[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-175":3,"related-175":65},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":8,"published_at":11,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":23,"tags":25,"primary_topic":8,"secondary_topics":31,"classification":32,"search_phrases":58,"slug":61,"view_count":62,"doi":63,"paper":8,"created_at":64},175,"Leveraging Large Language Models (LLMs) for GeoAI-enabled Digital Agro-advisory","https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract","International Rice Research Institute (India) 的 Shalini Gakhar 和 Raj Kumar Singh 在综述中系统分析 LLMs 在将复杂 GeoAI 数据转化为面向农户的可操作建议方面的潜力。区分了 LLM 已经验证的能力与农业遥感中的潜在应用，强调 RAG（检索增强生成）与 SFT（监督微调）在缓解误读风险中的重要性；探讨多模态方法在作物健康评估和生物物理异常识别的应用，并主张'人在回路'决策方式。综述最后展望了从田间结果中迭代学习的前瞻性研究范围和闭环系统。",null,"## Abstract\n\nDigital agriculture has undergone a profound transformation driven by rapid advances in Earth observation, unmanned aerial vehicles and advanced data processing techniques. Despite this progress, climate-smart agriculture management faces a critical challenge: a farmer-centric agro-advisory system. Farmers and extension workers have limited access to GeoAI-based agro-advisory services, primarily due to limited data access and insufficient technical support. This study synthesises the potential of Large Language Models to address this challenge by translating complex GeoAI data into actionable and human-centric advice. The review examines gaps, recent advancements, and the model architectures required to adapt general-purpose LLMs for remote-sensing-based crop monitoring and management. We distinguish between experimentally validated capabilities of LLMs and the broader prospective applications proposed for agricultural remote sensing, noting that many current advances are primarily driven by multimodal foundation models, computer vision, and GeoAI systems rather than standalone LLMs. The study highlights the importance of advanced Retrieval-Augmented Generation and Supervised Fine-Tuning in agronomic science and mitigating the risk of misinterpretation. Further, we examine the emerging capabilities of multimodal approaches, which can seamlessly integrate visualisation and textual reasoning to support stakeholders in assessing crop health conditions and biophysical anomalies. The representative case studies, spanning multiple geographies and including voice-based advisory systems, demonstrate the shift from static advisory tools to dynamic, interactive recommendation systems. We highlight current challenges, such as the need for region-specific fine-tuning, data governance, and operation in low-connectivity environments, and advocate for a “human-in-the-loop” approach to decision-making. Through this, the LLMs will function as co-pilots assisting multiple stakeholders rather than as autonomous decision-makers vulnerable to biased output or hallucination problems. The review concludes by outlining a forward-looking research scope and closed-loop systems that iteratively learn from field outcomes.\n\n## 1 Introduction\n\nOver the past decade, digital agriculture has undergone a profound transformation driven by rapid advances in Earth observation, unmanned aerial vehicles (UAVs), Internet of Things (IoT), and proximal sensing technologies. These developments have enabled continuous, multiscale, and unprecedented volumes of spatiotemporal data to be generated for monitoring crops, soils, hydro-ecology, and agroecosystem processes (). Operational satellite missions such as Landsat, Sentinel, MODIS, LISS, Cartosat, WorldView, PlanetScope, combined with very high-resolution UAV imagery, now form the backbone of precision agriculture, supporting applications including crop monitoring, yield estimation, cropping practices and the implementation of climate-smart agricultural practices (; ). Despite the growing maturity of remote sensing technologies, the increasing availability of data has exposed a fundamental limitation in digital agriculture: the interpretation and convergence bottleneck. Remote sensing data are typically expressed as spectral reflectance or backscatter values, vegetation indices, and derived biophysical parameters, which require substantial agronomic and contextual expertise to interpret and apply to improve crop management. Translating these quantitative indicators into actionable recommendations requires an understanding of crop type, phenological stage, soil characteristics, climatic conditions, and management history, which is complex (). This complexity constrains the timely delivery of decision-support information, particularly in smallholder-dominated agricultural systems where access to trained extension services remains limited and costly (). Recent advances in Large Language Models (LLMs) offer a promising framework for addressing these challenges. LLMs such as the Generative Pre-trained Transformer (GPT), Large Language Model Meta AI (LLaMA), and Pathways Language Model (PaLM) are based on transformer architectures that model long-range dependencies, support contextual reasoning, and enable large-scale knowledge synthesis (; ). Unlike conventional machine learning models, which are optimised for narrowly defined prediction tasks, LLMs are designed to interpret, explain, and adapt to complex information across domains. These capabilities position LLMs as an effective interface layer for translating remote-sensing-derived indicators into human-interpretable, decision-relevant guidance for agronomists, extension agents, and farmers.