Concept-Based Selective Passage Indexing for Off-the-Shelf Neural Ranking in Brazilian Portuguese Legal Retrieval
Resumo
This paper investigates an alternative to domain-specific fine-tuning for neural ranking models in information retrieval. We propose Concept-Based Selective Passage Indexing, a pre-encoding intervention in which each document is partitioned into sliding-window passages, and only those whose number of distinct human-curated concepts exceeds a calibrated threshold are retained for encoding. We evaluate this strategy on a legal corpus for which a high-quality set of human-curated concepts was available. Specifically, we experiment with four off-the-shelf neural rankers on a corpus of 4,268 Brazilian Federal Court decisions paired with 280 real user queries (242 in the primary evaluation set) and expert-graded relevance judgments. Across our experiments, Concept-Based Selective Passage Indexing improves NDCG relative to the full-indexing baseline for all evaluated rankers.
Palavras-chave:
Legal Information Retrieval, Neural Ranking, Late Interaction, ColBERT, Zero-Shot Retrieval, Brazilian Portuguese, Low-Resource IR
Referências
Atuhurra, J., Shindo, H., Kamigaito, H., and Watanabe, T. (2024). Introducing syllable tokenization for low-resource languages: A case study with swahili. arXiv preprint arXiv:2406.15358.
Berthélémé, A. (2024). Colbertv1.0-bertugues-base-portuguese-mmarcopt. [link]. 0-BERTugues-base-portuguese-mmarcoPT.
Bonifacio, L., Jeronymo, V., Abonizio, H. Q., Campiotti, I., Fadaee, M., Lotufo, R., and Nogueira, R. (2021). mmarco: A multilingual version of the ms marco passage ranking dataset. arXiv preprint arXiv:2108.13897.
Chalkidis, I., Fergadiotis, M., Malakasiotis, P., Aletras, N., and Androutsopoulos, I. (2020). Legal-bert: The muppets straight out of law school. In Findings of the association for computational linguistics: EMNLP 2020.
Chari, A., MacAvaney, S., and Ounis, I. (2025). Improving low-resource retrieval effectiveness using zero-shot linguistic similarity transfer. In European Conference on Information Retrieval.
Dai, Z. and Callan, J. (2019). Deeper text understanding for ir with contextual neural language modeling. In Proceedings of the 42nd international ACM SIGIR conference on research and development in information retrieval, pages 985–988.
Guo, J., Cai, Y., Fan, Y., Sun, F., Zhang, R., and Cheng, X. (2022). Semantic models for the first-stage retrieval: A comprehensive review. ACM Transactions on Information Systems, 40.
Khattab, O. and Zaharia, M. (2020). Colbert: Efficient and effective passage search via contextualized late interaction over bert. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. Association for Computing Machinery.
Louis, A., Saxena, V. K., van Dijck, G., and Spanakis, G. (2025a). Colbert-xm: A modular multi-vector representation model for zero-shot multilingual information retrieval. In Proceedings of the 31st International Conference on Computational Linguistics.
Louis, A., van Dijck, G., and Spanakis, G. (2025b). Know when to fuse: Investigating non-english hybrid retrieval in the legal domain. In Proceedings of the 31st International Conference on Computational Linguistics.
Lu, W., Jiao, J., and Zhang, R. (2020). Twinbert: Distilling knowledge to twin-structured compressed bert models for large-scale retrieval. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management.
Mazza Zago, R. and Agnoletti dos Santos Pedotti, L. (2024). Bertugues: A novel bert transformer model pre-trained for brazilian portuguese. Semina: Ciências Exatas e Tecnológicas.
Mentzingen, H., António, N., and Bacao, F. (2023). Automation of legal precedents retrieval: Findings from a literature review. International Journal of Intelligent Systems.
Nguyen, C. M., Tran, H. A. P., Thai, P. T., Nguyen, H. D., and Nguyen, V. T. (2025a). Lightweight semantic search for low-resource languages: A case study in vietnamese information retrieval. In International Conference on Similarity Search and Applications. Springer.
