基于大语言模型的《中国小麦品种志》信息提取

doi:10.11871/jfdc.issn.2096-742X.2025.01.013

数据与计算发展前沿 ›› 2025, Vol. 7 ›› Issue (1): 175-185.

CSTR: 32002.14.jfdc.CN10-1649/TP.2025.01.013

doi: 10.11871/jfdc.issn.2096-742X.2025.01.013

基于大语言模型的《中国小麦品种志》信息提取

韦一金^1,²(),陈彦清³,王秀东^4,^5,^*(),樊景超^1,²

1.中国农业科学院农业信息研究所，北京 100081
2.国家农业科学数据中心，北京 100081
3.中国农业科学院作物科学研究所，北京 100081
4.中国农业科学院农业经济与发展研究所，北京 100081
5.中国农业科学院战略研究中心，北京 100081

收稿日期:2024-05-23 出版日期:2025-02-20 发布日期:2025-02-21
通讯作者: *王秀东（E-mail: wangxiudong@caas.cn）
作者简介:韦一金，中国农业科学院农业信息研究所，硕士研究生，研究方向为农业信息技术。负责实验与论文撰写。
WEI Yijin, is a master’s student at the Institute of Agricultural Information, Chinese Academy of Agricultural Sciences (CAAS). Her research interests include agricultural information technology.
In this paper, she is responsible for conducting experiments and writing the manuscript.
E-mail: weiyijin0816@163.com|王秀东，中国农业科学院农业经济与发展研究所，博士，研究员，研究方向为粮食安全及农业发展战略。
负责论文审定，参与小麦种质信息指标构建。
WANG Xiudong, Ph.D., is a professor at the Institute of Agricultural Economics and Development, Chinese Academy of Agricultural Sciences (CAAS). His research interests include food security and agricultural development strategies.
In this paper, he is responsible for the finalization of papers and the construction of wheat germplasm information indicators.
E-mail: wangxiudong@caas.cn
基金资助:
中国农业科学院农业信息研究所科技创新工程(CAAS-ASTIP-2024-AII)

Information Extraction from Chinese Wheat Varieties Journal Based on Large Language Model

WEI Yijin^1,²(),CHEN Yanqing³,WANG Xiudong^4,^5,^*(),FAN Jingchao^1,²

1. Agriculture Information Institution of CAAS, Beijing 100081, China
2. National Agriculture Science Data Center, Beijing 100081, China
3. Institute of Crop Sciences, CAAS, Beijing 100081, China
4. Institution of Agricultural Economics and Development, CAAS, Beijing 100081, China
5. Center for Strategic Studies, CAAS, Beijing 100081, China

Received:2024-05-23 Online:2025-02-20 Published:2025-02-21

摘要/Abstract

摘要：

【目的】为促进小麦种质资源向小麦产业优势转化、提高小麦遗传背景丰富性，本文基于大语言模型（Large Language Model, LLM）和提示词工程，针对已出版的三卷《中国小麦品种志》进行信息挖掘。【方法】扫描《中国小麦品种志》纸质版文稿并进行OCR识别等数据处理工作以获取小麦品种数据，构建面向育种工作需求的小麦品种数据关键提取指标和相应的大语言模型提示词，以调用商业LLM api接口的方式对小麦品种数据的关键信息进行自动化提取，并形成一套成熟的基于大语言模型的小麦品种信息提取工作方案。【结果】以信息提取任务中的实际存在关系个数、识别出的关系个数、正确识别的关系个数进行精确率、召回率和F1值的计算，结果表明该小麦品种志信息提取方案在已出版的三卷《中国小麦品种志》信息提取中均达到了0.89以上的准确率、0.73以上的召回率和0.84以上的F1值。【结论】小麦品种志信息提取方案的高准确率表明其完全有能力实现精准信息提取，但是召回率又表明该方案存在部分信息无法识别的问题，因此虽然综合F1值而言该方案整体可行，但仍需对提取结果进行进一步的人工核验及审查。

关键词: 大语言模型, 农业, 小麦, 信息挖掘, 种质资源

Abstract:

