本体增强行业智能体的关键技术及应用
Technologies and Applications of Ontology-Enhanced Domain-Specific Agents
行业智能体是大语言模型与行业知识、数据系统及业务工具深度融合的产物,正由面向知识问答的辅助应用向面向复杂任务的执行系统演进。随着行业数据来源不断扩展,业务规则日益复杂,外部工具持续增加,仅依赖大语言模型难以维持任务理解、资源调用与结果判断之间的语义一致。本体增强通过形式化描述领域对象及其关系,并将业务条件、操作能力和运行状态纳入统一语义框架,可为行业智能体提供稳定的知识基础与可验证的任务依据。本文界定了本体增强行业智能体的基本内涵及适用范围;构建了相应的本体语义模型与分层技术体系;梳理了本体构建、知识获取、工具调用、任务规划及结果校验等关键技术;并以某大型制造企业的生产数据为对象开展应用验证。研究表明,本体增强能够通过显式表达业务对象、关系路径、指标口径、时间条件和状态规则,减少多约束查询中的语义偏移,为行业智能体连接异构业务数据提供可验证的语义支撑。
Domain-specific agents emerge from the deep integration of large language models (LLMs) with domain knowledge, data systems, and business tools. They are evolving from auxiliary applications for knowledge question answering into execution systems for complex tasks. As sources of domain data diversify, business rules grow more complex, and external tools proliferate, LLMs alone struggle to maintain semantic consistency across task understanding, resource invocation, and result assessment. Ontology enhancement formally represents domain objects and their relations and incorporates business conditions, operational capabilities, and runtime states into a unified semantic framework. It thus provides domain-specific agents with a stable knowledge foundation and a verifiable basis for task execution. This study first defines the concept and scope of ontology-enhanced domain-specific agents, develops an ontology semantic model and a layered technical framework, and reviews key technologies for ontology construction, knowledge acquisition, tool invocation, task planning, and result verification. It then conducts an application validation from a large manufacturing enterprise. The findings indicate that ontology enhancement can reduce semantic deviations in multi-constraint queries by explicitly representing business objects, relation paths, metric definitions, temporal conditions, and state rules, thereby providing verifiable semantic support for domain-specific agents accessing heterogeneous business data.
| [1] |
郑逸宁,余镇,李不凡, 大语言模型的工具使用综述[J]. 自动化学报,2025,51(11):2371-2386. |
| [2] |
Zheng Y N,Yu Z,Li B F,et al. Survey of tool use in large language models[J]. Acta Automatica Sinica,2025,51(11):2371-2386. |
| [3] |
Gruber T R. A translation approach to portable ontology specifications[J]. Knowledge Acquisition,1993,5(2):199-220. |
| [4] |
Pan S R,Luo L H,Wang Y F,et al. Unifying large language models and knowledge graphs:A roadmap[J]. IEEE Transactions on Knowledge and Data Engineering,2024,36(7):3580-3599. |
| [5] |
Lewis P,Perez E,Piktus A,et al. Retrieval-augmented generation for knowledge-intensive NLP tasks[C]//Advances in Neural Information Processing Systems 33,2020:9459-9474. |
| [6] |
Albertoni R,Browning D,Cox S J D,et al. Data catalog vocabulary (DCAT)—Version 3[EB/OL]. (2024-08-22)[2026-07-21]. https://www.w3.org/TR/2024/REC-vocab-dcat-3-20240822/. |
| [7] |
杜小勇,李曼,王珊. 本体学习研究综述[J]. 软件学报,2006,17(9):1837-1847. |
| [8] |
Du X Y,Li M,Wang S. A survey on ontology learning research[J]. Journal of Software,2006,17(9):1837-1847. |
| [9] |
任飞亮,沈继坤,孙宾宾, 从文本中构建领域本体技术综述[J]. 计算机学报,2019,42(3):654-676. |
| [10] |
Ren F L,Shen J K,Sun B B,et al. A review for domain ontology construction from text[J]. Chinese Journal of Computers,2019,42(3):654-676. |
| [11] |
Motik B,Patel-Schneider P F,Parsia B. OWL 2 web ontology language:Structural specification and functional-style syntax (second edition)[EB/OL]. (2012-12-11)[2026-07-21]. https://www.w3.org/TR/2012/REC-owl2-syntax-20121211/. |
| [12] |