Nguyen, T. N., Hai, N. L., Hieu, N. D., Nguyen, D. A., Van, L. N., Nguyen, T. H., and Dinh, S. (2025b). Improving Vietnamese-English cross-lingual retrieval for legal and general domains. In Chiruzzo, L., Ritter, A., and Wang, L., editors, Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers). Association for Computational Linguistics.
Nogueira, R. and Cho, K. (2019). Passage re-ranking with bert. arXiv preprint arXiv:1901.04085.
Nogueira, R. F., Yang, W., Cho, K., and Lin, J. (2019). Multi-stage document ranking with BERT. CoRR.
Souza, F., Nogueira, R., and Lotufo, R. (2020). BERTimbau: pretrained BERT models for Brazilian Portuguese. In 9th Brazilian Conference on Intelligent Systems, BRACIS, Rio Grande do Sul, Brazil, October 20-23 (to appear).
Su, W., Ai, Q., Wu, Y., Xie, A., Wang, C., Ma, Y., Li, H., Wu, Z., Liu, Y., and Zhang, M. (2025). Pre-training for legal case retrieval based on inter-case distinctions. ACM Transactions on Information Systems.
Thakur, N., Reimers, N., Rücklé, A., Srivastava, A., and Gurevych, I. (2021). Beir: A heterogenous benchmark for zero-shot evaluation of information retrieval models. arXiv preprint arXiv:2104.08663.
Tien, S. P., Doan, H. N., Dai, A. N., and Viet, S. D. (2024). Improving vietnamese legal document retrieval using synthetic data. arXiv preprint arXiv:2412.00657.
Vera, H. S., Dua, S., Zhang, B., Salz, D., Mullins, R., Panyam, S. R., Smoot, S., Naim, I., Zou, J., Chen, F., et al. (2025). Embeddinggemma: Powerful and lightweight text representations. arXiv preprint arXiv:2509.20354.
Yates, A., Nogueira, R., and Lin, J. (2021). Pretrained transformers for text ranking: Bert and beyond. In Proceedings of the 14th ACM International Conference on web search and data mining, pages 1154–1156.
Zhou, Y., Huang, H., and Wu, Z. (2023). Boosting legal case retrieval by query content selection with large language models. In Proceedings of the Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacific Region, SIGIR-AP ’23. Association for Computing Machinery.
Berthélémé, A. (2024). Colbertv1.0-bertugues-base-portuguese-mmarcopt. [link]. 0-BERTugues-base-portuguese-mmarcoPT.
Bonifacio, L., Jeronymo, V., Abonizio, H. Q., Campiotti, I., Fadaee, M., Lotufo, R., and Nogueira, R. (2021). mmarco: A multilingual version of the ms marco passage ranking dataset. arXiv preprint arXiv:2108.13897.
Chalkidis, I., Fergadiotis, M., Malakasiotis, P., Aletras, N., and Androutsopoulos, I. (2020). Legal-bert: The muppets straight out of law school. In Findings of the association for computational linguistics: EMNLP 2020.
Chari, A., MacAvaney, S., and Ounis, I. (2025). Improving low-resource retrieval effectiveness using zero-shot linguistic similarity transfer. In European Conference on Information Retrieval.
Dai, Z. and Callan, J. (2019). Deeper text understanding for ir with contextual neural language modeling. In Proceedings of the 42nd international ACM SIGIR conference on research and development in information retrieval, pages 985–988.
Guo, J., Cai, Y., Fan, Y., Sun, F., Zhang, R., and Cheng, X. (2022). Semantic models for the first-stage retrieval: A comprehensive review. ACM Transactions on Information Systems, 40.
Khattab, O. and Zaharia, M. (2020). Colbert: Efficient and effective passage search via contextualized late interaction over bert. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. Association for Computing Machinery.
