1.华中农业大学信息学院,武汉430070
2.果蔬园艺作物种质创新与利用全国重点实验室,武汉430070
3.农业智能技术教育部工程研究中心,武汉430070
刘琪,E-mail:2024317120031@webmail.hzau.edu.cn
徐娟,E-mail: xujuan@mail.hzau.edu.cn
魏小梅,E-mail:may@mail.hzau.edu.cn
收稿:2025-09-30,
纸质出版:2026-05-25
移动端阅览
刘琪,王皓冉,方婉茹,等.FruitQA-Sys:面向农业果蔬领域的专家问答系统[J].华中农业大学学报,2026,45(3):68‐76.
LIU Qi,WANG Haoran,FANG Wanru,et al.FruitQA-Sys: an expert question-answering system for fruit and vegetable domain[J].Journal of Huazhong Agricultural University,2026,45(03):68-76.
刘琪,王皓冉,方婉茹,等.FruitQA-Sys:面向农业果蔬领域的专家问答系统[J].华中农业大学学报,2026,45(3):68‐76. DOI: 10.13300/j.cnki.hnlkxb.2026.03.006.
LIU Qi,WANG Haoran,FANG Wanru,et al.FruitQA-Sys: an expert question-answering system for fruit and vegetable domain[J].Journal of Huazhong Agricultural University,2026,45(03):68-76. DOI: 10.13300/j.cnki.hnlkxb.2026.03.006.
为解决现有大语言模型在农业果蔬种植等专业细分场景中知识覆盖不足、难以满足实际智能化应用需求的问题,采用领域数据构建、参数高效微调与检索增强生成相结合的方法,提出了一种面向农业果蔬领域的专家问答系统FruitQA-Sys。该系统通过生成器-打分器模块构建大规模高质量果蔬问答数据集,基于LoRA策略对领域大模型FruitBOT进行微调,并结合引导式检索增强生成技术IntroRAG,实现果蔬专业知识的精准问答。结果显示,在自动评估中,FruitQA-Sys的BERTScore为0.76,较DeepSeek-V3提升1.3%,与Qwen-3.0持平; GLEU为0.37,较DeepSeek-V3和Qwen-3.0分别提升2.7%;在人工评估中,系统综合得分达到8.56,较DeepSeek-V3和Qwen-3.0分别提升11.5%和15.7%。结果表明,领域微调大模型与IntroRAG的协同机制能够有效提升果蔬种植场景下智能问答系统的专业性与准确性。
A domain-specific expert question-answering system,FruitQA-Sys,was proposed by integrating domain data construction,parameter-efficient fine-tuning,and retrieval-augmented generation to solve the problems of insufficient knowledge coverage and difficulty in meeting the practical requirements of intelligent applications in specialized sub-scenarios including the planting of fruit and vegetable with existing large language models.The proposed system constructed a large-scale and high-quality fruit and vegetable question-answer dataset through a generator-scorer module,finely tuned the domain-specific large model FruitBOT based on the LoRA strategy,and incorporated the guided retrieval-augmented generation technique IntroRAG to achieve accurate question and answer of professional knowledge of fruit and vegetable.The results show that in automatic evaluation,the FruitQA-Sys achieves a BERTScore of 0.76, which is a 1.3% improvement over the DeepSeek-V3, and is on par with another mainstream general-purpose large model Qwen-3.0; its GLEU score is 0.37, representing a 2.7% improvement over both DeepSeek-V3 and Qwen-3.0. The overall score in human evaluation reached 8.56, representing improvements of 11.5% and 15.7% over DeepSeek-V3 and Qwen-3.0, respectively.It is indicated that the collaborative mechanism of domain-specific fine-tuning and IntroRAG can effectively improve the professionalism and accuracy of intelligent question-answering systems in scenarios of planting fruit and vegetable.
赵瑞雪,杨晨雪,郑建华,等.农业智能知识服务研究现状及展望[J].智慧农业(中英文),2022,4(4):105-125. ZHAO R X,YANG C X,ZHENG J H,et al.Agricultural intelligent knowledge service:overview and future perspectives[J].Smart agriculture,2022,4(4):105-125 (in Chinese with English abstract) .
QIU X P,SUN T X,XU Y G,et al.Pre-trained models for natural language processing:a survey[J].Science China technological sciences,2020,63(10):1872-1897.
ZHANG Y Q,FAN Q J,CHEN X,et al.IPM-AgriGPT:a large language model for pest and disease management with a G-EA framework and agricultural contextual reasoning[J/OL].Mathematics,2025,13(4):566[2025-09-30].https://doi.org/10.3390/math13040566.
王婷,王娜,崔运鹏,等.基于人工智能大模型技术的果蔬农技知识智能问答系统[J].智慧农业(中英文),2023,5(4):105-116.WANG T,WANG N,CUI Y P,et al.Agricultural technology knowledge intelligent question-answering system based on large language model[J].Smart agriculture,2023,5(4):105-116 (in Chinese with English abstract) .
YANG S T,LIU Z H,MAYER W,et al.ShizishanGPT:an agricultural large language model integrating tools and resources[C]//Web Information Systems Engineering(WISE 2024). December 2-5,2024 ,Doha,Qatar.Cham:Springer,2024:284-298.
YANG B,ZHANG Y,FENG L,et al.AgriGPT: a large language model ecosystem for agriculture [PP/OL].arXiv,2023:2308.15085[2025-09-30].https://doi.org/10.48550/arXiv.2508.08632.
ZHAO X Y,CHEN B Y,JI M X,et al.Implementation of large language models and agricultural knowledge graphs for efficient plant disease detection[J/OL].Agriculture,2024,14(8):1359[2025-09-30].https://doi.org/10.3390/agriculture14081359.