Navigli R,Velardi P. Learning domain ontologies from document warehouses and dedicated web sites[J]. Computational Linguistics,2004,30(2):151-179. |
| [13] |
Cimiano P,Völker J. Text2Onto[M]. Berlin:Springer,2005:227-238. |
| [14] |
Velardi P,Faralli S,Navigli R. OntoLearn reloaded:A graph-based algorithm for taxonomy induction[J]. Computational Linguistics,2013,39(3):665-707. |
| [15] |
Demir C,Baci A,Kouagou N J,et al. Ontolearn—A framework for large-scale OWL class expression learning in Python[J]. Journal of Machine Learning Research,2025,26(63):1-6. |
| [16] |
Babaei Giglou H,D’Souza J,Auer S. LLMs4OL:Large language models for ontology learning[C]//The Semantic Web-ISWC 2023,2023:408-427. |
| [17] |
Jiménez-Ruiz E,Cuenca Grau B. LogMap:Logic-based and scalable ontology matching[C]//The Semantic Web-ISWC 2011,2011:273-288. |
| [18] |
He Y,Chen J Y,Antonyrajah D,et al. BERTMap:A BERT-based ontology alignment system[C]//The AAAI Conference on Artificial Intelligence,2022:4184-4191. |
| [19] |
Hertling S,Paulheim H. OLaLa:Ontology matching with large language models[C]//The 12th Knowledge Capture Conference 2023,2023:131-139. |
| [20] |
Karpukhin V,Oguz B,Min S,et al. Dense passage retrieval for open-domain question answering[C]//The 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP),2020:6769-6781. |
| [21] |
Wu L,Petroni F,Josifoski M,et al. Scalable zero-shot entity linking with dense entity retrieval[C]//The 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP),2020:6397-6407. |
| [22] |
Li B Z,Min S,Iyer S,et al. Efficient one-pass end-to-end entity linking for questions[C]//The 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP),2020:6433-6441. |
| [23] |
Sung M,Jeon H,Lee J,et al. Biomedical entity representations with synonym marginalization[C]//The 58th Annual Meeting of the Association for Computational Linguistics,2020:3641-3650. |
| [24] |
Liu F Y,Shareghi E,Meng Z Q,et al. Self-alignment pretraining for biomedical entity representations[C]//The 2021 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies,2021:4228-4238. |
| [25] |
Yahya M,Berberich K,Elbassuoni S,et al. Natural language questions for the web of data[C]//The 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning,2012:379-390. |
| [26] |
Dubey M,Banerjee D,Chaudhuri D,et al. EARL:Joint entity and relation linking for question answering over knowledge graphs[C]//The Semantic Web-ISWC 2018,2018:108-126. |
| [27] |
Guo J Q,Zhan Z C,Gao Y,et al. Towards complex text-to-SQL in cross-domain database with intermediate representation[C]//The 57th Annual Meeting of the Association for Computational Linguistics,2019:4524-4535. |
| [28] |
Wang B L,Shin R,Liu X D,et al. RAT-SQL:Relation-aware schema encoding and linking for text-to-SQL parsers[C]//The 58th Annual Meeting of the Association for Computational Linguistics,2020:7567-7578. |
| [29] |
Yih W T,Chang M W,He X D,et al. Semantic parsing via staged query graph generation:Question answering with knowledge base[C]//The 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing,2015:1321-1331. |
| [30] |
Massai L. Evaluation of semantic relations impact in query expansion-based retrieval systems[J]. Knowledge-Based Systems,2024,283:111183. |
| [31] |
Abe K,Takeoka K,Kato M P,et al. LLM-based query expansion fails for unfamiliar and ambiguous queries[C]//The 48th International ACM SIGIR Conference on Research and Development in Information Retrieval,2025:3035-3039. |
| [32] |
Xia Y,Wu J D,Kim S,et al. Knowledge-aware query expansion with large language models for textual and relational retrieval[C]//The 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics:Human Language Technologies,2025:4275-4286. |