\n\nThe potential of LLM-enabled agricultural advisory systems is further enhanced by recent advances in multimodal foundation models and vision-language frameworks that integrate visual, textual, and geospatial information. It is important, however, to distinguish between standalone LLMs and multimodal computer vision systems. Standalone LLMs are primarily designed for language understanding, knowledge synthesis, reasoning, and natural-language generation, whereas remote sensing image interpretation is largely performed by computer vision, vision-language, and multimodal foundation models. Recent vision-language models (VLMs) and geospatial foundation models have demonstrated increasing capability in analysing remote sensing imagery, field photographs, and UAV observations for applications such as crop stress detection, disease identification, canopy trait assessment, phenological monitoring, and interpretation of spectral anomalies (; ). These models generate structured visual insights that can subsequently be translated into actionable recommendations through LLM-based reasoning and conversational interfaces. This distinction is particularly important given the exponential growth of Earth observation imagery, driven by the expansion of satellite constellations and the widespread adoption of UAV platforms for agricultural management. While multimodal foundation models and GeoAI systems increasingly perform perception and analytical tasks associated with image understanding, LLMs primarily contextualise these outputs, integrate agronomic knowledge, retrieve relevant information from external knowledge bases, and communicate recommendations in a human-interpretable form. Therefore, the emerging opportunity lies not in replacing remote sensing analytics with LLMs, but in integrating multimodal perception systems with language-based reasoning frameworks to support end-to-end agricultural decision-making. Despite this rapid progress, a systematic understanding of how LLMs can be effectively integrated into remote sensing-based agricultural workflows remains lacking. The existing literature predominantly focuses on remote sensing for crop monitoring or the application of LLMs in broader agricultural contexts, leaving a crucial gap in current knowledge. To date and to our knowledge, no comprehensive review has synthesised how LLMs specifically enhance the interpretation, communication, and decision-support functionalities of remote sensing in digital agriculture. Moreover, critical deployment challenges such as the risk of model “hallucination”, the necessity for region-specific fine-tuning, and strategies for effective implementation remain largely unexplored. Existing studies tend to focus either on remote sensing-based crop monitoring and modelling (; ) or on the application of LLMs in broader agricultural advisory and knowledge systems () with relatively little synthesis across these domains. As agricultural monitoring systems increasingly transition from descriptive assessment to operational advisory services, the need for intelligent interpretation frameworks becomes more acute. LLMs are efficient in data mining through historical observations and practices, textual reports, and agronomic knowledge bases, thereby reducing human effort. The Earth observation foundation models highlight the growing integration of geospatial AI, multimodal learning, and LLM-assisted agricultural analytics. For instance, the IBM-NASA _Prithvi_ model uses multi-temporal satellite imagery and transformer-based learning to support applications such as crop monitoring, land-use mapping, and environmental assessment (). Similarly, the European Space Agency (ESA) _WorldCereal_ initiative applies AI-driven analysis of Sentinel satellite data to generate global crop maps for major cereals, including wheat and maize (). Although these systems are not standalone LLMs, they demonstrate how foundation models and multimodal geospatial AI are advancing scalable agricultural monitoring and advisory systems. Future integration of these Earth observation models with conversational LLM frameworks could further improve farmer-centric decision support and agro-advisory services (). Therefore, a coherent understanding of how LLMs can be systematically integrated into remote sensing-driven agricultural workflows will improve farm management and decision support.\n\nThis review aims to address the following vital gaps: (a) To synthesise the potential for integrating advanced remote sensing tools\u002Ftechniques and LLMs for precision\u002Fdigital agriculture and advisory generation. (b) To present representative case studies illustrating the integration of LLMs with satellite- and UAV-based crop monitoring workflows, focusing on rice-based cropping systems. (c) To critically examine technical, operational, and governance challenges, including reliability, validation, and ethical considerations associated with LLM-based advisory. (d) To outline a forward-looking roadmap\u002Fframework that prioritise multimodal model development, regionally adaptive fine-tuning, and edge or low-bandwidth deployment strategies.\n\nBy bridging remote sensing analytics with LLM-based knowledge representation, this review argues that foundation models have the potential to significantly enhance the scalability, contextual relevance, and inclusivity of digital agriculture. When appropriately grounded in agronomic knowledge and validated against observational data, LLMs can transform geospatial information from an expert-centric analytical resource into a human-centred and actionable advisory ecosystem that supports informed decision-making (). [Figure 1](https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract#F1) illustrates the knowledge translation pipeline from remote sensing data acquisition through biophysical indicator derivation to LLM-driven farm-level advisory, highlighting the interpretation gap that LLMs are positioned to bridge.\n\nFIGURE 1\n\n### 1.1 Scope and methodology of this review\n\nThis review follows a structured narrative and conceptual review approach to map the intersection of LLMs, remote sensing, and agricultural advisory systems. The literature search was conducted across Web of Science, Scopus, Google Scholar, and the arXiv preprint repository using Boolean combinations of key terms including “Large Language Model,” “LLM,” “Generative AI,” “remote sensing,” “digital agriculture,” “precision farming,” “crop advisory,” “farm advisory chatbot,” “RAG agriculture,” and “multimodal geospatial AI.” The search covered publications from 2010 to 2025, with emphasis on post-2020 literature to capture the rapidly evolving LLM landscape. Inclusion criteria required that studies: (i) explicitly addressed the application of LLMs or conversational AI in agricultural contexts, (ii) incorporated remote sensing data or geospatial analytics, or (iii) presented technical architectures (SFT, RAG, RLHF, multimodal fusion) with relevance to agro-advisory generation. Studies that focused exclusively on conventional machine learning