Research on Question Answering System of Oral History Resources Based on Knowledge Graph and Large Language Models

Abstract

Abstract:

Humanities institutions such as archives and libraries have gradually formed a rich and diverse collection of organized oral historical archives. By introducing a question answering system can display the reasoning ability of knowledge graphs through interactions. This study integrates knowledge graph and large language models, fully leveraging the advantages of accuracy and transparency of knowledge graph and reducing the problem of response inaccuracies and high development costs often associated with large language models, and attempting to construct a question answering system for oral historical archives resources. The article elaborates on the system design concept, construction process, and the core technical features. It takes the oral history of the CUSPEA theme housed in the T.D.Lee Library as the research object and conducts practical application. The practice has verified the feasibility of the question answering system, showcasing its ability to integrate and mine knowledge from oral history archives. The system effectively assists humanities scholars and history enthusiasts in understanding and gaining deeper insights into the essence of oral history.

Sun Yi, Liu Yin (Shanghai Jiao Tong University Library). Research on Question Answering System of Oral History Resources Based on Knowledge Graph and Large Language Models[J]. Libraly Journal, 2025, 44(405): 98-107.

References

[ 1 ] 顾静，丁华东. 口述历史档案社会记忆再生产的基本特点与思考[J]. 北京档案，2023（6）： 19–23.

[ 2 ] 高冕. 中美英国家图书馆记忆工程中的口述历史资源建设比较研究[J]. 图书馆学研究，2020（23）： 14–22.

[ 3 ] 毛海帆，李鹏达，田丹华，等.《录音录像类电子档案元数据方案》解读[J]. 中国档案，2018（5）： 34–35.

[ 4 ] 口述史料采集与管理规范[EB/OL]. [2024- 05-10]. https://www.saac.gov.cn/daj/hybz/ 201806/f5622dbc469340799aaccdd7e42a0150/ files/6b5376b5788b4ce680d08d1cb1b8946d.pdf.

[ 5 ] 录音录像档案管理规范[EB/OL]. [2024-05-10]. https://www.saac.gov.cn/daj/hybz/201912/5b3adc 136d6c4428bca60517658ac8e9.shtml.

[ 6 ] 闫静. 档案与记忆：中共党史研究的两个维度[J]. 档案学通讯，2021，259（3）：12–17.

[ 7 ] 邓君，王阮. 口述历史档案资源知识图谱与多维知识发现研究[J]. 图书情报工作，2022，66（7）： 4–16.

[ 8 ] 闫悦，郭晓然，王铁君，等. 问答系统研究综述[J]. 计算机系统应用，2023，32（8）：1–18.

[ 9 ] MIT. START natural language question answering system [EB/OL]. [2024-05-10]. http://start.csail. mit.edu.

[10] Singhal A. Introducing the knowledge graph： Things，not strings [EB/OL]. [2024-05-01]. https://blog.google/products/search/introducing-knowledge-graph-things-not/.

[11] 中国开放知识图谱[EB/OL]. [2024-05-15]. http:// openkg.cn/.

[12] 夏润泽，李丕绩.ChatGPT大模型技术发展与应用[J]. 数据采集与处理，2023，38（5）：1017–1034.

[13] Lewis P，Perez E，Piktus A，et al. Retrieval- Augmented Generation for Knowledge-Intensive NLP Tasks[J]. Advances in Neural Information Processing Systems，2020，33： 9459–9474.

[14] 王寰，孙雷，吴斌，等. 基于阅读理解智能问答的RPR 融合模型研究[J]. 计算机应用研究， 2022，39（3）：726–731；738.

[15] 萨日娜，李艳玲，林民. 知识图谱推理问答研究综述[J]. 计算机科学与探索，2022，16（8）： 1727–1741.

[16] Wei J，Tay Y，Bommasani R，et al. Emergent abilities of large language models[J]. arXiv： 2206.07682，2022.

[17] 张鹤译，王鑫，韩立帆，等. 大语言模型融合知识图谱的问答系统研究[J]. 计算机科学与探索， 2023，17（10）：2377–2388.

[18] Stephen Gilbert，Jakob Nikolas Kather，Aidan Hogan. Augmented Non-Hallucinating Large Language Models as Medical Information Curators [J]. NPJ Digital Medicine，2024，7（4）：1–5.

[19] Wang H，Liu C，Xi N，et al. HuaTuo： tuning LLaMA model with Chinese medical knowledge[J]. ar Xiv：2304.06975，2023.

[20] 杨喆，许甜，靳哲，等. 基于知识图谱的羊群疾病问答系统的构建与实现[J]. 华中农业大学学报，2023，42（3）：63–70.

[21] Zhentao Xu，Mark Jerome Cruz，Matthew Guevara，et al. Retrieval-Augmented Generation with Knowledge Graphs for Customer Service Question Answering [EB/OL]. [2024-05-15]. https://arxiv.org/abs/2404.17723.

[22] LangChain [EB/OL]. [2024-05-10]. https://python. langchain.com/.

[23] Yi Sun，Wanru Yang，Yin Liu. The Application of Constructing Knowledge Graph of Oral History Archive Resources Based on LLM-RAG [C] // 2024 the 8th International Conference on Information System and Data Mining，ICISDM 2024，2024： 1–10.

[24] Abrams M H，Harpham G. A glossary of literary terms[M]. Boston：Cengage Learning，2011： 50–52.

[25] 宋宁远，王晓光. 基于情节本体的叙事性文本语义结构化表示方法研究[J]. 中国图书馆学报， 2020，46（2）：96–113.

[26] 曾蕾，王晓光，范炜. 图档博领域的智慧数据及其在数字人文研究中的角色[J]. 中国图书馆学报，2018，44（1）：17–34.

[27] Van Hage W R，Malaisé V，Segers R，et al. Design and Use of the Simple Event Model （SEM）[J]. Journal of Web Semantics，2011，9（2）： 128–136.

[28] Gottschalk S，Demidova E. EventKg： A Multilingual Event-Centric Temporal Knowledge Graph[C]//European Semantic Web Conference. Springer，Cham，2018： 272–287.

[29] Corda I，Bennett B，Dimitrova V. A Logical Model of an Event Ontology for Exploring Connections in Historical Domains[C]// Workshop on Detection， Representation and Exploitation of Events in Semantic Web （Derive 2011），Tenth International Semantic Web Conference （ISWC’11）. 2011： 22–31.

[30] 刘宁静，孙翌，刘音，等. 基于口述历史资源的名人历史事件语义模型构建及实证研究[J/OL]. 现代情报：1–12[2024-05-13]. http://kns.cnki.net/ kcms/detail/22.1182.g3.20240402.1528.002.html.

[31] Tomaz Bratanic. Using a Knowledge Graph to Implement a RAG Application [EB/OL]. [2024-03- 01]. https://neo4j.com/developer-blog/knowledge-graph-rag-application/.