大模型驱动的矿产资源智能预测超级智能体构建方法探索

doi:10.13745/j.esf.sf.2025.7.1

地学前缘 ›› 2025, Vol. 32 ›› Issue (4): 38-45.DOI: 10.13745/j.esf.sf.2025.7.1

大模型驱动的矿产资源智能预测超级智能体构建方法探索

王永志¹^,²^,³^,^*(), 温世博², 李博文¹, 陈星宇², 董宇浩¹, 田江涛³, 王斌⁴^,⁵, Muhammed Atif BILAL², 纪政², 孙丰月⁶

1.吉林大学综合信息矿产预测研究所, 吉林长春 130061
2.吉林大学地球探测科学与技术学院, 吉林长春 130061
3.新疆维吾尔自治区地质研究院, 新疆乌鲁木齐 830057
4.自然资源部深部金矿勘查开采技术创新中心, 山东威海 264209
5.山东省地质矿产勘查开发局第六地质大队, 山东威海 264209
6.吉林大学地球科学学院, 吉林长春 130061

收稿日期:2024-12-10 修回日期:2025-05-26 出版日期:2025-07-25 发布日期:2025-08-04
通信作者: *王永志(1974—),男,博士,教授,主要从事地球科学大数据分析与挖掘、矿产资源智能预测等理论与应用研究工作。 E-mail: wangyongzhi@jlu.edu.cn
基金资助:
国家重点研发计划项目(2023YFC2906903);国家重点研发计划项目(2021YFC2901801);国家重点研发计划项目(2023YFC2907105);国家自然科学基金重点项目(42230810);自然资源部科技支撑项目(ZKKJ202419);山东省地质矿产局科技攻关项目(KY202502)

Construction technology of super-agents for intelligent mineral resources prediction driven by large model

WANG Yongzhi¹^,²^,³^,^*(), WEN Shibo², LI Bowen¹, CHEN Xingyu², DONG Yuhao¹, TIAN Jiangtao³, WANG Bin⁴^,⁵, Muhammed Atif BILAL², JI Zheng², SUN Fengyue⁶

1. Institute of Integrated Information for Mineral Resources Prediction, Jilin University, Changchun 130061, China
2. College of Geoexploration Science and Technology, Jilin University, Changchun 130061, China
3. Xinjiang Academy of Geological Research, ürümqi 830057, China
4. Ministry of Natural Resources Technology Innovation Center for Deep Gold Resources Exploration and Mining, Weihai 264209, China
5. No.6 Geological Team of Shandong Provincial Bureau of Geology and Mineral Resources, Weihai 264209, China
6. College of Earth Sciences, Jilin University, Changchun 130061, China

Received:2024-12-10 Revised:2025-05-26 Online:2025-07-25 Published:2025-08-04

摘要/Abstract

摘要：

矿产资源预测是数学地球科学领域的一项重要研究内容,需使用多种软件处理跨专业地学数据,面临处理过程复杂、工作量巨大、语义难对齐等诸多问题,给研究人员带来巨大挑战。随着新一代生成式人工智能的大模型、智能体等出现,极大地推动了各行业的变革性发展,亦赋能矿产资源预测向智能预测跨越。本文提出一种大模型驱动的矿产资源智能预测超级智能体方法,以多模态大模型(如DeepSeek、通义千问)为基础底座,依托通用智能体技术创建由管理智能体和智能体群构成的超级智能体。智能体群包括地质智能体群、地球物理智能体群、地球化学智能体群、遥感智能体群等,每个智能体群含有多个单一智能体或小型智能体群,每个智能体访问具体的工具(本地自定义、网络及自动生成)、数据等。智能预测超级智能体自动感知外界发送的预测要求,由管理智能体串行或并行调用多个智能体群、单一智能体(如生成二维图)、工具(如插值)、访问数据等完成矿产资源智能化预测任务。以地球化学图生成为例,深度剖析通过智能体与大模型交互完成任务的内部运行机制,一键式智能生成一种或多种地球化学图,证明智能计算方法的有效性。通过将大模型、智能体与矿产资源预测业务三者深度融合,在输入为文字或语音时即可完成零代码的预测任务,为创建矿产资源智能预测新范式提供有益探索。

关键词: 矿产资源预测, 智能预测, 大模型, 大语言模型, 智能体, 超级智能体

Abstract:

Mineral resource prediction is a key area of research in mathematical geoscience, involving the integration of diverse, cross-disciplinary geoscientific datasets. This process involves significant computational complexity and workload, along with semantic inconsistencies, which pose significant challenges to researchers. The emergence of next-generation generative artificial intelligence, particularly large language models (LLMs) and intelligent agents, is catalyzing transformative advances across industries, facilitating the transition toward intelligent mineral prediction. This study proposes a super-agent framework for AI-driven mineral resource forecasting, built upon multimodal large language models (MLLMs) (e.g., DeepSeek, Qianwen) and agent orchestration frameworks. The super-agent framework comprises a management agent and multiple specialized agent groups (geological, geophysical, geochemical, remote sensing), each comprising multiple atomic agents or lightweight agent collectives. Individual agents are capable of accessing and invoking specific tools—including those locally deployed, cloud-based, or dynamically generated—as well as datasets and services. Upon receiving external prediction tasks, the management agent dynamically coordinates the workflow. This involves orchestrating specialized agent groups, atomic agents (e.g., for geochemical map generation), analytical tools (e.g., interpolation algorithms), and relevant data sources, executing tasks either sequentially or in parallel as needed. Using geochemical map generation as a case study, we elucidate the internal mechanisms of agent-model collaboration, enabling one-click generation of numerous geochemical element maps (up to 39 different elements), thereby validating the effectiveness of this AI-driven approach. By seamlessly integrating multimodal LLMs, intelligent agents, and domain-specific mineral prediction workflows, this framework enables zero-code operation with multimodal interaction through natural language (text or voice), and presents a promising approach for establishing a new paradigm in intelligent mineral resource prediction.

Key words: mineral resources prediction, intelligent prediction, large models, large language models, agent, super-agents

中图分类号:

王永志, 温世博, 李博文, 陈星宇, 董宇浩, 田江涛, 王斌, Muhammed Atif BILAL, 纪政, 孙丰月. 大模型驱动的矿产资源智能预测超级智能体构建方法探索[J]. 地学前缘, 2025, 32(4): 38-45.

WANG Yongzhi, WEN Shibo, LI Bowen, CHEN Xingyu, DONG Yuhao, TIAN Jiangtao, WANG Bin, Muhammed Atif BILAL, JI Zheng, SUN Fengyue. Construction technology of super-agents for intelligent mineral resources prediction driven by large model[J]. Earth Science Frontiers, 2025, 32(4): 38-45.

图/表 5

图1 智能体基本工作原理(据文献[51] 修改)

Fig.1 Mechanism of agent. Modified after [51].

图2 大模型与矿产预测智能体关系

Fig.2 Relationship between LLM agents and mineral resources prediction

图3 基于大模型的智能预测超级智能体框架

Fig.3 Framework of mineral resources prediction super-agent based on LLMs

图4 大模型智能体驱动的地球化学图自动生成流程

Fig.4 Automatic generation of geochemical contour maps driven by large model agents

图5 大模型智能体驱动智能生成的地球化学图

Fig.5 Geochemical contour map generated intelligently driven by large model agents

