Bias Detection and Optimization of Recruitment Texts Based on BERT
Main Article Content
Abstract
This study proposes an end-to-end framework for automatically detecting and mitigating implicit social bias in recruitment texts, providing an effective technical solution for fair and intelligent information processing in digital recruitment systems. As trustworthy semantic analysis and automated decision support become increasingly important for intelligent communication environments and information transmission applications, improving the neutrality and interpretability of textual content is also valuable for broader engineering-oriented information systems. Based on BERT, a bias detection model integrating BiLSTM and a hierarchical attention mechanism is developed to achieve fine-grained bias classification through word-level and sentence-level attention analysis. Furthermore, a T5-based bias optimization model is fine-tuned using biased–unbiased parallel corpora to accomplish semantic-preserving text rewriting. Experimental results demonstrate that the proposed detection model achieves an accuracy of 0.882 and an F1 score of 0.842 on the test set, outperforming the strongest baseline DeBERTa-v3 by 0.021 in F1 score. Ablation studies further verify that the BiLSTM and hierarchical attention modules improve the F1 score by 0.023 and 0.017, respectively. The optimization model obtains a neutrality score of 4.7 and an information retention score of 4.5 in human evaluation, while achieving BLEU and ROUGE-L scores of 0.65 and 0.71, respectively. The proposed framework provides a reliable and interpretable approach for automated fairness correction of recruitment texts and offers practical guidance for intelligent text processing and trustworthy information management in engineering applications.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
References
T. Sachs, A. C. Homan, and B. Lancee, “Impression formation of majority and minority applicants during resume screening—Does processing more information reduce prejudice? Journal of Applied Social Psychology,” 2024; 54(7):387-404, doi: 10.1111/jasp.13047.
M. Adamovic, “Analyzing discrimination in recruitment: A guide and best practices for resume studies,” International Journal of Selection and Assessment, vol. 28, no. 4, pp. 445-464, 2020, doi: 10.1111/ijsa.12298.
S. Celik and M. V. Turker, “Can eye movements be a predictor of implicit attitudes? Discrimination against disadvantaged individuals during the recruitment process,” Istanbul Business Research, vol. 51, no. 2, pp. 459-489, 2022, doi: 10.26650/ibr.2022.51.837555.
S. Hennekam, J. Peterson, L. Tahssain-Gay, and J. P. Dumazert, “Recruitment discrimination: How organizations use social power to circumvent laws and regulations,” The International Journal of Human Resource Management, vol. 32, no. 10, pp. 2213-2241, 2021, doi: 10.1080/09585192.2019.1579251.
D. Hangartner, D. Kopp, and M. Siegenthaler, “Monitoring hiring discrimination through online recruitment platforms,” Nature, vol. 589, no. 7843, pp. 572-576, 2021, doi: 10.1038/s41586-020-03136-0.
A. Waling, A. Lyons, B. Alba, V. Minichiello, C. Barrett, M. Hughes, et al., “Recruiting stigmatised populations and managing negative commentary via social media: A case study of recruiting older LGBTI research participants in Australia,” International Journal of Social Research Methodology, vol. 25, no. 2, pp. 157-170, 2022, doi: 10.1080/13645579.2020.1863545.
R. Bhardwaj, N. Majumder, and S. Poria, “Investigating gender bias in bert,” Cognitive Computation, vol. 13, no. 4, pp. 1008-1018, 2021, doi: 10.1007/s12559-021-09881-2.
J. R. Minot, N. Cheney, M. Maier, D. C. Elbers, C. M. Danforth, and P. S. Dodds, “Interpretable bias mitigation for textual data: Reducing genderization in patient notes while maintaining classification performance,” ACM Transactions on Computing for Healthcare, vol. 3, no. 4, pp. 1-41, 2022, doi: 10.1145/3524887.
R. Qasim, W. H. Bangyal, M. A. Alqarni, and A. Ali Almazroi, “A fine-tuned BERT-based transfer learning approach for text classification,” Journal of Healthcare Engineering, vol. 2022, no. 1, Art. no. 3498123, 2022, doi: 10.1155/2022/3498123.
