[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-5194":3,"related-5194":92},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":6,"published_at":11,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":24,"tags":26,"primary_topic":32,"secondary_topics":38,"classification":39,"search_phrases":65,"slug":68,"view_count":22,"doi":69,"paper":70,"created_at":91},5194,"A Comprehensive Review on Multi-Modal Large Language Models and Agentic Architectures for Precision Agricultural Decision Support","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23411181","This article reviews recent developments in multi-modal large language models and agentic architectures for agricultural decision support. Forty-five representative studies published between 2023 and 2026 were selected across perception, language, orchestration, soil intelligence and governance, organised under a four-layer taxonomy, and classified against nine functional capabilities using a stated, reproducible procedure. The comparative analysis exposes a pronounced imbalance. Crop and disease perception is covered by 36 per cent of the selected studies, while dose-bearing prescription appears in 7 per cent, profit-aware recommendation in 4 per cent, and persistent per-field memory in a single study. Advisory tools in this literature therefore remain unable to exchange state: image-based scanners are rarely conditioned on measured soil chemistry, and crop-recommendation models optimise agronomic suitability without reference to farm economics. On this basis the article synthesises a conceptual reference architecture from the reviewed literature. It is a literature-derived synthesis rather than an implemented system, and calibrated abstention, evaluation protocols for longitudinal memory, and vernacular voice coverage emerge as the most urgent gaps.","本文综述了面向农业决策支持的多模态大语言模型与智能体架构的最新进展。按照可复现的既定流程，选取2023年至2026年间发表的45项代表性研究，涵盖感知、语言、编排、土壤智能与治理等方向，将其纳入一个四层分类体系，并依据九项功能能力进行分类。比较分析揭示出明显的不均衡：作物与病害感知在所选研究中占36%，而含剂量处方的研究仅占7%，利润感知推荐占4%，持久性田块记忆仅见于一项研究。因此，该领域的咨询工具仍无法交换状态：基于图像的扫描器很少以实测土壤化学状况为条件，作物推荐模型在优化农艺适宜性时也未考虑农场经济性。在此基础上，本文综合所综述文献提出一个概念性参考架构。该架构是文献衍生的综合框架，而非已实现的系统；校准弃权、纵向记忆的评估协议以及方言语音覆盖被列为最紧迫的空白。",null,"Zenodo (CERN European Organization for Nuclear Research)","2026-10-16T00:00:00Z","论文",10,false,63,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},12,18,16,8,9,1,"系统性综述揭示处方决策与农场经济维度严重缺位，对农业AI研发方向有参考价值，但属文献综述而非实证突破，影响范围有限。",[25],{"name":10,"url":6},[27,28,29,30,31],"智慧农业","农业人工智能","智能体","多模态大模型","农业决策支持",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":40,"locked":14,"version":41,"method":42,"evidence":43},"classified","2026-10-22.2","rules",{"candidates":44,"agricultural_evidence":47},[45,52,60],{"code":33,"score":46,"supported":47,"title":48,"abstract":50,"tags":51},14,true,[49],"decision support",[49],[28,30,29],{"code":53,"score":54,"supported":47,"title":55,"abstract":56,"tags":59},"H2",6,[],[57,58],"土壤","soil",[],{"code":61,"score":22,"supported":14,"title":62,"abstract":63,"tags":64},"I1",[],[],[27],[66,67],"多模态大模型 农业决策支持","农业人工智能 农业决策支持 多模态大模型 智慧农业","多模态大模型农业决策支持-5194","10.5281\u002Fzenodo.23411181",{"doi":69,"openalex_id":71,"authors":72,"venue":10,"cited_by_count":83,"oa_url":6,"card":84,"direction":88,"ingested_from":90},"W7222605322",[73,75,77,79,81],{"name":74,"orcid":9},"Tanmay Bhamre",{"name":76,"orcid":9},"Yash Barde",{"name":78,"orcid":9},"Devendra Yeole",{"name":80,"orcid":9},"Gajanan Chaudhary",{"name":82,"orcid":9},"Amarsinh