for crop classification, without an advisory or interpretive component, were excluded. Grey literature, including technical reports from CGIAR centres, FAO, and national agricultural extension bodies, was also consulted to capture operationally deployed systems not yet represented in peer-reviewed literature. A total of over 330 sources were screened, of which 97 were retained for detailed synthesis ([Figure 2](https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract#F2)). The review is organised thematically rather than chronologically, progressing from the remote sensing foundation ([Section 2](https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract#s2)) through LLM adaptation pathways ([Section 3](https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract#s3)), representative field deployments ([Section 4](https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract#s4)), deployment challenges ([Section 5](https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract#s5)), and future research frontiers ([Section 6](https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fremote-sensing\u002Farticles\u002F10.3389\u002Ffrsen.2026.1839369\u002Fabstract#s6)).\n\nFIGURE 2\n\n## 2 Overview of remote sensing in digital agriculture and associated challenges\n\nRemote sensing has become a cornerstone of modern digital agriculture, providing the spatial and temporal multi-dimensional information needed to monitor, assess, and manage complex agroecosystems. The expansion of various satellite platforms, UAV-based imaging systems, and sensor-integrated field networks has revolutionised our ability to observe crop conditions, map spatiotemporal variability, and optimise resource use (; ). These advances fuel precision agriculture, enabling data-driven decisions from crop suitability, fertilisation and pest management, crop health monitoring, water resource utilization, and yield forecasting. Although a large proportion of the latest remote sensing-based crop management uses machine learning models, most of these trained models operate in silos and require human intervention for consistent interpretation and translation into actionable advice for farmers, extension workers, and policymakers.\n\n### 2.1 Landscaping the diverse remote sensing applications in agriculture\n\nSatellite remote sensing remains the backbone of national, regional, and global agricultural monitoring systems, providing consistent, spatially explicit observations of cropland dynamics over large areas. A diverse constellation of publicly accessible and commercial Earth observation platforms, including multispectral optical sensors such as Sentinel-2, Landsat-8\u002F9, and PlanetScope, MODIS, WorldView, LISS, as well as synthetic aperture radar (SAR) systems such as Sentinel-1, EOS-04, and hyperspectral data such as HySI and PRISMA, and satellite data-derived products such as soil moisture (by SMAP), now deliver high-frequency data streams essential for operational agricultural applications (; ; ). Despite these advances, a key limitation of satellite remote sensing lies in its spatial, spectral, and temporal resolutions. Most available data have revisit times of 5–6 days–20 days and moderate spatial resolution, which limits their application during the crucial crop-monitoring phase. While freely available sensors typically operate at spatial resolutions of 10–30 m, which are often insufficient to capture within-field variability in small farmlands with field sizes below 1 ha (; ). As a result, satellite products are highly effective for strategic monitoring and policy-level assessments in homogeneous fields. However, they are less suited for heterogeneous small field-level decision-making without costly high-resolution imagery.\n\nTo overcome the spatial resolution limitations of satellite-based observations, UAVs have emerged as a critical component of field-scale agricultural intelligence. UAV platforms provide centimetre-level spatial detail with flexible deployment schedules, allowing crop condition monitoring at user-defined temporal frequencies in key phenological stages (; ). Advanced UAV systems can be equipped with a diverse suite of lightweight sensors, each contributing complementary information on crop status. High-resolution RGB cameras generate detailed orthomosaics that support plant stand assessment, lodging detection, weed mapping, and canopy structure analysis (). Recent advances in deep learning have further expanded the capabilities of UAV and mobile vision systems beyond traditional vegetation index mapping. Convolutional neural networks and real-time object detection frameworks, particularly the YOLO family of models, have enabled automated crop detection, fruit localisation, segmentation, counting, and growth-stage recognition at field and greenhouse scales. For example, developed an integrated YOLO-based system for capsicum detection, segmentation, growth-stage classification, fruit counting, and real-time mobile identification. Such perception-oriented computer vision models provide detailed crop-level intelligence that can complement satellite-derived biophysical indicators and serve as important upstream information sources for multimodal advisory systems. In future GeoAI-enabled agricultural platforms, outputs from UAV-based detection and crop-stage recognition models could be integrated with LLM reasoning frameworks to generate more context-aware recommendations on harvesting, crop management, and resource optimisation. In parallel, thermal infrared sensors enable the mapping of canopy temperature, which serves as a proxy for plant water status and transpiration efficiency, supporting irrigation scheduling and drought stress detection at fine spatial scales (). These high-fidelity datasets are already reshaping agricultural research and innovation pipelines. For farmers and extension agents, a detailed crop health or thermal anomaly map may indicate variability. However, it does not inherently convey why stress is occurring or what management action should follow. Bridging this interpretive gap remains a challenge in translating UAV-based insights into timely, actionable agricultural decisions.