Louis, A., Saxena, V. K., van Dijck, G., and Spanakis, G. (2025a). Colbert-xm: A modular multi-vector representation model for zero-shot multilingual information retrieval. In Proceedings of the 31st International Conference on Computational Linguistics.
Louis, A., van Dijck, G., and Spanakis, G. (2025b). Know when to fuse: Investigating non-english hybrid retrieval in the legal domain. In Proceedings of the 31st International Conference on Computational Linguistics.
Lu, W., Jiao, J., and Zhang, R. (2020). Twinbert: Distilling knowledge to twin-structured compressed bert models for large-scale retrieval. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management.
Mazza Zago, R. and Agnoletti dos Santos Pedotti, L. (2024). Bertugues: A novel bert transformer model pre-trained for brazilian portuguese. Semina: Ciências Exatas e Tecnológicas.
Mentzingen, H., António, N., and Bacao, F. (2023). Automation of legal precedents retrieval: Findings from a literature review. International Journal of Intelligent Systems.
Nguyen, C. M., Tran, H. A. P., Thai, P. T., Nguyen, H. D., and Nguyen, V. T. (2025a). Lightweight semantic search for low-resource languages: A case study in vietnamese information retrieval. In International Conference on Similarity Search and Applications. Springer.
Nguyen, T. N., Hai, N. L., Hieu, N. D., Nguyen, D. A., Van, L. N., Nguyen, T. H., and Dinh, S. (2025b). Improving Vietnamese-English cross-lingual retrieval for legal and general domains. In Chiruzzo, L., Ritter, A., and Wang, L., editors, Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers). Association for Computational Linguistics.
Nogueira, R. and Cho, K. (2019). Passage re-ranking with bert. arXiv preprint arXiv:1901.04085.
Nogueira, R. F., Yang, W., Cho, K., and Lin, J. (2019). Multi-stage document ranking with BERT. CoRR.
Souza, F., Nogueira, R., and Lotufo, R. (2020). BERTimbau: pretrained BERT models for Brazilian Portuguese. In 9th Brazilian Conference on Intelligent Systems, BRACIS, Rio Grande do Sul, Brazil, October 20-23 (to appear).
Su, W., Ai, Q., Wu, Y., Xie, A., Wang, C., Ma, Y., Li, H., Wu, Z., Liu, Y., and Zhang, M. (2025). Pre-training for legal case retrieval based on inter-case distinctions. ACM Transactions on Information Systems.
Thakur, N., Reimers, N., Rücklé, A., Srivastava, A., and Gurevych, I. (2021). Beir: A heterogenous benchmark for zero-shot evaluation of information retrieval models. arXiv preprint arXiv:2104.08663.
Tien, S. P., Doan, H. N., Dai, A. N., and Viet, S. D. (2024). Improving vietnamese legal document retrieval using synthetic data. arXiv preprint arXiv:2412.00657.
Vera, H. S., Dua, S., Zhang, B., Salz, D., Mullins, R., Panyam, S. R., Smoot, S., Naim, I., Zou, J., Chen, F., et al. (2025). Embeddinggemma: Powerful and lightweight text representations. arXiv preprint arXiv:2509.20354.
Yates, A., Nogueira, R., and Lin, J. (2021). Pretrained transformers for text ranking: Bert and beyond. In Proceedings of the 14th ACM International Conference on web search and data mining, pages 1154–1156.
Zhou, Y., Huang, H., and Wu, Z. (2023). Boosting legal case retrieval by query content selection with large language models. In Proceedings of the Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacific Region, SIGIR-AP ’23. Association for Computing Machinery.
Publicado
19/10/2026
Como Citar
PALHETA, Matheus dos Santos; MOURA, Edleno Silva de.
Concept-Based Selective Passage Indexing for Off-the-Shelf Neural Ranking in Brazilian Portuguese Legal Retrieval. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 23. , 2026, Cuiabá/MT.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 193-204.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2026.29738.