| [33] |
Das S,Sundara S,Cyganiak R. R2RML:RDB to RDF mapping language[EB/OL]. (2012-09-27)[2026-07-21]. https://www.w3.org/TR/2012/REC-r2rml-20120927/. |
| [34] |
Calvanese D,De Giacomo G,Lembo D,et al. Tractable reasoning and efficient query answering in description logics:The DL-lite family[J]. Journal of Automated Reasoning,2007,39(3):385-429. |
| [35] |
W3C. OWL 2 web ontology language profiles:Second edition[EB/OL]. (2012-12-11)[2026-07-21]. https://www.w3.org/TR/2012/REC-owl2-profiles-20121211/. |
| [36] |
Calvanese D,Cogrel B,Komla-Ebri S,et al. Ontop:Answering SPARQL queries over relational databases[J]. Semantic Web,2016,8(3):471-487. |
| [37] |
Harris S,Seaborne A. SPARQL 1.1 Query language[EB/OL]. (2013-03-21)[2026-07-21]. https://www.w3.org/TR/2013/REC-sparql11-query-20130321/. |
| [38] |
Kharlamov E,Hovland D,Skjæveland M G,et al. Ontology based data access in statoil[J]. Journal of Web Semantics,2017,44:3-36. |
| [39] |
Kharlamov E,Mailis T,Mehdi G,et al. Semantic access to streaming and static data at Siemens[J]. Journal of Web Semantics,2017,44:54-74. |
| [40] |
Kalaycı E G,Grangel González I,Lösch F,et al. Semantic integration of Bosch manufacturing data using virtual knowledge graphs[C]//The Semantic Web—ISWC 2020,2020:464-481. |
| [41] |
Lebo T,Sahoo S,Mcguinness D. PROV-O:The PROV ontology[EB/OL]. (2013-04-30)[2026-07-21]. https://www.w3.org/TR/2013/REC-prov-o-20130430/. |
| [42] |
W3C. Data on the web best practices:Data quality vocabulary[EB/OL]. (2016-12-15)[2026-07-21]. https://www.w3.org/TR/2016/NOTE-vocab-dqv-20161215/. |
| [43] |
Iannella R,Villata S. ODRL information model 2.2[EB/OL]. (2018-02-15)[2026-07-21]. https://www.w3.org/TR/2018/REC-odrl-model-20180215/. |
| [44] |
Sharma K,Kumar P,Li Y Q. OG-RAG:Ontology-grounded retrieval-augmented generation for large language models[C]//The 2025 Conference on Empirical Methods in Natural Language Processing,2025:32962-32981. |
| [45] |
Schick T,Dwivedi-Yu J,Dessi R,et al. Toolformer:Language models can teach themselves to use tools[C]//Advances in Neural Information Processing Systems 36,2023:68539-68551. |
| [46] |
Qin Y J,Liang S H,Ye Y N,et al. ToolLLM:Facilitating large language models to master 16000+ real-world APIs[C]//The Twelfth International Conference on Learning Representations (ICLR),2024. |
| [47] |
Martin D,Burstein M,Hobbs J,et al. OWL-S:Semantic markup for web services[EB/OL]. (2004-11-22)[2026-07-21]. https://www.w3.org/submissions/2004/SUBM-OWL-S-20041122/. |
| [48] |
Farrell J,Lausen H. Semantic annotations for WSDL and XML schema[EB/OL]. (2007-08-28)[2026-07-21]. https://www.w3.org/TR/2007/REC-sawsdl-20070828/. |
| [49] |
Yuan S Y,Song K T,Chen J J,et al. EASYTOOL:Enhancing LLM-based agents with concise tool instruction[C]//The 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics:Human Language Technologies,2025:951-972. |
| [50] |
Shi Z L,Wang Y H,Yan L Y,et al. Retrieval models aren’t tool-savvy:Benchmarking tool retrieval for large language models[C]//The Association for Computational Linguistics:ACL 2025,2025:24497-24524. |
| [51] |
Klusch M,Fries B,Sycara K. OWLS-MX:A hybrid semantic web service matchmaker for OWL-S services[J]. Web Semantics:Science,Services and Agents on the World Wide Web,2009,7(2):121-133. |
| [52] |
Yao S Y,Zhao J,Yu D,et al. ReAct:Synergizing reasoning and acting in language models[C]//The Eleventh International Conference on Learning Representations (ICLR 2023),2023:1-21. |