\n\nRecent advances in UAV and mobile vision systems have further extended field-scale monitoring capabilities through deep learning-based crop detection, segmentation, counting, and growth-stage classification. demonstrated the application of YOLO-based frameworks for real-time capsicum detection, segmentation, growth-stage identification, and mobile deployment, highlighting the growing convergence between computer vision and precision horticultural management.\n\nRemote sensing observations from satellite and UAV platforms are rarely interpreted in their raw spectral form. Instead, they are systematically transformed into standardized biophysical and agroecological indicators that provide an interpretable description of crop condition, canopy functioning, and landscape-scale processes. Optical satellite sensors capture surface reflectance across visible, near-infrared, and shortwave infrared wavelengths, enabling the development of widely adopted vegetation indices. Metrics such as the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Soil-Adjusted Vegetation Index (SAVI), Normalized Difference Red Edge Index (NDRE), Photochemical Reflectance Index (PRI), and Normalized Difference Moisture Index (NDMI), are routinely used to characterize crop vigour, biomass accumulation, canopy development, and leaf water content throughout the growing season (). Beyond spectral greenness, structural and radiative parameters offer a more process-oriented perspective on crop growth. Leaf Area Index (LAI) characterises canopy structure by quantifying leaf surface area available for light interception, while the fraction of absorbed Photosynthetically Active Radiation (fAPAR) represents the efficiency with which vegetation captures incoming solar energy. These v","Frontiers in Remote Sensing 2026, Data Fusion and Assimilation section","2026-07-07T00:00:00Z","论文",10,false,74,{"impact":17,"substance":18,"depth":17,"authority":19,"freshness":20,"relevant":21,"comment":22},18,22,14,2,1,"综述论文探讨LLM在GeoAI农业咨询中的应用，具有前瞻性，但时效性较低。",[24],{"name":10,"url":6},[26,27,28,29,30],"大语言模型","GeoAI","数字农情咨询","RAG","人在回路",[],{"status":33,"locked":14,"version":34,"method":35,"evidence":36},"needs_review","2026-10-22.2","rules",{"candidates":37,"agricultural_evidence":41,"reason":57},[38,46,52],{"code":39,"score":40,"supported":41,"title":42,"abstract":43,"tags":45},"A1",3,true,[],[44],"rice",[],{"code":47,"score":40,"supported":41,"title":48,"abstract":49,"tags":51},"I3",[],[50],"遥感",[],{"code":53,"score":21,"supported":14,"title":54,"abstract":55,"tags":56},"I2",[],[],[26],"多个产业证据相当，需要确认主归属",[59,60],"数字农情咨询 大语言模型 人在回路 GeoAI","数字农情咨询 大语言模型","数字农情咨询大语言模型人在回路GeoAI-175",9,"10.3389\u002Ffrsen.2026.1839369\u002Fabstract","2026-08-01T23:53:44.541424Z",{"total":66,"page":21,"page_size":66,"items":67},6,[68,153,236,311,384,456],{"id":69,"title":70,"url":71,"summary":72,"summary_zh":73,"content":8,"source_name":74,"source_url":71,"published_at":75,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":76,"score_detail":77,"sources":82,"tags":84,"primary_topic":89,"secondary_topics":95,"classification":100,"search_phrases":124,"slug":127,"view_count":128,"doi":129,"paper":130,"created_at":152},5376,"Agri-Synergy: A Hardware–Software Co-Designed Hybrid IoT System for Sensor Interface Standardization and LLM-Driven Interactive Decision Support","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fagriengineering8100446","This paper presents Agri-Synergy, a hybrid Internet of Things (IoT) system featuring a hardware–software co-design to tackle sensor data heterogeneity and the barriers to effective AI-driven decision support in smart agriculture. Unlike software-based data harmonization, we propose a universal sensor printed circuit board (PCB) with RS485\u002FModbus interface for hardware unification, complemented by on-node edge processing that ensures standardized data from source. To lower the entry barrier for farmers, the system integrates the DeepSeek large language model (LLM) API, combining static agricultural knowledge with real-time sensor data, for context-aware, natural language-based decision support, functioning without local model deployment. Field deployment showed reference-verified sensor accuracy predominantly within manufacturer specifications, stable 4G connectivity during the 4-h benchmark but intermittent dropouts during the 20-day deployment, and remote-control success rates of 96% (normal) and 80% (weak signal at 300 m). Multi-node concurrent transmission resulted in average latency of 238.0 ± 27.1 ms and packet loss below 0.5%. The system achieves a 3–5× cost reduction compared to commercial platforms. These results demonstrate that addressing data heterogeneity from source through hardware unification and software edge processing, while leveraging AI to lower the barrier to system adoption, provides a more reliable foundation for data synchronization and accessible AI-driven analytics.","本文提出Agri-Synergy，一种采用软硬件协同设计的混合物联网（IoT）系统，旨在解决智慧农业中传感器数据异构性问题以及有效AI驱动决策支持的障碍。与基于软件的数据协调方式不同，我们提出了一种带有RS485\u002FModbus接口的通用传感器印刷电路板（PCB）以实现硬件统一，并辅以节点端边缘处理，确保从源头获得标准化数据。为降低农民的使用门槛，该系统集成了DeepSeek大语言模型（LLM）API，将静态农业知识与实时传感器数据相结合，提供上下文感知的、基于自然语言的决策支持，且无需本地部署模型即可运行。现场部署表明，经参考验证的传感器精度基本处于制造商规格范围内；在4小时基准测试期间4G连接稳定，但在20天部署期间出现间歇性断连；远程控制成功率分别为96%（正常信号）和80%（300 m处弱信号）。多节点并发传输的平均延迟为238.0 ± 27.1 ms，丢包率低于0.5%。与商业平台相比，该系统实现了3–5倍的成本降低。这些结果表明，通过硬件统一和软件边缘处理从源头解决数据异构性问题，同时利用AI降低系统采用门槛，可为数据同步和可及的AI驱动分析提供更可靠的基础。","AgriEngineering","2026-10-19T00:00:00Z",76,{"impact":78,"substance":79,"depth":17,"authority":80,"freshness":62,"relevant":21,"comment":81},16,21,12,"硬件统一接口与边缘处理结合大模型决策支持的农业物联网系统论文，方法新颖、有实地部署数据，但属细分技术进展，公共影响有限。",[83],{"name":74,"url":71},[85,86,87,26,88],"智慧农业","农业人工智能","农业物联网","传感器标准化",{"code":90,"slug":91,"name":92,"path":93,"parent_code":94},"I5","agricultural-iot","物联网传感与边缘计算","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-iot","I",[96],{"code":53,"slug":97,"name":98,"path":99,"parent_code":94},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai",{"status":101,"locked":14,"version":34,"method":35,"evidence":102},"classified",{"candidates":103,"agricultural_evidence":41},[104,111,117],{"code":90,"score":17,"supported":41,"title":105,"abstract":107,"tags":110},[106],"IoT",[108,109,106],"物联网","internet of things",[87],{"code":53,"score":78,"supported":41,"title":112,"abstract":114,"tags":116},[113],"decision support",[115,113],"large language