Verpe",0,{"tldr":85,"method":86,"finding":87,"direction":88,"opportunity":89},"综述多模态大模型与智能体架构在精准农业决策支持中的应用与不足。","筛选2023-2026年45篇代表性研究，按四层分类与九项功能能力归类比较。","感知研究占36%，处方仅7%、利润推荐4%、田块记忆仅1篇，系统难以交换状态。","农业人工智能与决策模型","可研究跨模态状态交换、纵向记忆评测协议与方言语音覆盖，填补处方与经济决策空白。","openalex","2026-10-17T23:30:55.282694Z",{"total":54,"page":22,"page_size":54,"items":93},[94,161,199,257,313,356],{"id":95,"title":96,"url":97,"summary":98,"summary_zh":99,"content":9,"source_name":100,"source_url":97,"published_at":101,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":102,"score_detail":103,"sources":106,"tags":108,"primary_topic":111,"secondary_topics":112,"classification":113,"search_phrases":126,"slug":129,"view_count":83,"doi":130,"paper":131,"created_at":160},5562,"pxGPT: A schema-constrained multimodal LLM framework for plant phenotyping","https:\u002F\u002Fdoi.org\u002F10.1093\u002Fbioinformatics\u002Fbtag781","Abstract Summary Deep learning can scale image-based plant phenotyping but typically targets predefined traits, whereas manual phenotypic description is more flexible but difficult to scale. We present Phenotype eXplorer Generative Plant Traits (pxGPT), a Python framework that converts multi-view plant images into structured phenotype records. The workflow combines free-text description, dataset-specific schema generation with human review, and schema-constrained information extraction. We evaluated the extraction stage on three Brassica datasets using 78,973 annotations from ten annotators who had no access to pxGPT predictions. For 38-45 qualitative traits per dataset, pxGPT-human agreement ranged from 0.572 to 0.712, compared with 0.620 to 0.739 agreement among human annotators. All 32,736 required trait fields were present; 93 values were empty, and no qualitative descriptor violations were recorded. For selected morphological structures, pooled false-positive rates were 0.5-8.0%, while sensitivity varied across structure types in one mature-plant dataset (37.5-100%). The curated schemas were applied unchanged with two commercial models and one locally hosted open-weight model, with differences in mean accuracy of 0.03 or less. In the evaluated Brassica datasets, these results support schema-guided extraction of visible plant traits without separate model training for each trait. Human-guided schema development offers a path to extend pxGPT to other crops and imaging conditions, with accuracy requiring validation in each new setting. Availability and Implementation The software is available at https:\u002F\u002Fgithub.com\u002FYaoChengLab\u002Fpxgpt and an archival version of the source code is available on Zenodo (https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.22931745). Curated schemas used in this study, annotation guides, phenotype records and human annotations are available at https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.22907554. Supplementary information Supplementary data are available at Bioinformatics online.","摘要 深度学习能够扩展基于图像的植物表型分析，但通常针对预定义性状，而人工表型描述更为灵活却难以规模化。我们提出Phenotype eXplorer Generative Plant Traits（pxGPT），一个将多视角植物图像转换为结构化表型记录的Python框架。该工作流程结合了自由文本描述、数据集特定模式生成与人工审核，以及模式约束的信息提取。我们在三个芸薹属（Brassica）数据集上评估了提取阶段，使用了来自十名标注者的78,973条标注，这些标注者无法获取pxGPT的预测结果。对于每个数据集38-45个定性性状，pxGPT与人工标注者之间的一致性范围为0.572至0.712，而人工标注者之间的一致性为0.620至0.739。所有32,736个所需性状字段均已存在；93个值为空，未记录到定性描述符违规。对于选定的形态结构，合并假阳性率为0.5-8.0%，而在一个成熟植株数据集中，灵敏度因结构类型而异（37.5-100%）。经整理的模式在未作修改的情况下应用于两个商业模型和一个本地部署的开放权重模型，平均准确率差异不超过0.03。在所评估的芸薹属数据集中，这些结果支持在无需为每个性状单独训练模型的情况下，对可见植物性状进行模式引导的提取。人工引导的模式开发为将pxGPT扩展至其他作物和成像条件提供了途径，其准确性需在每个新设置中加以验证。可用性与实现 该软件可在https:\u002F\u002Fgithub.com\u002FYaoChengLab\u002Fpxgpt获取，源代码的存档版本可在Zenodo（https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.22931745）获取。本研究中使用的经整理模式、标注指南、表型记录和人工标注可在https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.22907554获取。