NIRANJAN P Y,RAJPUROHIT V S,SANNAKKI S S.Question answering system for agriculture domain using machine learning techniques:literature survey and challenges[J/OL].International journal of computational systems engineering,2020,6(2):91[2025-09-30].https://doi.org/10.1504/IJCSYSE.2020.111208.
郭旺,杨雨森,吴华瑞,等.农业大模型:关键技术、应用分析与发展方向[J].智慧农业(中英文),2024,6(2):1-13.GUO W,YANG Y S,WU H R,et al.Big models in agriculture:key technologies,application and future directions[J].Smart agriculture,2024,6(2):1-13 (in Chinese with English abstract) .
张云慧,张智,刘杰.病虫测报智能化研究进展[J].现代农药,2023,22(5):9-16.ZHANG Y H,ZHANG Z,LIU J.Research progress of intelligent in forecasting for diseases and insect pests[J].Modern agrochemicals,2023,22(5):9-16 (in Chinese with English abstract) .
WANG Y Z,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. July 7-9,2023 ,Toronto,Canada.Toronto:ACL,2023:13484-13508.
GUO D,YANG D,ZHANG H,et al.Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning [PP/OL].arXiv,2025:2501.12948[2025-09-30].https://doi.org/10.1038/s41586-025-09422-z.
PURI R,SPRING R,SHOEYBI M,et al.Training question answering models from synthetic data[PP/OL].arXiv:2002.09599[2025-09-30].https://doi.org/10.48550/arXiv.2002.095996.
HU E J,SHEN Y L,WALLIS P,et al.LoRA:low-rank adaptation of large language models[PP/OL].arXiv,2021:2106.09685[2025-09-30].https://doi.org/10.48550/arXiv.2106.09685.
刘铮,朱轶.农业大语言模型的部署及应用开发探索[J].中国农业文摘-农业工程,2025(6): 60-64.LIU Z,ZHU Y.Exploration on deployment and application development of agricultural large language models[J].China agricultural abstracts-agricultural engineering,2025(6): 60-64 (in Chinese) .
BIDERMAN D,PORTES J,ORTIZ J J G,et al.LoRA learns less and forgets less[PP/OL].arXiv,2024:2405.09673[2025-09-30].https://doi.org/10.48550/arXiv.2405.09673.
张钦彤,王昱超,王鹤羲,等.大语言模型微调技术的研究综述[J].计算机工程与应用,2024,60(17):17-33.ZHANG Q T,WANG Y C,WANG H X,et al.Comprehensive review of large language model fine-tuning[J].Computer engineering and applications,2024,60(17):17-33 (in Chinese with English abstract) .
JI Z W,LEE N,FRIESKE R,et al.Survey of hallucination in natural language generation[J].ACM computing surveys,2023,55(12):1-38.
LEWIS P,PEREZ E,PIKTUS A,et al.Retrieval-augmented generation for knowledge-intensive NLP tasks[PP/OL].arXiv,2005.11401[2025-09-30].https://doi.org/10.48550/arXiv.2005.11401.
JIANG J C,YAN L,LIU H F,et al.Knowledge assimilation:Implementing knowledge-guided agricultural large language model[J/OL].Knowledge-based systems,2025,314:113197[2025-09-30].https://doi.org/10.1016/j.knosys.2025.113197.
REIMERS N,GUREVYCH I.Sentence-BERT:sentence embeddings using Siamese BERT-networks[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). November 3-7,2019 ,Hong Kong,China.Hong Kong:ACL,2019:3980-3990.
KARPUKHIN V,OGUZ B,MIN S,et al.Dense passage retrieval for open-domain question answering[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). November 16-20,2020 ,Stroudsburg,PA,USA.Stroudsburg:ACL,2020:6769-6781.
于合龙,沈金梦,毕春光,等.基于知识图谱的水稻病虫害智能诊断系统[J].华南农业大学学报,2021,42(5):105-116.YU H L,SHEN J M,BI C G,et al.Intelligent diagnostic system for rice diseases and pests based on knowledge graph[J].Journal of South China Agricultural University,2021,42(5):105-116 (in Chinese with English abstract) .
WANG C X,CHENG S R,GUO Q P,et al.Evaluating open-QA evaluation[PP/OL].arXiv,2023:2305.12421[2025-09-30].https://doi.org/10.48550/arXiv.2305.12421.
ZHAO J M,XU L X,MA Z Z,et al.A review of plant leaf disease identification by deep learning algorithms[J/OL].Frontiers in plant science,2025,16:1637241[2025-09-30].https://doi.org/10.3389/fpls.2025.1637241.
符鹏,杨海平.大语言模型能力评价体系构建及问题研究[J].情报探索,2024(11):34-40.FU P,YANG H P.Construction and investigation of evaluation system for large language model capabilities[J].Information research,2024(11):34-40 (in Chinese with English abstract)
KAUR R,BHANKHAR A S,SALH J S,et al.Leveraging synthetic data for question answering with multilingual LLMs in the agricultural domain[PP/OL].arXiv,2025:2507.16974[2025-09-30].https://doi.org/10.48550/arXiv.2507.16974.
ZHANG T Y,KISHORE V,WU F,et al.BERTScore:evaluating text generation with BERT[PP/OL].arXiv, 2019 :1904.09675[2025-09-30].https://doi.org/10.48550/arXiv.1904.09675.
MUTTON A,DRAS M,WAN S,et al.GLEU:automatic evaluation of sentence-level fluency[C]//Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics.June 25-27,2007.Prague,Czech Republic.Prague: ACL,2007:344-351.
LIN C Y.Rouge: a package for automatic evaluation of summaries[C]//Proceedings of the Text Summarization Branches Out. July 25-26,2004 ,Barcelona,Spain.Barcelona:ACL,2004:74-81.
0
浏览量
3
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621