model",[86,26],{"code":118,"score":119,"supported":41,"title":120,"abstract":121,"tags":123},"I1",7,[],[85,122],"smart agriculture",[85],[125,126],"Agri-Synergy 物联网 传感器","DeepSeek 决策支持","Agri-Synergy物联网传感器-5376",0,"10.3390\u002Fagriengineering8100446",{"doi":129,"openalex_id":131,"authors":132,"venue":74,"cited_by_count":128,"oa_url":71,"card":145,"direction":149,"ingested_from":151},"W7222920424",[133,135,137,139,141,143],{"name":134,"orcid":8},"Leqi Lai",{"name":136,"orcid":8},"Langtao Duan",{"name":138,"orcid":8},"Yuhao Chen",{"name":140,"orcid":8},"Yuhang Zheng",{"name":142,"orcid":8},"Jinyu Yang",{"name":144,"orcid":8},"Mingyao Wen",{"tldr":146,"method":147,"finding":148,"direction":149,"opportunity":150},"提出软硬件协同的混合物联网系统，统一传感器接口并用大模型提供交互式决策支持。","通用RS485\u002FModbus传感器PCB、节点边缘处理、DeepSeek大模型A","传感器精度达标，多节点延迟238ms、丢包低于0.5%，成本降低3-5倍。","智慧农业 \u002F 农业物联网","可研究弱信号下通信可靠性提升与边缘-云端LLM协同的长期田间鲁棒性。","openalex","2026-10-20T23:30:08.319698Z",{"id":154,"title":155,"url":156,"summary":157,"summary_zh":158,"content":8,"source_name":159,"source_url":156,"published_at":160,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":161,"score_detail":162,"sources":164,"tags":166,"primary_topic":169,"secondary_topics":170,"classification":171,"search_phrases":218,"slug":221,"view_count":128,"doi":222,"paper":223,"created_at":235},5232,"Current applications of artificial intelligence for advancing agriculture, sustainability and climate mitigation","https:\u002F\u002Fdoi.org\u002F10.1007\u002Fs44163-026-02540-1","Artificial intelligence (AI) is reshaping agriculture by enabling data-driven decision-making across crop production systems and supporting the transition toward more sustainable, resilient, and climate-smart farming. This review critically synthesizes advances in AI applications spanning precision irrigation, nutrient and soil management, yield prediction, pest and disease monitoring, computer vision, autonomous agricultural robotics, and large language model (LLM)-based decision support systems. Beyond summarizing technological developments, the review evaluates their practical deployment, emphasizing that successful agricultural AI depends not only on predictive accuracy but also on model generalizability, explainability, economic feasibility, interoperability, and integration with existing agronomic knowledge. Emerging technologies including Vision Transformers, foundation models, Digital Twins, multimodal AI, Retrieval-Augmented Generation, and federated and continual learning—are discussed as important directions for overcoming limitations related to data scarcity, environmental variability, and scalable field implementation. The review highlights that responsible AI adoption requires transparent governance frameworks addressing data ownership, privacy, algorithmic fairness, cybersecurity, and equitable access, particularly for smallholder farming systems. Although numerous AI models report outstanding performance, widespread agricultural adoption remains constrained by cross-region validation, fragmented datasets, infrastructure requirements, and insufficient real-world deployment. Future progress will depend on multidisciplinary collaboration integrating advances in artificial intelligence, sensing technologies, robotics, crop science, and agricultural engineering to develop robust, trustworthy, and farmer-centered decision-support systems. By critically evaluating technological innovations and implementation challenges, this review provides a framework for understanding the current state of AI in agriculture while identifying research priorities that will shape the next generation of sustainable and climate-resilient agricultural systems.","人工智能（AI）正在重塑农业，通过在作物生产系统中实现数据驱动的决策，并支持向更可持续、更具韧性和气候智能型农业转型。本综述批判性地综合了AI应用的进展，涵盖精准灌溉、养分与土壤管理、产量预测、病虫害监测、计算机视觉、自主农业机器人以及基于大语言模型（LLM）的决策支持系统。除总结技术发展外，本综述还评估了其实际部署情况，强调农业AI的成功不仅取决于预测准确性，还取决于模型泛化能力、可解释性、经济可行性、互操作性以及与现有农学知识的整合。文中讨论了包括视觉Transformer、基础模型、数字孪生、多模态AI、检索增强生成以及联邦学习和持续学习在内的新兴技术，将其视为克服数据稀缺、环境变异性和田间可扩展实施等局限的重要方向。本综述强调，负责任的AI采用需要透明的治理框架，以解决数据所有权、隐私、算法公平性、网络安全和公平获取等问题，尤其是对小农耕作系统而言。尽管众多AI模型报告了卓越的性能，但农业领域的广泛采用仍受限于跨区域验证、碎片化数据集、基础设施要求以及现实世界部署不足。未来的进展将取决于多学科协作，整合人工智能、传感技术、机器人技术、作物科学和农业工程领域的进展，以开发稳健、可信且以农民为中心的决策支持系统。通过批判性评估技术创新和实施挑战，本综述提供了一个理解AI在农业中现状的框架，同时确定了将塑造下一代可持续和气候韧性农业系统的研究优先事项。","Discover Artificial Intelligence","2026-10-17T00:00:00Z",83,{"impact":18,"substance":18,"depth":17,"authority":80,"freshness":62,"relevant":21,"comment":163},"该综述系统梳理了AI在农业中的应用与落地瓶颈，覆盖精准灌溉、机器人、大模型决策支持等方向，信息密度高，但属综述类论文，产业级突破性有限，可入精选。",[165],{"name":159,"url":156},[85,86,26,167,168],"数字孪生","气候智慧农业",{"code":53,"slug":97,"name":98,"path":99,"parent_code":94},[],{"status":101,"locked":14,"version":34,"method":35,"evidence":172},{"candidates":173,"agricultural_evidence":41},[174,182,189,196,201,207,213],{"code":53,"score":175,"supported":41,"title":176,"abstract":178,"tags":181},19,[177],"artificial intelligence",[179,177,180,115,113],"计算机视觉","computer vision",[86,26],{"code":183,"score":66,"supported":41,"title":184,"abstract":185,"tags":188},"H2",[],[186,187],"土壤","soil",[],{"code":190,"score":66,"supported":41,"title":191,"abstract":192,"tags":195},"H3",[],[193,194],"灌溉","irrigation",[],{"code":118,"score":197,"supported":41,"title":198,"abstract":199,"tags":200},5,[],[167],[167,85],{"code":202,"score":40,"supported":41,"title":203,"abstract":204,"tags":206},"F6",[],[205],"农业机器人",[],{"code":208,"score":40,"supported":41,"title":209,"abstract":210,"tags":212},"G2",[],[211],"smallholder",[],{"code":214,"score":21,"supported":14,"title":215,"abstract":216,"tags":217},"H7",[],[],[168],[219,220],"气候智慧农业 AI","农业人工智能 气候智慧农业 大语言模型 数字孪生","气候智慧农业AI-5232","10.1007\u002Fs44163-026-02540-1",{"doi":222,"openalex_id":224,"authors":225,"venue":159,"cited_by_count":128,"oa_url":156,"card":229,"direction":149,"ingested_from":151},"W7222823256",[226],{"name":227,"orcid":228},"Mostafa Farajpour","https:\u002F\u002Forcid.org\u002F0000-0003-4223-502X",{"tldr":230,"method":231,"finding":232,"direction":233,"opportunity":234},"综述农业人工智能在精准灌溉、产量预测、病虫害监测等应用及落地挑战。","文献综述，覆盖视觉Transformer、基础模型、数字孪生、多模态AI与联邦学","农业AI落地受限于跨区域验证、数据碎片化与基础设施，需兼顾可解释性与公平治理。","农业人工智能与决策模型","可研究小农户场景下跨区域可泛化、可解释且低成本的农业大模型与联邦学习部署。","2026-10-18T23:30:07.666127Z",{"id":237,"title":238,"url":239,"summary":240,"summary_zh":241,"content":8,"source_name":242,"source_url":239,"published_at":243,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":128,"score_detail":244,"sources":246,"tags":248,"primary_topic":8,"secondary_topics":252,"classification":253,"search_phrases":283,"slug":286,"view_count":21,"doi":287,"paper":288,"created_at":310},4443,"Geospatial