premise for granted: mitigating artifacts in natural language inference. arXiv preprint arXiv:1907.04380 (2019)
Bender, E.M., Gebru, T., McMillan-Major, A., Shmitchell, S.: On the dangers of stochastic parrots: can language models be too big?. In: Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, pp. 610–623 (2021)
Brown, T., et al.: Language models are few-shot learners. Adv. Neural. Inf. Process. Syst. 33, 1877–1901 (2020)
Caton, S., Haas, C.: Fairness in machine learning: a survey. ACM Comput. Surv. (2023)
Clark, D.M.: Drag them: a brief etymology of so-called “cancel culture". Commun. Pub. 5(3–4), 88–92 (2020)
Davani, A.M., Díaz, M., Prabhakaran, V.: Dealing with disagreements: looking beyond the majority vote in subjective annotations. Trans. Assoc. Comput. Linguist. 10, 92–110 (2022)
Davidson, T., Warmsley, D., Macy, M., Weber, I.: Automated hate speech detection and the problem of offensive language. In: Proceedings of the International AAAI Conference on Web and Social Media, vol. 11, pp. 512–515 (2017)
Dillion, D., Tandon, N., Gu, Y., Gray, K.: Can AI language models replace human participants? Trends Cogn. Sci. (2023)
Feng, S.Y., et al.: A survey of data augmentation approaches for NLP. In: Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pp. 968–988 (2021)
Gardner, M., et al.: Evaluating models’ local decision boundaries via contrast sets. arXiv preprint arXiv:2004.02709 (2020)
Garg, T., Masud, S., Suresh, T., Chakraborty, T.: Handling bias in toxic speech detection: a survey. ACM Comput. Surv. 55(13s), 1–32 (2023)
Gibson, A.: Free speech and safe spaces: how moderation policies shape online discussion spaces. Soc. Media+ Soc. 5(1), 2056305119832588 (2019)
Howard, P., Singer, G., Lal, V., Choi, Y., Swayamdipta, S.: Neurocounterfactuals: beyond minimal-edit counterfactuals for richer data augmentation. In: Findings of the Association for Computational Linguistics: EMNLP 2022, pp. 5056–5072 (2022)
Jahan, M.S., Oussalah, M.: A systematic review of hate speech automatic detection using natural language processing. Neurocomputing, 126232 (2023)
Kaushik, D., Hovy, E., Lipton, Z.: Learning the difference that makes a difference with counterfactually-augmented data. In: International Conference on Learning Representations (2019)
Kenton, J.D.M.W.C., Toutanova, L.K.: Bert: pre-training of deep bidirectional transformers for language understanding. In: Proceedings of NAACL-HLT, vol. 1, p. 2 (2019)
Kumar, A., Tan, C., Sharma, A.: Probing classifiers are unreliable for concept removal and detection. arXiv preprint arXiv:2207.04153 (2022)
Le, T.T., Fu, W., Moore, J.H.: Scaling tree-based automated machine learning to biomedical big data with a feature set selector. Bioinformatics 36(1), 250–256 (2020)
Le Bras, R., et al.: Adversarial filters of dataset biases. In: Proceedings of the 37th International Conference on Machine Learning, pp. 1078–1088 (2020)
Madiega, T.A.: Artificial intelligence act. European Parliamentary Research Service, European Parliament (2021)
Min, B., et al.: Recent advances in natural language processing via large pre-trained language models: a survey. arXiv preprint arXiv:2111.01243 (2021)
Motoki, F., Neto, V.P., Rodrigues, V.: More human than human: measuring ChatGPT political bias. Public Choice, pp. 1–21 (2023)
Ouyang, L., et al.: Training language models to follow instructions with human feedback. Adv. Neural. Inf. Process. Syst. 35, 27730–27744 (2022)
Pearl, J.: Causal and counterfactual inference. the handbook of rationality (2019)
Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., Sutskever, I., et al.: Language models are unsupervised multitask learners. OpenAI Blog 1(8), 9 (2019)
Ramponi, A., Tonelli, S.: Features or spurious artifacts? data-centric baselines for fair and robust hate speech detection. In: Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 3027–3040. Association for Computational Linguistics (2022)
Ranasinghe, T., Zampieri, M., Hettiarachchi, H.: Brums at HASOC 2019: deep learning models for multilingual hate speech and offensive language identification. In: FIRE (working notes), pp. 199–207 (2019)
Ray, P.P.: ChatGPT: a comprehensive review on background, applications, key challenges, bias, ethics, limitations and future scope. Internet Things and Cyber-Phys. Syst. (2023)
Ross, A., Wu, T., Peng, H., Peters, M.E., Gardner, M.: Tailor: generating and perturbing text with semantic controls. In: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 3194–3213 (2022)
Samory, M., Sen, I., Kohne, J., Flöck, F., Wagner, C.: “call me sexist, but...”: revisiting sexism detection using psychological scales and adversarial samples. In: Proceedings of the International AAAI Conference on Web and Social Media, vol. 15, pp. 573–584 (2021)
Sarwar, S.M., Murdock, V.: Unsupervised domain adaptation for hate speech detection using a data augmentation approach. In: Proceedings of the International AAAI Conference on Web and Social Media, vol. 16, pp. 852–862 (2022)
Schlangen, D.: Targeting the benchmark: on methodology in current natural language processing research. In: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers), pp. 670–674 (2021)
Sen, I., Assenmacher, D., Samory, M., Augenstein, I., van der Aalst, W., Wagne, C.: People make better edits: measuring the efficacy of LLM-generated counterfactually augmented data for harmful language detection. arXiv preprint arXiv:2311.01270 (2023)
Sen, I., Samory, M., Flöck, F., Wagner, C., Augenstein, I.: How does counterfactually augmented data impact models for social computing constructs? arXiv preprint arXiv:2109.07022 (2021)
Sen, I., Samory, M., Wagner, C., Augenstein, I.: Counterfactually augmented data and unintended bias: the case of sexism and hate speech detection. In: Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 4716–4726 (2022)
Toraman, C., Şahinuç, F., Yilmaz, E.: Large-scale hate speech detection with cross-domain transfer. In: Proceedings of the Thirteenth Language Resources and Evaluation Conference, pp. 2215–2225. European Language Resources Association, Marseille, France (2022)
Vidgen, B., Thrush, T., Waseem, Z., Kiela, D.: Learning from the worst: dynamically generated datasets to improve online hate detection. arXiv preprint arXiv:2012.15761 (2020)
Wu, T., Ribeiro, M.T., Heer, J., Weld, D.S.: Polyjuice: generating counterfactuals for explaining, evaluating, and improving models. In: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pp. 6707–6723 (2021)
Acknowledgments
This publication has emanated from research conducted with the financial support of Science Foundation Ireland under grant no. 13/RC/2106_P2 at the ADAPT SFI Research Centre at Technological University Dublin.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Qureshi, M.A., Younus, A., Caton, S. (2024). Inclusive Counterfactual Generation: Leveraging LLMs in Identifying Online Hate. In: Stefanidis, K., Systä, K., Matera, M., Heil, S., Kondylakis, H., Quintarelli, E. (eds) Web Engineering. ICWE 2024. Lecture Notes in Computer Science, vol 14629. Springer, Cham. https://doi.org/10.1007/978-3-031-62362-2_3
Download citation
DOI: https://doi.org/10.1007/978-3-031-62362-2_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-62361-5
Online ISBN: 978-3-031-62362-2
eBook Packages: Computer ScienceComputer Science (R0)