A. K. Nandanwar and J. Choudhary, “Contextual embeddings-based web page categorization using the fine-tune bert model,” Symmetry, vol. 15, no. 2, pp. 395, 2023, doi: 10.3390/sym15020395.
K. Liu, Y. Feng, L. Zhang, R. Wang, W. Wang, X. Yuan, et al., “An effective personality-based model for short text sentiment classification using BiLSTM and self-attention,” Electronics, vol. 12, no. 15, pp. 3274, 2023, doi: 10.3390/electronics12153274.
L. Enamoto, A. R. A. S. Santos, R. Maia, L. Weigang, and G. P. R. Filho, “Multi-label legal text classification with BiLSTM and attention,” International Journal of Computer Applications in Technology, vol. 68, no. 4, pp. 369-378, 2022, doi: 10.1504/IJCAT.2022.125186.
W. Wang, Y. Zhang, Y. Sui, Y. Wan, Z. Zhao, J. Wu, et al., “Reinforcement-learning-guided source code summarization using hierarchical attention,” IEEE Transactions on Software Engineering, vol. 48, no. 1, pp. 102-119, 2020, doi: 10.1109/TSE.2020.2979701.
M. W. B. D. Satya, A. Luthfiarta, and M. N. Althoff, “Comparative Analysis of T5 Model Performance for Indonesian Abstractive Text Summarization,” SISTEMASI, vol. 14, no. 3, pp. 1092-1106, 2025, doi: 10.32520/stmsi.v14i3.4884.
E. Daraghmi, L. Atwe, and A. Jaber, “A Comparative Study of PEGASUS, BART, and T5 for Text Summarization Across Diverse Datasets,” Future Internet, vol. 17, no. 9, pp. 389, 2025, doi: 10.3390/fi17090389.
C. Raffel, N. Shazeer, A. Roberts, K. Lee, S. Narang, M. Matena, et al., “Exploring the limits of transfer learning with a unified text-to-text transformer,” Journal of Machine Learning Research, vol. 21, no. 140, pp. 1-67, 2020, doi: 10.48550/arXiv.1910.10683.
H. Zhang, H. Song, S. Li, M. Zhou, and D. Song, “A survey of controllable text generation using transformer-based pre-trained language models,” ACM Computing Surveys, vol. 56, no. 3, pp. 1-37, 2023, doi: 10.1145/3617680.
R. Rao, S. Sharma, and N. Malik, “Automatic text summarization using transformer-based language models,” International Journal of System Assurance Engineering and Management, vol. 15, no. 6, pp. 2599-2605, 2024, doi: 10.1007/s13198-024-02280-4.
J. C. Timoneda and S. V. Vera, “BERT, RoBERTa, or DeBERTa? Comparing Performance Across Transformers Models in Political Science Text,” The Journal of Politics, vol. 87, no. 1, pp. 347-364, 2025, doi: 10.1086/730737.
W. Liao, B. Zeng, X. Yin, and P. Wei, “An improved aspect-category sentiment analysis model for text sentiment analysis based on RoBERTa,” Applied Intelligence, vol. 51, no. 6, pp. 3522-3533, 2021, doi: 10.1007/s10489-020-01964-1.
W. Safira, B. Prabaswara, A. Stevens Karnyoto, and B. Pardamean, “Leveraging albert for sentiment classification of long-form chatgpt reviews on twitter,” International Journal of Computing and Digital Systems, vol. 17, no. 1, pp. 1-12, 2024, doi: 10.12785/ijcds/1570999256.
M. La Quatra and L. Cagliero, “Bart-it: An efficient sequence-to-sequence model for italian text summarization,” Future Internet, vol. 15, no. 1, pp. 15, 2022, doi: 10.3390/fi15010015.
R. K. Dey and A. K. Das, “Modified term frequency-inverse document frequency based deep hybrid framework for sentiment analysis,” Multimedia Tools and Applications, vol. 82, no. 21, pp. 32967-32990, 2023, doi: 10.1007/s11042-023-14653-1.