Abstract
Pre-training large language models on vast amounts of web-scraped texts is a current trend in natural language processing. While the resulting models are capable of generating convincing text, they also reproduce harmful social biases. Several modern studies have demonstrated that societal biases have a significant impact on the outcome of information retrieval systems. Our study is among the recent studies aimed at developing methods for quantifying and mitigating bias in search results and applying them to retrieval and recommendation systems based on transformer-based language models. This paper explores expressions of bias in Arabic text generation. Analyses will be performed on samples produced by the generative model AraGPT2 (a GPT-2 fine-tuned for Arabic). An Arabic bias classifier (Regard Classifier) based on new transformer model AraBERT (a BERT fine-tuned for Arabic) will be used to captures the social Bias of an Arabic-generated sentence or text. For the development of this classifier, a dataset will be crowd-sourced, cleaned, and independently annotated. AraGPT2 will be used to generate more biased descriptions from the standard prompts. Our Bias Detection Model (BDM) will be based on the combination of the two transformers (AraGPT2-AraBERT) models. In addition to our quantitative evaluation study, we will also conduct a qualitative study to understand how our system would compare to others approaches where users try to find bias in Arabic texts generated using AraGPT2 model. Our proposed model has achieved very encouraging results by reaching an accuracy percentage of 81%.
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
References
Hern, A., Facebook translates’ good morning’into’attack them’, leading to arrest. The Guardian, 24 (2017)
Caliskan, A., Bryson, J.J., Narayanan, A.: Semantics derived automatically from language corpora contain human-like biases. Science 356(6334), 183–186 (2017)
Zhou, X: Challenges in automated debiasing for toxic language detection. University of Washington (2021)
Pasquale, F.: The Black Box Society: The Secret Algorithms that Control Money and Information. Harvard University Press (2015). https://doi.org/10.4159/harvard.9780674736061
Antoun, W., Baly, F., Hajj, H.: AraBERT: transformer-based model for Arabic language understanding. arXiv preprint arXiv:2003.00104 (2020)
Antoun, W., Baly, F., Hajj, H., Aragpt2: pre-trained transformer for Arabic language generation. arXiv preprint arXiv:2012.15520 (2020)
Sheng, E., Chang, K.W., Natarajan, P., Peng, N.: Societal biases in language generation: progress and challenges. arXiv preprint arXiv:2105.04054 (2021)
Brown, T., et al.: Language models are few-shot learners. Adv. Neural. Inf. Process. Syst. 33, 1877–1901 (2020)
Liang, P.P., Wu, C., Morency, L.P., Salakhutdinov, R.: Towards understanding and mitigating social biases in language models. In: International Conference on Machine Learning, pp. 6565–6576. PMLR (2021)
Xu, A., Pathak, E., Wallace, E., Gururangan, S., Sap, M., Klein, D.: Detoxifying language models risks marginalizing minority voices. arXiv preprint arXiv:2104.06390 (2021)
Vaidya, A., Mai, F., Ning, Y.: Empirical analysis of multi-task learning for reducing identity bias in toxic comment detection. In: Proceedings of the International AAAI Conference on Web and Social Media, vol. 14, pp. 683–693 (2020)
Dathathri, S.: Plug and play language models: a simple approach to controlled text generation. arXiv preprint arXiv:1912.02164 (2019)
Kraft, A.: Triggering models: measuring and mitigating bias in German language generation, Doctoral dissertation, Master’s thesis, University of Hamburg (2021)
Gururangan, S., Marasović, A., Swayamdipta, S., Lo, K., Beltagy, I., Downey, D., Smith, N.A.: Don't stop pretraining: adapt language models to domains and tasks. arXiv preprint arXiv:2004.10964 (2020)
Schick, T., Udupa, S., Schütze, H.: Self-diagnosis and self-debiasing: a proposal for reducing corpus-based bias in NLP. Trans. Assoc. Comput. Linguist. 9, 1408–1424 (2021)
SMOTE for Imbalanced Classification with Python - MachineLearningMastery.com. Accessed 09 Jan 2023
Papineni, K., Roukos, S., Ward, T., Zhu, W.J.: BLEU: a method for automatic evaluation of machine translation. In: Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pp. 311–318 (2002)
Lin, C.Y.: ROUGE: a package for automatic evaluation of summaries. In: Text Summarization Branches Out, pp. 74–81 (2004)
Wang, J., Xia, B.: Relationships of Cohen's Kappa, sensitivity, and specificity for unbiased annotations. In: Proceedings of the 4th International Conference on Biomedical Signal and Image Processing (ICBIP 2019), pp. 98–101 (2019)
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
Harrag, F., Mahdadi, C., Ziad, A.N. (2023). Detecting and Measuring Social Bias of Arabic Generative Models in the Context of Search and Recommendation. In: Boratto, L., Faralli, S., Marras, M., Stilo, G. (eds) Advances in Bias and Fairness in Information Retrieval. BIAS 2023. Communications in Computer and Information Science, vol 1840. Springer, Cham. https://doi.org/10.1007/978-3-031-37249-0_13
Download citation
DOI: https://doi.org/10.1007/978-3-031-37249-0_13
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-37248-3
Online ISBN: 978-3-031-37249-0
eBook Packages: Computer ScienceComputer Science (R0)