[Objective] In order to promote the transformation of wheat germplasm resources to wheat industry advantages and to improve the richness of wheat genetic background, this paper presents a study of information mining from the published three-volume Chinese Wheat Variety Journal based on the Large Language Model (LLM) and cue word engineering. [Methods] This project involves scanning the paper version of the Chinese Wheat Variety Journal and performing OCR recognition and other data processing tasks to obtain wheat variety data. We aim to develop key extraction indices for wheat variety data and the corresponding prompt words of the LLMs for the needs of breeding work. By calling commercial LLM API interfaces, the key information of wheat variety data will be automatically extracted. The result will be a well-established workflow for extracting wheat variety information using large language models. [Results] The calculation of precision rate, recall rate, and F1 value in terms of the number of actually existing relations, the number of recognized relations, and the number of correctly recognized relations in the information extraction task show that this wheat varietal journal information extraction scheme achieved more than 0.89 precision rate, 0.73 recall rate, and 0.84 F1 value in the information extraction for the three volumes of Chinese Wheat Varietal Journal that have been published. [Conclusions] The high accuracy of this wheat varietal journal information extraction scheme indicates that it is fully capable of achieving precise information extraction, but the recall rate also indicates that the scheme has the problem that some information cannot be recognized. Though the scheme is overall feasible in terms of the combined F1 score, further manual verification and review of the extraction results is still required.

Key words: Large Language Model (LLM), agriculture, wheat, information mining, genetic resources

韦一金, 陈彦清, 王秀东, 樊景超. 基于大语言模型的《中国小麦品种志》信息提取[J]. 数据与计算发展前沿, 2025, 7(1): 175-185.

WEI Yijin, CHEN Yanqing, WANG Xiudong, FAN Jingchao. Information Extraction from Chinese Wheat Varieties Journal Based on Large Language Model[J]. Frontiers of Data and Computing, 2025, 7(1): 175-185, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2025.01.013.