补充信息 补充数据可在Bioinformatics在线获取。","Bioinformatics","2026-10-20T00:00:00Z",81,{"impact":18,"substance":104,"depth":18,"authority":46,"freshness":21,"relevant":22,"comment":105},22,"提出基于模式约束的多模态大模型植物表型提取框架，在近8万条标注上验证，方法新颖、数据规模可观，对智慧育种与表型自动化有参考价值。",[107],{"name":100,"url":97},[27,28,109,30,110],"植物表型","芸薹属",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":114},{"candidates":115,"agricultural_evidence":47},[116,122],{"code":33,"score":20,"supported":47,"title":117,"abstract":118,"tags":121},[],[119,120],"深度学习","deep learning",[28,30],{"code":61,"score":22,"supported":14,"title":123,"abstract":124,"tags":125},[],[],[27],[127,128],"pxGPT 植物表型","Brassica 表型 多模态大模型","pxGPT植物表型-5562","10.1093\u002Fbioinformatics\u002Fbtag781",{"doi":130,"openalex_id":132,"authors":133,"venue":100,"cited_by_count":83,"oa_url":152,"card":153,"direction":159,"ingested_from":90},"W7223127960",[134,137,139,142,145,148,150],{"name":135,"orcid":136},"Po‐Xing Zheng","https:\u002F\u002Forcid.org\u002F0000-0001-5827-5865",{"name":138,"orcid":9},"Jheng-Yan Ou",{"name":140,"orcid":141},"Yu-Lin Wu","https:\u002F\u002Forcid.org\u002F0009-0003-8596-9047",{"name":143,"orcid":144},"Ruei-Shiuan Wang","https:\u002F\u002Forcid.org\u002F0009-0007-5159-1129",{"name":146,"orcid":147},"WU Ting-li","https:\u002F\u002Forcid.org\u002F0000-0001-7996-5515",{"name":149,"orcid":9},"Te-Chang Hsu",{"name":151,"orcid":9},"Yao-Cheng Lin","https:\u002F\u002Facademic.oup.com\u002Fbioinformatics\u002Fadvance-article-pdf\u002Fdoi\u002F10.1093\u002Fbioinformatics\u002Fbtag781\u002F72201337\u002Fbtag781.pdf",{"tldr":154,"method":155,"finding":156,"direction":157,"opportunity":158},"提出pxGPT框架，用多模态大模型将多视角植物图像转为结构化表型记录。","结合自由文本描述、人工审核的模式生成与模式约束信息抽取，在三个芸薹属数据集上评估","pxGPT与人工标注一致性0.572-0.712，接近人工间一致性，且无需逐性状训练模型。","农业遥感与作物表型","可探索将模式引导抽取迁移至更多作物与复杂成像条件，并研究自动化模式生成以降低人工审核成本。","智慧农业 \u002F 农业物联网","2026-10-22T11:20:22.173561Z",{"id":162,"title":163,"url":164,"summary":165,"summary_zh":9,"content":9,"source_name":166,"source_url":9,"published_at":167,"category":168,"cover_url":9,"hotness":13,"is_selected":14,"score":169,"score_detail":170,"sources":174,"tags":176,"primary_topic":179,"secondary_topics":180,"classification":181,"search_phrases":194,"slug":197,"view_count":22,"doi":9,"paper":9,"created_at":198},4933,"中国农科院信息所提出层次知识引导枸杞害虫识别框架","https:\u002F\u002Fnews.qq.com\u002Frain\u002Fa\u002F20261013A07CL800","中国农业科学院农业信息研究所科学数据研究室提出一种层次知识引导的枸杞害虫识别框架,有效解决真实农业场景下样本稀缺与类间混淆的双重挑战。相关研究成果发表在《人工智能的工程应用》(Engineering Applications of Artificial Intelligence)上。该研究依托长文本-图像检索模型,把专家撰写的长文本描述解析为粗、中、细三个语义层次,并与多尺度视觉特征实现精准跨模态对齐,为枸杞等特色作物的虫害智能监测与精准防控提供新的技术路径。","中国农业科学院农业信息研究所 腾讯新闻 2026 年 10 月 13 日","2026-10-13T00:00:00Z","报道",78,{"impact":18,"substance":171,"depth":172,"authority":46,"freshness":20,"relevant":22,"comment":173},21,17,"国家级科研机构在农业AI领域的方法创新，跨模态对齐思路对特色作物虫害监测有推广价值，值得进入每日精选。",[175],{"name":166,"url":164},[27,28,30,177,178],"害虫识别","枸杞",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":182},{"candidates":183,"agricultural_evidence":47},[184,190],{"code":33,"score":185,"supported":47,"title":186,"abstract":187,"tags":189},5,[],[188],"artificial intelligence",[28,30],{"code":61,"score":22,"supported":14,"title":191,"abstract":192,"tags":193},[],[],[27],[195,196],"中国农科院信息所 枸杞害虫识别","枸杞 害虫识别 视觉语言模型","中国农科院信息所枸杞害虫识别-4933","2026-10-15T00:06:01.623759Z",{"id":200,"title":201,"url":202,"summary":203,"summary_zh":204,"content":9,"source_name":205,"source_url":202,"published_at":206,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":102,"score_detail":207,"sources":209,"tags":211,"primary_topic":214,"secondary_topics":215,"classification":216,"search_phrases":231,"slug":234,"view_count":83,"doi":235,"paper":236,"created_at":256},4695,"AgroNLP-SimAssist: a multilingual transformer-based decision support simulator for crop disease diagnosis, advisory generation, and