参考文献 56

[1]	成秋明. 极端地质事件定量模拟与预测[J]. 中国科学: 地球科学, 2022, 52(6): 975-991.
[2]	SINGER D A. Progress in integrated quantitative mineral resource assessments[J]. Ore Geology Reviews, 2010, 38(3): 242-250.
[3]	赵鹏大, 陈建平, 张寿庭. “三联式” 成矿预测新进展[J]. 地学前缘, 2003, (2): 455-463.
[4]	王世称. 综合信息矿产预测理论与方法体系新进展[J]. 地质通报, 2010, 29(10): 1399-1403.
[5]	王世称, 陈永良, 夏立显, 等. 综合信息矿产预测理论与方法[M]. 北京: 科学出版社, 2000.
[6]	CHENG Q. Non-linear theory and power-law models for information integration and mineral resources quantitative assessments[J]. Mathematical Geosciences, 2008, 40(5): 503-532.
[7]	肖克炎, 樊铭静, 孙莉, 等. 矿床成矿系列综合信息预测理论方法及其应用[J]. 地球学报, 2023, 44(5): 769-780.
[8]	何文娜. 大数据时代基于物联网和云计算的地质信息化研究[D]. 长春: 吉林大学, 2013.
[9]	肖克炎, 李程, 唐瑞, 等. 大数据智能预测评价[J/OL]. 地学前缘, 1-18[2025-6-28]. https://doi.org/10.13745/j.esf.sf.2025.4.58.
[10]	周永章, 肖凡. 管窥人工智能与大数据地球科学研究新进展[J]. 地学前缘, 2024, 31(4): 1-6. DOI
[11]	成秋明. 什么是数学地球科学及其前沿领域?[J]. 地学前缘, 2021, 28(3): 6-25. DOI
[37]	LIN Z, DENG C, ZHOU L, et al. GeoGalactica: a scientific large language model in Geoscience[J]. arXiv preprint, 2023.
[38]	王嘉翔. 基于大语言模型的斑岩型铜矿找矿标志智能识别研究[D]. 长春: 吉林大学, 2024.
[39]	ZHAO W X, ZHOU K, LI J, et al. A survey of large language models[J]. arXiv preprint, 2023. DOI: arXiv:2303.18223.
[40]	ZHANG D, YU Y, DONG J, et al. MM-LLMs: recent advances in multimodal large language models[J]. arXiv preprint, 2024. DOI: arXiv:2401.13601.
[41]	ZHANG W, CAI M, ZHANG T, et al. EarthGPT: a universal multimodal large language model for multisensor image comprehension in remote sensing domain[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 1-20.
[42]	CHEN J, LIN B, XU R, et al. MapGPT: map-guided prompting with adaptive path planning for vision-and-language navigation[J]. arXiv preprint, 2024. DOI: arXiv:2401.07314.
[43]	ZHAN Y, XIONG Z, YUAN Y. SkyEyeGPT: unifying remote sensing vision-language tasks via instruction tuning with large language model[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2025, 221: 64-77.
[44]	DAVIES R S, TROTT M, GEORGI J, et al. Artificial intelligence and machine learning to enhance critical mineral deposit discovery[J]. Geosystems and Geoenvironment, 2025, 4(2): 100361.
[45]	XIE J, CHEN Z, ZHANG R, et al. Large multimodal agents: a survey[J]. arXiv preprint, 2024. DOI: arXiv:2402.15116.
[46]	ALZUBI S, BROOKS C, CHINIYA P, et al. Open deep search: democratizing search with open-source reasoning agents[J]. arXiv preprint, 2025. DOI: arXiv:2503.20201.
[47]	GUO S, DENG C, WEN Y, et al. Ds-agent: automated data science by empowering large language models with case-based reasoning[J]. arXiv preprint, 2024. DOI: arXiv:2402.17453.
[12]	YANG F, ZUO R, KREUZER O P. Artificial intelligence for mineral exploration: a review and perspectives on future directions from data science[J]. Earth-Science Reviews, 2024: 104941.
[13]	XIAO F, CHEN W, WANG J, et al. A hybrid logistic regression: gene expression programming model and its application to mineral prospectivity mapping[J]. Natural Resources Research, 2022, 31(4): 2041-2064.
[14]	WANG J, ZUO R, XIONG Y. Mapping mineral prospectivity via semi-supervised random forest[J]. Natural Resources Research, 2020, 29(1): 189-202.
[15]	ABEDI M, NOROUZI G H, BAHROUDI A. Support vector machine for multi-classification of mineral prospectivity areas[J]. Computers & Geosciences, 2012, 46: 272-283.
[16]	XIONG Y, ZUO R. A positive and unlabeled learning algorithm for mineral prospectivity mapping[J]. Computers & Geosciences, 2021, 147: 104667.
[17]	YANG F, ZUO R. Geologically constrained convolutional neural network for mineral prospectivity mapping[J]. Mathematical Geosciences, 2024, 56(8): 1605-1628.
[18]	XU Y, ZUO R. An interpretable graph attention network for mineral prospectivity mapping[J]. Mathematical Geosciences, 2024, 56(2): 169-190.
[19]	唐赧钰, 申俊峰, 陈强, 等. 基于深度学习的安徽黄屯铜金矿床黄铜矿定量分析及其意义[J]. 现代地质, 2024, 38(6): 1585-1593.
[20]	ZHANG Y, WEI C, WU S, et al. GeoGPT: understanding and processing geospatial tasks through an autonomous gpt[J]. arXiv preprint, 2023. DOI: 2307.07930.
[21]	李强. 2025政府工作报告[R]. [2025-03-05]. https://www.gov.cn/zhuanti/2025qglh/2025zfgzbgjd/index.htm.