图/表 8

图1

表1

图2

图3

表2

图4

表3

表4

参考文献 45

[1]	唐建卫, 高艳, 胡润雨, 等. 近十年基于重要小麦亲本周麦22的遗传改良[J]. 麦类作物学报, 2021, 41(3): 263-271.
[2]	郝晨阳, 董玉琛, 王兰芬, 等. 我国普通小麦核心种质的构建及遗传多样性分析[J]. 科学通报, 2008(8): 908-915.
[3]	相吉山, 杨欣明, 李秀全, 等. 小麦骨干亲本南大2419对衍生品种(系)HMW-GS的贡献分析[J]. 植物遗传资源报, 2013, 14(6): 1053-1058.
[4]	潘玉朋, 李立群, 郑锦娟, 等. 黄淮麦区近年大面积推广小麦品种的遗传多样性分析[J]. 西北农业学报, 2011, 20(4): 47-52.
[5]	SHANAHAN M. Talking about large language models[J]. Communications of the ACM, 2024, 67(2): 68-79.
[6]	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.
[7]	CHOWDHERY A, NARANG S, DEVLIN J, et al. Palm: Scaling language modeling with pathways[J]. Journal of Machine Learning Research, 2023, 24(240): 1-113.
[8]	TOUVRON H, LAVRIL T, IZACARD G, et al. Llama: Open and efficient foundation language models[J]. arXiv preprint arXiv:2302.13971, 2023.
[9]	QIN C, ZHANG A, ZHANG Z, et al. Is ChatGPT a general-purpose natural language processing task solver?[J]. arXiv preprint arXiv:2302.06476, 2023.
[10]	OpenAI R. Gpt-4 technical report. arxiv 2303.08774[J]. View in Article, 2023, 2: 3.
[11]	HAN X, ZHANG Z, DING N, et al. Pre-trained models: Past, present and future[J]. AI Open, 2021, 2: 225-250.
[12]	WEI J, TAY Y, BOMMASANI R, et al. Emergent abilities of large language models[J]. arXiv preprint arXiv:2206.07682, 2022.
[13]	HOFFMANN J, BORGEAUD S, MENSCH A, et al. Training compute-optimal large language models[J]. arXiv preprint arXiv:2203.15556, 2022.
[14]	BIAN J, ZHENG J, ZHANG Y, et al. Inspire the Large Language Model by External Knowledge on BioMedical Named Entity Recognition[J]. arXiv preprint arXiv:2309.12278, 2023.
[15]	POLAK M P, MORGAN D. Extracting accurate materials data from research papers with conversational language models and prompt engineering[J]. Nature Communications, 2024, 15(1): 1569.
[16]	GILARDI F, ALIZADEH M, KUBLI M. Chatgpt outperforms crowd-workers for text-annotation tasks[J]. arXiv preprint arXiv:2303.15056, 2023.
[17]	HUANG J, GU S S, HOU L, et al. Large language models can self-improve[J]. arXiv preprint arXiv:221 0.11610, 2022.
[18]	TANG R, HAN X, JIANG X, et al. Does synthetic data generation of llms help clinical text mining?[J]. arXiv preprint arXiv:2303.04360, 2023.
[19]	JEBLICK K, SCHACHTNER B, DEXL J, et al. ChatGPT makes medicine easy to swallow: an exploratory case study on simplified radiology reports[J]. European Radiology, 2023: 1-9.
[20]	CHOI J H, HICKMAN K E, MONAHAN A, et al. Chatgpt goes to law school[J]. Available at SSRN, 2023.
[21]	NOV O, SINGH N, MANN D M. Putting ChatGPT’s medical advice to the (Turing) test[J]. medRxiv, 2023: 2023.01. 23.23284735.
[22]	BLAIR-STANEK A, HOLZENBERGER N, VAN D URME B. Can GPT-3 perform statutory reasoning?[J]. arXiv preprint arXiv:2302.06100, 2023.
[23]	BANG Y, CAHYAWIJAYA S, LEE N, et al. A multitask, multilingual, multimodal evaluation of chatgpt on reasoning, hallucination, and interactivity[J]. arXiv preprint arXiv:2302.04023, 2023.
[24]	FLORIDI L, CHIRIATTI M. GPT-3: Its nature, scope, limits, and consequences[J]. Minds and Machines, 2020, 30: 681-694.
[25]	ACHIAM J, ADLER S, AGARWAL S, et al. Gpt-4 technical report[J]. arXiv preprint arXiv:2303.08774, 2023.
[26]	KURIBAYASHI T, OSEKI Y, TAIEB S B, et al. Large Language Models Are Human-Like Internally[J]. arXiv preprint arXiv:2502.01615, 2025.
[27]	RADFORD A, WU J, CHILD R, et al. Language models are unsupervised multitask learners[J]. OpenAI Blog, 2019, 1(8): 9.
[28]	TOUVRON H, LAVRIL T, IZACARD G, et al. Llama: Open and efficient foundation language models[J]. arXiv preprint arXiv:2302.13971, 2023.
[29]	TOUVRON H, MARTIN L, STONE K, et al. Llama 2: Open foundation and fine-tuned chat models[J]. arXiv preprint arXiv:2307.09288, 2023.
[30]	MIN B, ROSS H, SULEM E, et al. Recent advances in natural language processing via large pre-trained language models: A survey[J]. ACM Computing Surveys, 2023, 56(2): 1-40.
[31]	YOO K M, PARK D, KANG J, et al. Gpt3mix: Leveraging large-scale language models for text augmentation[J]. arXiv preprint arXiv:2104.08826, 2021.
[32]	ALBRECHT J, KITANIDIS E, FETTERMAN A J. Despite “super-human” performance, current LLMs are unsuited for decisions about ethics and safety[J]. arXiv preprint arXiv:2212.06295, 2022.
[33]	LIANG P P, WU C, MORENCY L P, et al. Towards understanding and mitigating social biases in language models[C]// International Conference on Machine Learning. PMLR, 2021: 6565-6576.
[34]	ALVI M, ZISSERMAN A, NELLÅKER C. Turning a blind eye: Explicit removal of biases and variation from deep neural network embeddings[C]// Proceedings of the European Conference on Computer Vision (ECCV) Workshops, 2018.
[35]	ZHANG J, VERMA V. Discover Discriminatory Bias in High Accuracy Models Embedded in Machine Learning Algorithms[C]// The International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery. Cham: Springer International Publishing, 2020: 1537-1545.
[36]	HAJIKHANI A, COLE C. A critical review of large language models: Sensitivity, bias, and the path toward specialized ai[J]. Quantitative Science Studies, 2024: 1-22.
[37]	GARTLEHNER G, KAHWATI L, HILSCHER R, et al. Data extraction for evidence synthesis using a large language model: A proof-of-concept study[J]. Research Synthesis Methods, 2024.
[38]	BELTAGY I, LO K, COHAN A. SciBERT: A pretrained language model for scientific text[J]. arXiv preprint arXiv:1903.10676, 2019.
[39]	LEWIS P, OTT M, DU J, et al. Pretrained language models for biomedical and clinical tasks: understanding and extending the state-of-the-art[C]// Proceedings of the 3rd clinical natural language processing workshop, 2020: 146-157.
[40]	LIN Z. How to write effective prompts for large language models[J]. Nature Human Behaviour, 2024: 1-5.
[41]	LIU P, YUAN W, FU J, et al. Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing[J]. ACM Computing Surveys, 2023, 55(9): 1-35.
[42]	DeepSeek-AI. DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model[J]. arxiv preprint arxiv:2405.04434, 2024.
[43]	DAI D, DENG C, ZHAO C, et al. Deepseekmoe: Towards ultimate expert specialization in mixture-of-experts language models[J]. arxiv preprint arxiv:2401. 06066, 2024.
[44]	AINSLIE J, LEE-THORP J, DE JONG M, et al. Gqa: Training generalized multi-query transformer models from multi-head checkpoints[J]. arxiv preprint arxiv:2305.13245, 2023.
[45]	SHAZEER N. Fast transformer decoding: One write-head is all you need[J]. arxiv preprint arxiv:1911.021 50, 2019.