| [53] |
Wu Q Z,Liu W,Luan J,et al. ToolPlanner:A tool augmented LLM for multi granularity instructions with path planning and feedback[C]//The 2024 Conference on Empirical Methods in Natural Language Processing,2024:18315-18339. |
| [54] |
Lu J R,Holleis T,Zhang Y Z,et al. ToolSandbox:A stateful,conversational,interactive evaluation benchmark for LLM tool use capabilities[C]//The Association for Computational Linguistics:NAACL 2025,2025:1160-1183. |
| [55] |
Sirin E,Parsia B,Wu D,et al. HTN planning for web service composition using SHOP2[J]. Journal of Web Semantics,2004,1(4):377-396. |
| [56] |
Hao Y L,Chen Y C,Zhang Y,et al. Large language models can solve real-world planning rigorously with formal verification tools[C]//The 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics:Human Language Technologies,2025:3434-3483. |
| [57] |
Tabassi E. Artificial intelligence risk management framework (AI RMF 1.0):NIST AI 100-1[R]. Gaithersburg:National Institute of Standards and Technology,2023:1-42. |
| [58] |
Glimm B,Horrocks I,Motik B,et al. HermiT:An OWL 2 reasoner[J]. Journal of Automated Reasoning,2014,53(3):245-269. |
| [59] |
Horrocks I,Patel-Schneider P F,Boley H,et al. SWRL:A semantic web rule language combining OWL and RuleML[EB/OL]. (2004-05-21)[2026-07-21]. https://www.w3.org/submissions/2004/SUBM-SWRL-20040521/. |
| [60] |
Knublauch H,Kontokostas D. Shapes constraint language (SHACL)[EB/OL]. (2017-07-20)[2026-07-21]. https://www.w3.org/TR/2017/REC-shacl-20170720/. |
| [61] |
Chytas A,Gavriilides G,Kapetanakis A,et al. OpenPVSignal knowledge graph:Pharmacovigilance signal reports in a computationally exploitable FAIR representation[J]. Drug Safety,2025,48(4):425-436. |
| [62] |
Arya V,Colace F,Lombardi M,et al. DTSafe-O:A semantic digital twin ontology for construction safety[J]. International Journal on Semantic Web and Information Systems,2025,21(1):1-36. |
| [63] |
Thorne J,Vlachos A,Christodoulopoulos C,et al. FEVER:A large-scale dataset for fact extraction and VERification[C]//The 2018 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies,2018:809-819. |
| [64] |
Min S,Krishna K,Lyu X X,et al. FActScore:Fine-grained atomic evaluation of factual precision in long form text generation[C]//The 2023 Conference on Empirical Methods in Natural Language Processing,2023:12076-12100. |
| [65] |
Gao T Y,Yen H,Yu J T,et al. Enabling large language models to generate text with citations[C]//The 2023 Conference on Empirical Methods in Natural Language Processing,2023:6465-6488. |
| [66] |
Wang X,Tan M J,Jin Q,et al. MedCite:Can language models generate verifiable text for medicine?[C]//The Association for Computational Linguistics:ACL 2025,2025:18891-18913. |
| [67] |
Gao L Y,Dai Z Y,Pasupat P,et al. RARR:Researching and revising what language models say,using language models[C]//The 61st Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers),2023:16477-16508. |
| [68] |
Es S,James J,Espinosa Anke L,et al. RAGAs:Automated evaluation of retrieval augmented generation[C]//The 18th Conference of the European Chapter of the Association for Computational Linguistics:System Demonstrations,2024:150-158. |
| [69] |
Schlegel M,Sattler K U. Capturing end-to-end provenance for machine learning pipelines[J]. Information Systems,2025,132:102495. |
中国工程院咨询项目“国家数据资源管理调度体系”(2024-JZ-03)
国家自然科学基金项目(62441230)
Funding project: Chinese Academy of Engineering project "National Data Resource Management and Scheduling System"(2024-JZ-03)
National Natural Science Foundation of China Project(62441230)
/
| 〈 |
|
〉 |