计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (9): 92-100.doi: 10.11896/jsjkx.250600103

• 大语言模型技术研究及应用 • 上一篇    下一篇

基于领域适配与动态检索增强的历史学科大模型构建方法

郭宇扬1, 史磊1, 刘欢1, 董怡翔1, 李睿2   

  1. 1 西安交通大学计算机科学与技术学院 西安 710049
    2 西安交通大学继续教育学院 西安 710049
  • 收稿日期:2025-06-16 修回日期:2025-11-02 出版日期:2026-09-15 发布日期:2026-09-10
  • 通讯作者: 刘欢(huanliu@xjtu.edu.cn)
  • 作者简介:(yuyangguo@stu.xjtu.edu.cn)
  • 基金资助:
    国家自然科学基金(62137002,62192781)

Domain-adapted and Dynamically Retrieval-augmented Approach for Large-scale History Discipline Model Construction

GUO Yuyang1, SHI Lei1, LIU Huan1, DONG Yixiang1, LI Rui2   

  1. 1 School of Computer Science and Technology,Xi'an Jiaotong University,Xi'an 710049,China
    2 School of Continuing Education,Xi'an Jiaotong University,Xi'an 710049,China
  • Received:2025-06-16 Revised:2025-11-02 Published:2026-09-15 Online:2026-09-10
  • About author:GUO Yuyang,born in 1998,postgraduate.His main research interests include natural language processing and the application of LLM in smart education.
    LIU Huan,born in 1990,associate professor,Ph.D supervisor, is a member of CCF(No.E9713M).His main research interests include LLM and smart education.
  • Supported by:
    National Natural Science Foundation of China(62137002,62192781).

PDF (PC)

314

摘要： 当前,教育信息化持续深入发展,人工智能技术为学科教育提供了新的方法论,其中大语言模型在构建知识体系与辅助学科任务方面展现出巨大潜力。然而,通用大模型在学科教育场景下面临知识边界模糊、事实性偏差显著、任务适配不足等核心问题。为此,聚焦历史学科,提出领域大模型构建及动态知识增强生成方法。该方法包含学科大模型构建与生成优化两个核心模块。首先,通过语义去重和上下文学习构建增量预训练数据集,提出基于指令微调的领域适配框架,采用跨段落因果关联建模融合文档全局语义与局部特征进行学科模型训练,并构建多维度学科模型评测体系;其次,提出动态决策检索增强生成机制,通过指令重构模块优化用户意图解析,设计多场景动态检索决策模块实现按需检索的智能路由,采用思维链引导的两阶段过滤机制筛选高价值知识。在历史学科任务上的实验结果表明,所提方法相较于通用大模型,在历史选择题准确率提升1.43%~10.19%,生成任务BLEU值提升1.02%~43.35%;相比传统检索增强方法,选择与生成任务分别提升3.54%~20.47%和0.18%~6.33%。该方法有效解决了知识偏差与低频知识覆盖难题,为教育领域大模型构建提供了动态知识增强的新范式。

关键词: 大语言模型, 学科大模型, 指令微调, 大模型评测, 检索增强生成

Abstract: Currently,the informatization of educatio continues to develop in depth,artificial intelligence technologies are providing new methodologies for subject-specific education.Among them,large language models(LLMs) demonstrate great potential in constructing knowledge systems and supporting discipline-specific tasks.However,general-purpose LLMs face core challenges in educational settings,including blurred knowledge boundaries,significant factual deviations,and inadequate task adaptation.To address these issues,this paper focuses on the history discipline and proposes a domain-specific LLM construction and dynamic knowledge-enhanced generation approach.The proposed method consists of two core modules:subject-specific model construction and generation optimization.Firstly,an incremental pre-training dataset is built through semantic deduplication and in-context learning.A domain adaptation framework based on instruction tuning is introduced,utilizing a cross-paragraph causal coherence modeling strategy to integrate global document semantics with local features for model training.A multi-dimensional evaluation system for subject-specific models is also developed.Secondly,a dynamic decision-based retrieval-enhanced generation mechanism is proposed.This includes an instruction reconstruction module for better user intent parsing,and a multi-scenario dynamic retrieval decision module for intelligent,on-demand routing.Finally,a chain-of-thought-guided two-stage filtering mechanism is employed to extract high-value knowledge.Experiments on history-related tasks demonstrate that,compared with general LLMs,the proposed method improves multiple-choice question accuracy by 1.43% to 10.19% and boosts BLEU scores in generative tasks by 1.02% to 43.35%.Compared with traditional retrieval-augmented approaches,it achieves gains of 3.54% to 20.47% in selection tasks and 0.18% to 6.33% in generation tasks.These results effectively mitigate knowledge bias and the challenge of cove-ring low-frequency knowledge,offering a new paradigm of dynamic knowledge enhancement for building domain-specific educational LLMs.