Education at the GIS–GeoAI Interface: A PRISMA-Guided Bibliometric Mapping of a Scopus Corpus, 2002–2025","https:\u002F\u002Fdoi.org\u002F10.24114\u002Fjg.v18i2.75596","Geospatial information is used in education as disciplinary content and as a learning technology, while computational geospatial research increasingly employs artificial intelligence, machine learning, deep learning, and remote sensing. This study maps the publication performance and thematic structure of a broad Scopus corpus at the interface of geospatial technology and education from 2002 to 2025. A PRISMA-guided review and bibliometric analysis were conducted on 125 documents from 68 sources using bibliometrix and Biblioshiny in R. The corpus showed an annual publication growth rate of 10.99%, an average of 11.93 citations per document, and 2.81 authors per document. The Journal of Geography was the most productive source, while De Lázaro-Torres and De Miguel González were the most productive authors. Co-word analysis identified an educational GIS cluster, a geospatial technology and learning cluster, and a computational geospatial cluster containing GeoAI-related terms. The increased visibility of the computational cluster in 2022–2025 indicates an emerging thematic orientation, not evidence that GeoAI has replaced GIS-based education or already dominates educational practice. Because the search expression included the broad term learning and the corpus retained several technical geospatial records, the findings should be interpreted as an exploratory map of the education–technology interface. The proposed instructional sequence is therefore a theory-informed research proposition that requires empirical validation","地理空间信息在教育中既作为学科内容，也作为学习技术使用，而计算地理空间研究日益采用人工智能、机器学习、深度学习与遥感。本研究绘制了2002年至2025年间Scopus数据库中地理空间技术与教育交叉领域广泛文献的出版表现与主题结构。采用PRISMA导向的综述与文献计量分析，使用R语言中的bibliometrix和Biblioshiny对来自68个来源的125篇文献进行分析。该文献集显示年出版增长率为10.99%，篇均被引11.93次，篇均作者2.81人。《地理学杂志》（Journal of Geography）是最高产来源，而De Lázaro-Torres和De Miguel González是最高产作者。共词分析识别出教育GIS聚类、地理空间技术与学习聚类，以及包含GeoAI相关术语的计算地理空间聚类。计算聚类在2022—2025年可见度上升，表明一种新兴的主题取向，而非GeoAI已取代基于GIS的教育或已主导教育实践的证据。由于检索式包含宽泛术语“学习”，且文献集保留了一些技术性地地理空间记录，研究结果应被解释为教育—技术界面的探索性图谱。因此，所提出的教学序列是一种有理论依据的研究命题，尚需实证验证。","JURNAL GEOGRAFI","2026-10-06T00:00:00Z",{"impact":128,"substance":128,"depth":128,"authority":128,"freshness":128,"relevant":128,"comment":245},"该文为地理空间教育与GeoAI交叉领域的文献计量论文，与三农、农业信息化、智慧农业等主题无直接关联，不建议进入每日精选。",[247],{"name":242,"url":239},[27,249,250,251],"文献计量","地理信息系统","地理空间教育",[],{"status":33,"locked":14,"version":34,"method":35,"evidence":254},{"candidates":255,"agricultural_evidence":14,"reason":282},[256,263,269,277],{"code":257,"score":258,"supported":41,"title":259,"abstract":261,"tags":262},"I7",15,[260],"bibliometric",[249,260],[249],{"code":264,"score":80,"supported":41,"title":265,"abstract":267,"tags":268},"I4",[266],"GIS",[266],[250],{"code":53,"score":62,"supported":41,"title":270,"abstract":271,"tags":276},[],[272,273,274,275,177],"深度学习","机器学习","machine learning","deep learning",[],{"code":47,"score":66,"supported":41,"title":278,"abstract":279,"tags":281},[],[50,280],"remote sensing",[],"农业相关性需要复核",[284,285],"GeoAI 地理空间教育","Scopus 文献计量 地理","GeoAI地理空间教育-4443","10.24114\u002Fjg.v18i2.75596",{"doi":287,"openalex_id":289,"authors":290,"venue":242,"cited_by_count":128,"oa_url":303,"card":304,"direction":309,"ingested_from":151},"W7220698962",[291,294,297,300],{"name":292,"orcid":293},"Sony Nugratama Hijrawadi","https:\u002F\u002Forcid.org\u002F0000-0001-7694-3062",{"name":295,"orcid":296},"Epon Ningrum","https:\u002F\u002Forcid.org\u002F0000-0003-2952-8231",{"name":298,"orcid":299},"Nandi Nandi","https:\u002F\u002Forcid.org\u002F0000-0002-5994-1142",{"name":301,"orcid":302},"Lili Somantri","https:\u002F\u002Forcid.org\u002F0000-0002-6783-5948","https:\u002F\u002Fjurnal.unimed.ac.id\u002F2012\u002Findex.php\u002Fgeo\u002Farticle\u002Fdownload\u002F75596\u002F30929",{"tldr":305,"method":306,"finding":307,"direction":233,"opportunity":308},"用PRISMA与文献计量法梳理2002–2025年GIS与GeoAI教育交叉领域的研究格局。","PRISMA综述+文献计量，R的bibliometrix\u002FBiblioshiny","共词分析识别出教育GIS、地理空间技术与学习、计算地理空间三簇，GeoAI簇2022–2025年关注","可将GeoAI教育主题图谱延伸至农业信息化人才培养，检验所提教学序列的实证效果。","农业遥感与作物表型","2026-10-08T23:30:31.197022Z",{"id":312,"title":313,"url":314,"summary":315,"summary_zh":316,"content":8,"source_name":317,"source_url":314,"published_at":318,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":128,"score_detail":319,"sources":321,"tags":323,"primary_topic":8,"secondary_topics":327,"classification":328,"search_phrases":341,"slug":344,"view_count":20,"doi":345,"paper":346,"created_at":383},3859,"Large language models for social governance: An analytical framework and research agenda","https:\u002F\u002Fdoi.org\u002F10.1093\u002Fnsr\u002Fnwag602","Abstract As social governance becomes increasingly complex, traditional governance approaches face challenges such as information asymmetry, fragmented coordination, delayed risk response, and uneven public service delivery. Large language models (LLMs) offer new possibilities for knowledge processing, decision support, and human–machine interaction in this context. This article provides a comprehensive and framework-oriented review of research on LLMs, digital governance, public governance, and computational social science. It develops a governance-oriented analytical framework that links LLMs capabilities, governance functions, operating mechanisms, governance outcomes, and risk constraints. The review examines representative application scenarios, including emergency management, information propagation, intelligent rule of law, data governance, public security, healthcare, and e-government. The analysis highlights uneven evidence maturity across these scenarios: bounded tasks such as policy retrieval, public consultation, information summarization, and administrative document processing have clearer deployment pathways, whereas high-stakes simulation, legal reasoning, public security assessment, healthcare triage, and cross-departmental decision support require stronger validation and oversight. LLM-enabled social governance also remains constrained by hallucination, algorithmic bias, privacy leakage, weak explainability, resource inequality, accountability gaps, and institutional adaptation challenges. This article argues that LLMs should be understood as decision-support infrastructures embedded in accountable, auditable, and public value-oriented governance institutions.","摘要 随着社会治理日益复杂，传统治理方式面临信息不对称、协调碎片化、风险响应滞后、公共服务供给不均等挑战。