market intelligence","https:\u002F\u002Fdoi.org\u002F10.1186\u002Fs12870-026-10040-8","Abstract There are three major shortcomings of agricultural NLP and need addressing, regardless of the recent developments in the field: (i) use of proprietary or small-scale datasets that affect reproducibility; (ii) one-task models that only focus on intent recognition but lack integration for crop disease diagnosis, advice, and market information; and (iii) the absence of simulators for evaluating the system’s performance during actual deployment. The application of precision agriculture is constrained by limitations in diverse farming environments. The study introduces AgroNLP-SimAssist, which represents a unified multilingual transformer architecture augmented by an easily deployable agricultural decision support simulator, as the proposed solution to overcome the challenges. The five major components constituting the proposed framework are: (1) domain-specific multilingual transformer embedding; (2) data collection and cleaning from publicly available agricultural datasets; (3) multiple prediction tasks including trend price estimation, crop disease detection, and advisory recommendation; (4) SHAP value explainability for expert validation; and (5) advisory simulation engine for evaluating latency and scalability. Multilingually available data sources, which include multilingual agricultural reports from the FAO organization, crop-disease descriptions from the PlantVillage knowledge base, and price-market narratives from the Agmarknet website, have been used for training and validation purposes. Among the 3,200 multilingual farmer queries dataset (in English, Hindi, and Kannada), there was an assignment of 70%, 15%, and 15% to the training, validation, and testing sets, respectively. In terms of accuracy compared to the existing classifiers, results from experiments yielded 96.1% disease classification, 94.3% advisory relevance, and 92.7% named entity recognition with F1-measure. The simulation system scaled up to 10,000 concurrent users with an inference latency of 0.35 s\u002Fquery. The proposed data-driven AgroNLP-SimAssist framework facilitates extension services and precision agriculture through providing a scalable, reproducible, and interpretable solution for intelligent agricultural advisory systems.","尽管近年来农业自然语言处理（NLP）领域取得了进展，但仍存在三个主要不足亟需解决：(i) 使用专有或小规模数据集，影响可复现性；(ii) 单任务模型仅关注意图识别，缺乏对作物病害诊断、建议和市场信息的整合；(iii) 缺乏用于评估系统实际部署性能的模拟器。精准农业的应用受到多样化农业环境的限制。本研究提出了AgroNLP-SimAssist，一种统一的多语言Transformer架构，并辅以易于部署的农业决策支持模拟器，作为克服上述挑战的解决方案。该框架由五个主要组件构成：(1) 领域特定的多语言Transformer嵌入；(2) 从公开农业数据集中进行数据收集与清洗；(3) 多预测任务，包括趋势价格估计、作物病害检测和咨询推荐；(4) SHAP值可解释性用于专家验证；(5) 咨询模拟引擎用于评估延迟和可扩展性。多语言可用数据源已用于训练和验证，包括联合国粮农组织（FAO）的多语言农业报告、PlantVillage知识库的作物病害描述以及Agmarknet网站的价格市场叙述。在包含3200条多语言农民查询数据集（英语、印地语和卡纳达语）中，训练集、验证集和测试集的分配比例分别为70%、15%和15%。与现有分类器相比，实验结果显示病害分类准确率达96.1%，咨询相关性达94.3%，命名实体识别的F1值为92.7%。模拟系统可扩展至10000名并发用户，推理延迟为0.35秒\u002F查询。所提出的数据驱动AgroNLP-SimAssist框架通过为智能农业咨询系统提供可扩展、可复现且可解释的解决方案，促进了推广服务和精准农业的发展。","BMC Plant Biology","2026-10-10T00:00:00Z",{"impact":18,"substance":104,"depth":18,"authority":46,"freshness":21,"relevant":22,"comment":208},"多语言农业大模型整合病害诊断、农技咨询与市场情报，数据规模与仿真验证扎实，对智慧农业信息化有参考价值。",[210],{"name":205,"url":202},[27,28,212,31,213],"农业大模型","作物病害诊断",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":217},{"candidates":218,"agricultural_evidence":47},[219,224],{"code":33,"score":220,"supported":47,"title":221,"abstract":222,"tags":223},13,[49],[49],[28,212],{"code":61,"score":225,"supported":47,"title":226,"abstract":227,"tags":230},7,[],[228,229],"精准农业","precision agriculture",[27],[232,233],"AgroNLP-SimAssist 作物病害诊断","多语言农业大模型 