一级指标	二级指标	三级指标
来源	名称	资源名
		其他名称
		属于哪个变种
	品种来源	国外引进（原产国；原产地；引进时间；引进单位；原始代号；原始父本、母本；首先引进到国内哪些地区；其他关于引进的描述）
		品种改良（父本；母本；育种单位；育成时间；原始代号；其他相关描述）
		地方品种（原始分布地区；曾收集改良过该品种的单位；原始编号；其他相关描述）
	特性特征	最突出的特性；被种植或推广的原因
产量与分布	区域试验	首次区试鉴定时间
		首次区试地点
		首次区试单位
		区试结果；主要优点
	审定	是否审定（审定单位；审定时间）
	产量	亩产（一般亩产；最高亩产记录；最高亩产时间；最高亩产地点）
		增产（最高增产记录；最高增产记录的对照品种；最高增产时间；最高增产地点）
		其他产量相关的描述
	种植/推广	首次推广时间
		种植/推广区域
		种植/推广面积
		种植/推广面积统计
		历年面积变化
		最大推广面积
		最大推广面积年份
		最大推广面积时主要区域
		种植推广中发现的问题
		品质特征
		熟期物候
		抗性
		其他相关描述

项目	具体配置
图形处理器（GPU）	RTXA6000-48G*4
操作系统	Windows 11 Pro
编程语言	Python

《中国小麦品种志》	提取小麦种质信息数量（种）
1949-1961卷	623
1962-1982卷	472
1983-1993卷	413

《中国小麦品种志》	Precision	Recall	F-measure
1949-1961卷	1.0	0.73	0.84
1962-1982版	0.99	0.84	0.91
1983-1993版	0.89	0.99	0.94

基于大语言模型的《中国小麦品种志》信息提取

Information Extraction from Chinese Wheat Varieties Journal Based on Large Language Model

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 45

相关文章 9

编辑推荐

Metrics

本文评价

[1]	王子健, 李凯, 曹荣强, 周纯葆. 基于检查点的大模型弹性训练方法研究[J]. 数据与计算发展前沿, 2025, 7(1): 135-151.
[2]	马秋平, 张琪, 赵晓凡. 图表问答研究综述[J]. 数据与计算发展前沿, 2025, 7(1): 19-37.
[3]	韦一金,樊景超. 基于改进的BERT-BiGRU-Attention的农业科技政策分类模型[J]. 数据与计算发展前沿, 2024, 6(6): 53-61.
[4]	裴炳森,李欣,蒋章涛,刘明帅. 基于大语言模型的司法文本摘要生成与评价技术研究[J]. 数据与计算发展前沿, 2024, 6(6): 62-73.
[5]	赵小丹, 胡林. 基于深度学习的农业科技政策知识抽取方法研究[J]. 数据与计算发展前沿, 2024, 6(4): 106-115.
[6]	韦一金, 樊景超. 基于ChatGLM2-6B的农业政策问答系统[J]. 数据与计算发展前沿, 2024, 6(4): 116-127.
[7]	孙坦, 赵瑞雪, 杨晓蓉, 王剑. 计算型科研的理念、内涵及农业科研发展的启示[J]. 数据与计算发展前沿, 2023, 5(4): 77-85.
[8]	王文生,郭雷风. 大数据技术农业应用[J]. 数据与计算发展前沿, 2020, 2(2): 101-110.
[9]	陈雷,袁媛. 基于深度迁移学习的农业病害图像识别[J]. 数据与计算发展前沿, 2020, 2(2): 111-119.