针对巴西法律检索的多样性，提出混合训练提升模型泛化能力
Domain-Adaptive Dense Retrieval for Brazilian Legal Search
- 用法律数据与SQuAD-pt混合训练，平衡专业性与泛化性
- 混合模型在六数据集上平均NDCG@10提升至0.447，显著优于纯法律模型
- 适合需要应对多种检索类型（如问答）的法律AI应用
巴西法律检索涵盖判例、立法及问答式搜索，导致稠密检索器训练面临专业化与泛化性的权衡。本文基于Qwen3-Embedding-4B，比较三种训练方式：未微调的基础模型、仅使用法律数据训练的模型，以及结合法律数据与SQuAD-pt监督数据的混合模型。在JUÁ榜单五个法律数据集及额外的Quati葡萄牙语检索基准上评估。仅法律数据训练的模型在最专业任务中表现最佳；混合训练模型在保持法律数据强性能的同时，整体表现更优，平均NDCG@10从0.414升至0.447，MRR@10从0.586升至0.595，MAP@10从0.270升至0.308。在Quati上的提升尤为显著，混合模型明显优于纯法律模型。结果表明：纯法律训练适合深度专业化，混合训练则在多类型检索（尤其问答）中更具鲁棒性。两种适配模型已发布于Hugging Face。
原文摘要 · Abstract (English)
Brazilian legal retrieval is heterogeneous, covering case law, legislation, and question-based search. This makes training dense retrievers a trade-off between stronger domain specialization and broader robustness across retrieval types of search. In this paper, we explore this trade-off using three training setups based on Qwen3-Embedding-4B: a base model with no fine-tuning, a version trained only on legal data, and a mixed setup that combines legal data with SQuAD-pt supervised dataset. We evaluate these models on five legal datasets from the JUÁ leaderboard, along with Quati dataset as an extra Portuguese retrieval benchmark to test out-of-domain generalization. The legal-only model performs best on the most specialized legal tasks. The mixed setup keeps strong performance on legal data while offering a better overall balance, improving average NDCG@10 from 0.414 to 0.447, MRR@10 from 0.586 to 0.595, and MAP@10 from 0.270 to 0.308 across all six datasets. The biggest improvement appears on Quati, where the mixed model clearly outperforms the legal-only one. Overall, the results show that legal-only and mixed training lead to different strengths: the first is better for specialization, while the second is more robust across different types of search, especially question-based ones. Both adapted models are available on Hugging Face
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。