| 6 | 0 | 39 |
| 下载次数 | 被引频次 | 阅读次数 |
针对现有蕴含跨领域关联训练数据稀缺和单一文本向量难以满足跨领域语义检索的问题，本文提出一种基于大语言模型（Large Language Model,LLM）的跨领域关联数据生成方法和支持跨领域知识迁移的并行语义检索模型（Parallel Semantic Retrieval Model,PSRD）。通过提示词引导LLM生成跨领域关联标注样本，并利用bge-base-zh模型计算语义相似度并筛选关联标注样本，使用跨领域关联数据生成方法生成的数据集训练PSRD模型。根据在自建数据集上进行对比实验和消融实验，验证了本文所提跨领域关联数据生成方法与PSRD模型在跨领域知识检索中的有效性。开发了一种计算机辅助跨领域知识迁移创新语义检索系统，通过实例展示了系统的实用性。
Abstract：In response to the scarcity of existing training data with cross-domain associations and the difficulty of single text vectors in meeting the requirements of cross-domain semantic retrieval, this paper proposes a cross-domain association data generation method based on Large Language Model(LLM) and a parallel semantic retrieval model Parallel Semantic Retrieval Model(PSRD) that supports cross-domain knowledge transfer. By guiding LLM with prompt words to generate cross-domain association labeled samples and using the bge-base-zh model to filter the association labeled samples, the dataset generated by the cross-domain association data generation method is used to train the PSRD model. Through comparative experiments and ablation experiments on a self-built dataset, the effectiveness of the cross-domain association data generation method and the PSRD model in cross-domain knowledge retrieval is verified. A computer-aided cross-domain knowledge transfer innovative semantic retrieval system is developed, and its practicality is demonstrated through examples.
[1]姚佳，马超，秦旭东，等.基于TRIZ理论的新型卧式活套钢绳缩紧装置创新设计[J].机械，2019,46(1):56-59.
[2]石钎，熊艳，李彦，等.面向类比思维的创新问题求解过程[J].机械设计与制造，2018(12):5-9.
[3]郑浩，冯毅雄，高一聪，等.基于性能演化的复杂产品概念设计求解过程研究[J].机械工程学报，2018,54(9):214-223.
[4]那惠珍，李彦，熊艳，等.面向产品创新设计的个性化知识推送研究[J].机械设计与制造，2016(11):261-264.
[5]王春雨，蒋祖华，王福华，等.面向工业软件开发的半结构化知识语义检索方法[J].计算机集成制造系统，2021,27(8):2371-2381.
[6]FENG S Y,GANGAL V,WEI J,et al. A survey of data augmentation approaches for NLP:Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the11th International Joint Conference on Natural Language Processing[C].Stroudsburg:Association for Computational Linguistics,2021.
[7]VESELOVSKY V,RIBEIRO M H,ARORA A,et al. Generating faithful synthetic data with large language models:a case study in computational social science[PP/OL]. ar Xiv(2023-05-24)[2026-04-29].https://arxiv.org/abs/2305.15041.
[8]KIM Y. Convolutional neural networks for sentence classification:Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing[C]. Doha:Association for Computational Linguistics,2014.
[9]JOHNSONR,ZHANG T. Deep learning for semantic textual similarity[C]. Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics,2017:1369-1378.
[10]DEVLIN J,CHANG M W,LEE K,et al. BERT:Pre-training of deep bidirectional transformers for language understanding:Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics[C]. Minneapolis:Association for Computational Linguistics,2019.
[11]JOSHI M , CHEND , LIUY , et al. SpanBERT:Improving pre-training by representing and predicting spans[J]. Transactions of the Association for Computational Linguistics,2020(8):64-77.
[12]关威，曹健，赵海燕，等.基于大语言模型的业务流程语义异常检测方法[J/OL].计算机集成制造系统，2026:1-16[2026-03-26].https://doi.org/10.13196/j.cims.2025.BPM03.
[13]李煦，朱睿，陈小磊，等.视觉语言大模型的幻觉综述：成因、评估与治理[J].计算机研究与发展，2025,62(12):2929-2950.
[14]姚元璋，徐健.跨学科术语语义差异现象研究[J].数据分析与知识发现，2026,10(1):103-115.
[15]VASWANI A,SHAZEER N,PARMAR N,et al. Attention is all you need:Advances in Neural Information Processing Systems[C].Red Hook:Curran Associates,2017.
[16]侯平月.基于预训练模型的专利检索系统的研究与实现[D].哈尔滨：哈尔滨工业大学，2023.
基本信息:
中图分类号:TH122;TP18;TP391.3
引用信息:
[1]蔡虎,李文强,许国胜,等.面向创新设计的跨领域知识检索方法与系统[J].机械,2026,53(05):14-23+44.
基金信息:
国家自然科学基金（52075350）; 四川省重大科技专项（2022ZDZX0001）
2026-03-24
2026
2026-05-29
2026-05-29
2026
1
2026-05-15
2026-05-15