[22]	YU S, DENG H, LIU Z, et al. Identification of geochemical anomalies using an end-to-end transformer[J]. Natural Resources Research, 2024, 33(3): 973-994.
[23]	WANG Z, ZUO R, DONG Y. Mapping geochemical anomalies through integrating random forest and metric learning methods[J]. Natural Resources Research, 2019, 28(4): 1285-1298.
[24]	冯婷婷, 蔡诗柔, 张振杰. 基于知识图谱的碳酸岩型稀土矿成矿要素挖掘[J]. 地学前缘, 2025, 32(4): 262-279.
[25]	陈国雄, 张越鹏, 罗磊, 等. 数据驱动斑岩型矿床时空预测模型[J]. 地学前缘, 2025, 32(4): 46-59.
[26]	ZUO R, CARRANZA E J M. Machine learning-based mapping for mineral exploration[J]. Mathematical Geosciences, 2023, 55(7): 891-895.
[27]	FENG Y, WANG Y, WANG C, et al. A fuzzy gold mineral prediction model integrating with knowledge-driven and data-driven: a case study of the Hatu region in Xinjiang, China[J]. Minerals, 2024, 14(12): 1209.
[28]	MAO X, ZHONG H, LIU Z, et al. 3D numerical modeling for investigating structural controls on orogenic gold mineralization, Sanshandao gold belt, eastern China[J]. Natural Resources Research, 2024, 33(4): 1413-1437.
[29]	袁峰, 李晓晖, 田卫东, 等. 三维成矿预测关键问题[J]. 地学前缘, 2024, 31(4): 119-128. DOI
[30]	WANG Z, LI T, ZUO R. Leucogranite mapping via convolutional recurrent neural networks and geochemical survey data in the Himalayan orogen[J]. Geoscience Frontiers, 2024, 15(1): 101715.
[31]	罗杰, 周仲礼, 邹天一, 等. 基于PSO-CNN的深部找矿预测模型构建[J]. 成都理工大学学报(自然科学版), 2022, 49(6): 697-708.
[32]	王永昌, 刘彩云, 熊杰, 等. 基于生成对抗网络的半监督地震波阻抗反演[J]. 现代地质, 2024, 38(6): 1585-1593.
[33]	LACHOWYCZ S. Utility of artificial intelligence in geoscience[J]. Nature Geoscience, 2024, 17(10): 953-955.
[34]	WU T, HE S, LIU J, et al. A brief overview of ChatGPT: the history, status quo and potential future development[J]. IEEE/CAA Journal of Automatica Sinica, 2023, 10(5): 1122-1136.
[35]	LIU A, FENG B, WANG B, et al. Deepseek-v2: a strong, economical, and efficient mixture-of-experts language model[J]. arXiv preprint, 2024. DOI: arXiv:2405.04434.
[36]	TOUVRON H, LAVRIL T, IZACARD G, et al. Llama: open and efficient foundation language models[J]. arXiv preprint, 2023. DOI: arXiv:2302.139713.
[48]	SHEN M, YANG Q. From mind to machine: the rise of Manus AI as a fully autonomous digital agent[J]. arXiv preprint, 2025. DOI: arXiv:2505.02024.
[49]	NOVIKOV A, VŨ N, EISENBERGER M, et al. AlphaEvolve: a coding agent for scientific and algorithmic discovery[J]. arXiv preprint, 2025. DOI: arXiv:2506.13131.
[50]	DURANTE Z, HUANG Q, WAKE N, et al. Agent AI: surveying the horizons of multimodal interaction[J]. arXiv preprint, 2024. DOI: arXiv:2401.03568.
[51]	HÄNDLER T. A taxonomy for autonomous LLM-powered multi-agent architectures[C]. KMIS, 2023: 85-98.
[52]	GAO Y, XIONG Y, GAO X, et al. Retrieval-augmented generation for large language models: a survey[J]. arXiv preprint, 2023. DOI: arXiv:2312.10997.
[53]	AGTERBERG F P. Computer programs for mineral exploration[J]. Science, 1989, 245(4913): 76-81. PMID
[54]	LIU J, YANG C, LU Z, et al. Graph foundation models: concepts, opportunities and challenges[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025, 47(6): 5023-5044.
[55]	ALBOODY A, SLAMA R. Graph transformer mixture-of-experts (GTMoE) for 3D hand gesture recognition[C]// Intelligent systems and applications. Cham: Springer, 2024: 317-336.
[56]	DAI D, DENG C, ZHAO C, et al. DeepSeekMoE: towards ultimate expert specialization in mixture-of-experts language models[J]. arXiv preprint, 2024. DOI: arXiv:2401.06066.

大模型驱动的矿产资源智能预测超级智能体构建方法探索

Construction technology of super-agents for intelligent mineral resources prediction driven by large model

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 56

相关文章 5

编辑推荐

Metrics

本文评价

[1]	肖克炎, 李程, 唐瑞, 王瑶, 孙莉, 柳炳利, 樊铭静. 大数据智能预测评价[J]. 地学前缘, 2025, 32(4): 20-37.
[2]	李博文, 王永志, 丁正江, 王斌, 温世博, 董宇浩, 纪政. 基于大模型与GraphRAG的胶东金矿智能搜索技术[J]. 地学前缘, 2025, 32(4): 155-164.
[3]	成秋明. 面向人类智能与人工智能融合的矿产资源预测新范式[J]. 地学前缘, 2025, 32(4): 1-19.
[4]	周琦, 吴冲龙. 基于大数据的智慧探矿模式实验研究与进展[J]. 地学前缘, 2024, 31(6): 350-367.
[5]	周永章, 张前龙, 黄永健, 杨威, 肖凡, 吉俊杰, 韩枫, 唐磊, 欧阳冲, 沈文杰. 钦杭成矿带斑岩铜矿知识图谱构建及应用展望[J]. 地学前缘, 2021, 28(3): 67-75.