Key words: Large language models, Disciplinary large models, Instruction tuning, Large model evaluation, Retrieval-augmented generation

中图分类号: 

  • TP391
[1] WEI J,TAY Y,BOMMASANI R,et al.Emergent abilities oflarge language models[J].arXiv:2206.07682,2022.
[2] LEE J,YOON W,KIM S,et al.BioBERT:a pre-trained biomedical language representation model for biomedical text mining[J].Bioinformatics,2020,36(4):1234-1240.
[3] CHALKIDIS I,FERGADIOTIS M,MALAKASIOTIS P,et al.LEGAL-BERT:The muppets straight out of law school[C] //Findings of the Association for Computational Linguistics:EMNLP 2020.2020:2898-2904.
[4] BIANCHI F,HOVY D.On the gap between adoption and understanding in NLP[C] //Findings of the Association for Computational Linguistics:ACL-IJCNLP 2021.2021:3895-3901.
[5] LEWIS P,PEREZ E,PIKTUS A,et al.Retrieval-augmentedgeneration for knowledge-intensive nlp tasks[J].Advances in Neural Information Processing Systems,2020,33:9459-9474.
[6] ASAI A,WU Z,WANG Y,et al.Self-rag:Learning to retrieve,generate,and critique through self-reflection[C] //The Twelfth International Conference on Learning Representations.2023.
[7] FEI Z,SHEN X,ZHU D,et al.Lawbench:Benchmarking legal knowledge of large language models[C] //Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing.2024:7933-7962.
[8] HOU J,AO C,WU H,et al.E-eval:A comprehensive ChineseK-12 education evaluation benchmark for large language models[C] //Findings of the Association for Computational Linguistics(ACL 2024).2024:7753-7774.
[9] CHEN S F,GOODMAN J.An empirical study of smoothingtechniques for language modeling[J].Computer Speech & Language,1999,13(4):359-394.
[10] SCHMIDHUBER J,HOCHREITER S.Long short-term memory[J].Neural Computation,1997,9(8):1735-1780.
[11] DEVLIN J,CHANG M W,LEE K,et al.Bert:Pre-training of deep bidirectional transformers for language understanding[C] //Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies,Volume 1(long and short papers).2019:4171-4186.
[12] WEI J,BOSMA M,ZHAO V Y,et al.Finetuned language models are zero-shot learners[C] //International Conference on Learning Representations.2022.
[13] WANG Y,KORDI Y,MISHRA S,et al.Self-instruct:Aligning language models with self-generated instructions[C] //Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics(Volume 1:Long Papers).2023:13484-13508.
[14] OUYANG L,WU J,JIANG X,et al.Training language models to follow instructions with human feedback[J].Advances in Neural Information Processing Systems,2022,35:27730-27744.
[15] RADFORD A,WU J,CHILD R,et al.Language models are unsupervised multitask learners[J].OpenAI Blog,2019,1(8):9.
[16] BROWN T,MANN B,RYDER N,et al.Language models are few-shot learners[J].Advances in neural information processing systems,2020,33:1877-1901.
[17] LI X L,LIANG P.Prefix-tuning:Optimizing continuous prom-pts for generation[C] //Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing(Volume 1:Long Papers).2021:4582-4597.
[18] LESTER B,AL-RFOU R,CONSTANT N.The power of scale for parameter-efficient prompt tuning[C] //Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing.2021:3045-3059.
[19] DUAN X,TAN D,FANG L,et al.Reason-and-execute prompting:Enhancing multi-modal large language models for solving geometry questions[C] //Proceedings of the 32nd ACM International Conference on Multimedia.2024:6959-6968.
[20] YANG Z,QI P,ZHANG S,et al.HotpotQA:A dataset for diverse,explainable multi-hop question answering[C] //Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing.2018:2369-2380.
[21] CAI Y,WANG L,WANG Y,et al.Medbench:A large-scale chinese benchmark for evaluating medical large language models[C] //Proceedings of the AAAI Conference on Artificial Intelligence.2024:17709-17717.
[22] HEINZERLING B,INUI K.Language models as knowledge bases:On entity representations,storage capacity,and paraphrased queries[C] //Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics.2021:1772-1791.
[23] GUU K,LEE K,TUNG Z,et al.Retrieval augmented language model pre-training[C] //International Conference on Machine Learning.PMLR,2020:3929-3938.
[24] CHENG X,ZHANG Z,WANG J,et al.Education-OrientedGraph Retrieval-Augmented Generation for Learning Path Recommendation[J].arXiv:2506.22303,2025.
[25] IZACARD G,GRAVE E.Leveraging passage retrieval withgenerative models for open domain question answering[C] //Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics.2021.
[26] KHANDELWAL U,LEVY O,JURAFSKY D,et al.Generalization through memorization:Nearest neighbor language models[C] //International Conference on Learning Representations.2020.
[27] KRATHWOHL D R.A revision of Bloom’s taxonomy:An overview[J].Theory Into Practice,2002,41(4):212-218.
[28] HUANG Y,BAI Y,ZHU Z,et al.C-eval:A multi-level multi-discipline chinese evaluation suite for foundation models[J].Advances in Neural Information Processing Systems,2023,36:62991-63010.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发