大语言模型（LLMs）为此情境下的知识处理、决策支持与人机交互提供了新的可能。本文对LLMs、数字治理、公共治理与计算社会科学的研究进行了综合性、框架导向的综述，构建了一个面向治理的分析框架，将LLMs能力、治理功能、运行机制、治理结果与风险约束相联结。综述考察了代表性应用场景，包括应急管理、信息传播、智慧法治、数据治理、公共安全、医疗健康与电子政务。分析表明，这些场景的证据成熟度不均衡：政策检索、公众咨询、信息摘要与行政文书处理等边界明确的任务具有更清晰的部署路径，而高风险模拟、法律推理、公共安全评估、医疗分诊与跨部门决策支持则需要更强的验证与监督。LLMs赋能的社会治理还受到幻觉、算法偏见、隐私泄露、可解释性薄弱、资源不平等、问责缺口与制度适应挑战等制约。本文认为，应将LLMs理解为嵌入可问责、可审计且面向公共价值的治理机构之中的决策支持基础设施。","National Science Review","2026-09-27T00:00:00Z",{"impact":128,"substance":128,"depth":128,"authority":128,"freshness":128,"relevant":128,"comment":320},"该文探讨大语言模型在社会治理中的应用框架，属公共管理与计算社会科学范畴，与三农、农业信息化、智慧农业、数字乡村等主题无直接关联，不建议进入每日精选。",[322],{"name":317,"url":314},[26,324,325,326],"数字治理","社会治理","公共管理",[],{"status":33,"locked":14,"version":34,"method":35,"evidence":329},{"candidates":330,"agricultural_evidence":14,"reason":282},[331,336],{"code":53,"score":332,"supported":41,"title":333,"abstract":334,"tags":335},4,[],[113],[26],{"code":337,"score":21,"supported":14,"title":338,"abstract":339,"tags":340},"G5",[],[],[324],[342,343],"大语言模型 公共管理 数字治理 社会治理","大语言模型 公共管理","大语言模型公共管理数字治理社会治理-3859","10.1093\u002Fnsr\u002Fnwag602",{"doi":345,"openalex_id":347,"authors":348,"venue":317,"cited_by_count":128,"oa_url":376,"card":377,"direction":381,"ingested_from":151},"W7214589214",[349,351,354,355,357,359,360,363,366,368,369,370,373],{"name":350,"orcid":8},"Jun Hao",{"name":352,"orcid":353},"Jian Li","https:\u002F\u002Forcid.org\u002F0000-0002-4549-0694",{"name":352,"orcid":353},{"name":356,"orcid":8},"Linyuan Lü",{"name":358,"orcid":8},"Fang Wang",{"name":358,"orcid":8},{"name":361,"orcid":362},"Yijun Liu","https:\u002F\u002Forcid.org\u002F0009-0008-6601-9237",{"name":364,"orcid":365},"Jianjun Yu","https:\u002F\u002Forcid.org\u002F0000-0002-5203-171X",{"name":367,"orcid":8},"Changyong Liang",{"name":364,"orcid":365},{"name":367,"orcid":8},{"name":371,"orcid":372},"Xiaolong Zheng","https:\u002F\u002Forcid.org\u002F0000-0003-0405-5458",{"name":374,"orcid":375},"Zhixin Liu","https:\u002F\u002Forcid.org\u002F0000-0003-2342-4298","https:\u002F\u002Facademic.oup.com\u002Fnsr\u002Fadvance-article-pdf\u002Fdoi\u002F10.1093\u002Fnsr\u002Fnwag602\u002F71450964\u002Fnwag602.pdf",{"tldr":378,"method":379,"finding":380,"direction":381,"opportunity":382},"综述大语言模型在社会治理中的应用，构建分析框架并提出研究议程。","文献综述，整合LLM、数字治理与计算社会科学研究。","低风险任务部署路径清晰，高风险场景需验证与监管，LLM应作为决策支持基础设施。","数字乡村与农业信息化","可探索LLM在乡村治理、农业政策咨询等低风险场景的落地与评估框架。","2026-09-30T23:30:46.382519Z",{"id":385,"title":386,"url":387,"summary":388,"summary_zh":389,"content":8,"source_name":390,"source_url":387,"published_at":391,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":392,"score_detail":393,"sources":397,"tags":399,"primary_topic":402,"secondary_topics":408,"classification":409,"search_phrases":439,"slug":442,"view_count":21,"doi":443,"paper":444,"created_at":455},3665,"ASAS-NANP SYMPOSIUM: MATHEMATICAL MODELING IN ANIMAL NUTRITION: The Evolution of Large Language Models and Their Impact on Animal Sciences","https:\u002F\u002Fdoi.org\u002F10.1093\u002Fjas\u002Fskag306","The rapid rise of large language models (LLM) is reshaping the scientific landscape, transitioning from early statistical language models to advanced transformer-based architectures capable of synthesizing knowledge across disciplines. While their predictive capacity and scalability have opened new avenues in data analysis, hypothesis generation, and decision support, concerns remain regarding bias, hallucination, reproducibility, and ethical governance. In animal sciences, LLM are gradually applied to challenges in nutrition modeling, animal health, genetic selection, and sustainability. Precision nutrition has benefited from LLM-driven synthesis of feed and metabolic data, enabling individualized feeding strategies and improved resource efficiency. In animal health, domain-specific systems have demonstrated applications in diagnostics and epidemiological monitoring. LLM is augmenting genomic analyses to accelerate marker discovery and breeding optimization, while sustainability efforts employ them to model greenhouse gas emissions, feed additives, and adaptation to climatic stressors. Notably, decision-support platforms demonstrate how domain-specialized LLM can bridge mechanistic knowledge with predictive analytics, enhancing knowledge transfer and empowering livestock producers. However, risks associated with overreliance, recursive reuse of LLM outputs in model development, and pseudo-expertise underscore the critical importance of human oversight. Unlike mechanistic models, which embed biological causality, LLM are entirely data-driven and may confidently propagate errors if trained on ill-conditioned datasets. Responsible use requires transparent reporting, validation, and bias auditing, with domain-specific fine-tuning. Open-source models can enhance reproducibility and trust, but they also raise financial and security concerns. LLM must be guided by transparency, accountability, and fairness to ensure they complement, rather than replace, human expertise. By advancing inquiry and livestock management, LLM hold the potential to support sustainable food production systems if deployed responsibly. Rather than full training, most applications will rely on fine-tuning and augmentation, which are more sustainable and adaptable strategies. This review intends to promote critical literacy and responsible adoption of LLM in research, education, and livestock decision-support systems for animal scientists, nutrition modelers, and extension professionals.","大语言模型（LLM）的迅速崛起正在重塑科学格局，从早期的统计语言模型过渡到能够跨学科综合知识的先进基于Transformer的架构。尽管其预测能力和可扩展性在数据分析、假设生成和决策支持方面开辟了新途径，但围绕偏见、幻觉、可重复性和伦理治理的担忧依然存在。在动物科学领域，LLM正逐步应用于营养建模、动物健康、遗传选择和可持续性等方面的挑战。精准营养已受益于LLM驱动的饲料与代谢数据综合，从而实现个体化饲喂策略并提高资源效率。在动物健康方面，领域专用系统已在诊断和流行病学监测中展现出应用价值。LLM正在增强基因组分析，以加速标记发现和育种优化，而可持续性工作则利用其建模温室气体排放、饲料添加剂及对气候胁迫的适应性。值得注意的是，决策支持平台展示了领域专用LLM如何将机制性知识与预测分析相衔接，促进知识转化并赋能畜牧生产者。然而，过度依赖、在模型开发中递归复用LLM输出以及伪专业知识等风险，凸显了人类监督的至关重要性。与嵌入生物学因果关系的机制性模型不同，LLM完全由数据驱动，若在病态数据集上训练，可能自信地传播错误。负责任的使用需要透明报告、验证和偏见审计，并进行领域专用微调。开源模型可以增强可重复性和信任，但也引发财务和安全方面的担忧。LLM必须以透明性、问责制和公平性为指引，以确保其补充而非取代人类专业知识。通过在研究和畜牧管理中推动探究，LLM在负责任部署的前提下具有支持可持续食品生产系统的潜力。大多数应用将依赖微调和增强，而非完整训练，这是更具可持续性和适应性的策略。本综述旨在促进动物科学家、营养建模人员和推广专业人员在研究、教育和畜牧决策支持系统中对LLM的批判性素养和负责任采用。","Journal of Animal Science","2026-09-25T00:00:00Z",78,{"impact":17,"substance":394,"depth":17,"authority":19,"freshness":395,"relevant":21,"comment":396},20,8,"核心期刊综述，系统梳理大语言模型在动物营养、健康与育种中的应用与风险，专业深度高但属综述性论文，产业级影响有限。",[398],{"name":390,"url":387},[85,86,26,400,401],"精准营养","动物科学",{"code":403,"slug":404,"name":405,"path":406,"parent_code":407},"B7","protected-horticulture","设施园艺与植物工厂","\u002Ftopic\u002Fhorticulture\u002Fprotected-horticulture","B",[],{"status":101,"locked":14,"version":34,"method":35,"evidence":410},{"candidates":411,"agricultural_evidence":41},[412,418,425,429,435],{"code":403,"score":66,"supported":41,"title":413,"abstract":414,"tags":417},[],[415,416],"温室","greenhouse",[],{"code":419,"score":66,"supported":41,"title":420,"abstract":421,"tags":424},"C",[],[422,423],"畜牧","livestock",[],{"code":53,"score":197,"supported":41,"title":426,"abstract":427,"tags":428},[],[113],[86,26],{"code":430,"score":40,"supported":41,"title":431,"abstract":432,"tags":434},"C4",[],[433],"饲料",[],{"code":118,"score":21,"supported":14,"title":436,"abstract":437,"tags":438},[],[],[85],[440,441],"大语言模型 动物科学","LLM 精准营养","大语言模型动物科学-3665","10.1093\u002Fjas\u002Fskag306",{"doi":443,"openalex_id":445,"authors":446,"venue":390,"cited_by_count":128,"oa_url":449,"card":450,"direction":149,"ingested_from":151},"W7214347972",[447],{"name":448,"orcid":8},"Luis O Tedeschi","https:\u002F\u002Facademic.oup.com\u002Fjas\u002Fadvance-article-pdf\u002Fdoi\u002F10.1093\u002Fjas\u002Fskag306\u002F71398322\u002Fskag306.pdf",{"tldr":451,"method":452,"finding":453,"direction":233,"opportunity":454},"综述大语言模型在动物科学中的应用、风险与负责任使用路径。","文献综述，梳理LLM在营养、健康、遗传和可持续性中的应用。","LLM可辅助精准营养与决策支持，但需人类监督、验证和领域微调。","可探索面向畜牧生产的领域专用LLM微调与可验证决策支持系统，并建立偏差审计与可复现评估框架。","2026-09-28T23:30:12.831387Z",{"id":457,"title":458,"url":459,"summary":460,"summary_zh":461,"content":8,"source_name":462,"source_url":459,"published_at":463,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":128,"score_detail":464,"sources":466,"tags":468,"primary_topic":8,"secondary_topics":472,"classification":473,"search_phrases":490,"slug":493,"view_count":20,"doi":494,"paper":495,"created_at":512},3088,"A Hybrid COBIT 2019–RAG Framework to Enhance Consumer Protection Compliance in Digital Finance","https:\u002F\u002Fdoi.org\u002F10.26877\u002Fasset.v8i4.2088","The rapid growth of digital financial services has increased the need for effective regulatory compliance mechanisms, particularly for consumer protection. This study develops and evaluates a hybrid COBIT 2019–Retrieval-Augmented Generation (RAG) framework for compliance with Indonesia’s OJK Regulation No. 22\u002F2023 in a digital multifinance context. Using a design science research methodology, COBIT 2019 design factors were applied to tailor governance objectives, while a RAG pipeline was evaluated on 100 synthetic compliance queries using RAGAS metrics and expert-validated reference answers. The governance assessment prioritized EDM03, APO12, and MEA03, each assessed at capability Level 2 against a target of Level 4, indicating gaps in risk optimization, risk management, and external compliance. The optimal RAG configuration achieved an overall score of 0.8257, with context recall of 0.9217, faithfulness of 0.8502, and semantic similarity of 0.8629. It also outperformed baseline keyword search in retrieving semantically relevant regulatory passages, although broader benchmarking remains limited. The findings show that integrating structured IT governance with AI-assisted regulatory retrieval can strengthen accountability, regulatory interpretation, and compliance decision support. The framework offers a scalable approach for regulated industries, subject to further validation using real operational data and broader institutional settings.","数字金融服务的快速增长，提升了对有效监管合规机制的需求，尤其是在消费者保护方面。本研究开发并评估了一个融合COBIT 2019与检索增强生成（Retrieval-Augmented Generation, RAG）的混合框架，用于在数字多金融（multifinance）场景下遵守印度尼西亚OJK第22\u002F2023号条例。采用设计科学研究方法，应用COBIT 2019设计因子来定制治理目标，同时使用RAGAS指标和专家验证的参考答案，在100条合成合规查询上评估了一条RAG流水线。治理评估优先考虑EDM03、APO12和MEA03，三者均被评定为能力等级2，而目标为等级4，表明在风险优化、风险管理和外部合规方面存在差距。最优RAG配置的总体得分为0.8257，其中上下文召回率为0.9217，忠实度为0.8502，语义相似度为0.8629。在检索语义相关的监管段落方面，它也优于基线关键词搜索，尽管更广泛的基准测试仍然有限。研究结果表明，将结构化IT治理与AI辅助的监管检索相结合，可以加强问责制、监管解释和合规决策支持。该框架为受监管行业提供了一种可扩展的方法，但仍有待使用真实运营数据和更广泛的制度环境进行进一步验证。","Advance Sustainable Science Engineering and Technology","2026-09-18T00:00:00Z",{"impact":128,"substance":128,"depth":128,"authority":128,"freshness":128,"relevant":128,"comment":465},"研究数字金融消费者保护合规的IT治理与RAG框架，与三农、农业信息化无关，不予入选。",[467],{"name":462,"url":459},[29,469,470,471],"数字金融","合规科技","IT治理",[],{"status":33,"locked":14,"version":34,"method":35,"evidence":474},{"candidates":475,"agricultural_evidence":14,"reason":282},[476,480,485],{"code":53,"score":40,"supported":41,"title":477,"abstract":478,"tags":479},[],[113],[],{"code":257,"score":40,"supported":41,"title":481,"abstract":482,"tags":484},[],[483],"研究方法",[],{"code":486,"score":21,"supported":14,"title":487,"abstract":488,"tags":489},"G3",[],[],[469],[491,492],"COBIT 2019 RAG 合规","OJK 22 2023 消费者保护","COBIT2019RAG合规-3088","10.26877\u002Fasset.v8i4.2088",{"doi":494,"openalex_id":496,"authors":497,"venue":462,"cited_by_count":128,"oa_url":506,"card":507,"direction":233,"ingested_from":151},"W7213555882",[498,501,503],{"name":499,"orcid":500},"Edy Salim","https:\u002F\u002Forcid.org\u002F0009-0007-6264-0117",{"name":502,"orcid":8},"Mohammad Achmad Amin Soetomo",{"name":504,"orcid":505},"Eka Budiarto","https:\u002F\u002Forcid.org\u002F0009-0000-6642-260X","https:\u002F\u002Fjournal2.upgris.ac.id\u002Findex.php\u002Fasset\u002Farticle\u002Fdownload\u002F2088\u002F1736",{"tldr":508,"method":509,"finding":510,"direction":233,"opportunity":511},"构建COBIT 2019与RAG混合框架，提升数字金融消费者保护合规能力。","设计科学研究法，COBIT 2019设计因子与RAG管道，100条合成合规查询及","治理目标存在能力差距，最优RAG配置得分0.8257，优于关键词检索。","可将该混合治理与RAG框架迁移至农业数字金融或农业数据合规场景，探索领域适配与真实数据验证。","2026-09-21T23:30:45.181565Z"]