农业决策支持","AgroNLP-SimAssist作物病害诊断-4695","10.1186\u002Fs12870-026-10040-8",{"doi":235,"openalex_id":237,"authors":238,"venue":205,"cited_by_count":83,"oa_url":202,"card":251,"direction":88,"ingested_from":90},"W7221339938",[239,241,243,245,247,249],{"name":240,"orcid":9},"Priya Thirugnanam",{"name":242,"orcid":9},"Sandeep kumar Mathivanan",{"name":244,"orcid":9},"Sundara Kumar M R",{"name":246,"orcid":9},"Sangeetha Ramaswamy",{"name":248,"orcid":9},"Shruthi T S",{"name":250,"orcid":9},"Sangeetha S.K.B",{"tldr":252,"method":253,"finding":254,"direction":88,"opportunity":255},"提出多语言Transformer农业决策支持模拟器，统一实现作物病害诊断、农技建议与市场情报。","基于FAO、PlantVillage、Agmarknet多语数据训练Transf","病害分类96.1%、建议相关性94.3%、NER F1 92.7%，万级并发延迟0.35秒。","可扩展至更多语种与作物，并引入真实田间部署反馈闭环优化模拟器。","2026-10-11T23:30:56.064196Z",{"id":258,"title":259,"url":260,"summary":261,"summary_zh":262,"content":9,"source_name":10,"source_url":260,"published_at":206,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":263,"score_detail":264,"sources":266,"tags":268,"primary_topic":271,"secondary_topics":272,"classification":273,"search_phrases":297,"slug":300,"view_count":22,"doi":301,"paper":302,"created_at":312},4657,"Agriculture and Farmer Digital Assistant: An AI-Based Digital Decision Support Platform for Smart Farming","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23278187","Abstract Farmers frequently make decisions under uncertain weather, incomplete agricultural information, changing soil conditions, and limited access to timely expert guidance. The Agriculture and Farmer Digital Assistant (AFDA) is a web-based digital decision-support platform designed to bring agricultural data, artificial intelligence, recommendation services, weather information, and role-based administration into a single system. The platform accepts agricultural inputs from farmers and processes them to provide crop recommendations and basic farming guidance. The design also considers soil analysis, pest and disease support, market information, notifications, multilingual assistance, and future sensor integration. The implemented project version focuses on user registration and login, agricultural data input, crop recommendation, basic weather guidance, and a recommendation dashboard, while several advanced capabilities remain future extensions. A layered architecture separates the user interface, application logic, AI processing, database, and notification\u002Frecommendation functions. The project further defines functional and non-functional requirements, data-flow and entity relationships, user workflows, testing procedures, and maintenance considerations. The reported evaluation is functional rather than numerical: the major input, authentication, recommendation, navigation, and output-display workflows were checked during development, while no measured accuracy or response-time benchmark is claimed. AFDA therefore provides a foundation for a connected and extensible agricultural assistance system intended to support better-informed and more sustainable farming decisions. Keywords Agriculture and Farmer Digital Assistant, AFDA, artificial intelligence, machine learning, crop recommendation, smart farming, agricultural decision support, weather guidance, role-based access control.","农户常常在天气不确定、农业信息不完整、土壤条件不断变化以及难以及时获得专家指导的情况下作出决策。农业与农户数字助手（Agriculture and Farmer Digital Assistant，AFDA）是一个基于网络的数字决策支持平台，旨在将农业数据、人工智能、推荐服务、天气信息和基于角色的管理整合到一个系统中。该平台接收农户输入的农业数据，并对其进行处理，以提供作物推荐和基本农耕指导。其设计还考虑了土壤分析、病虫害支持、市场信息、通知、多语言辅助以及未来的传感器集成。已实现的项目版本侧重于用户注册与登录、农业数据输入、作物推荐、基本天气指导以及推荐仪表板，而若干高级功能仍留作未来扩展。分层架构将用户界面、应用逻辑、人工智能处理、数据库以及通知\u002F推荐功能相互分离。该项目进一步定义了功能性与非功能性需求、数据流与实体关系、用户工作流程、测试程序以及维护方面的考虑。所报告的评估是功能性的而非数值性的：在开发过程中检查了主要输入、身份验证、推荐、导航和输出显示等工作流程，但未声称有任何经过测量的准确性或响应时间基准。因此，AFDA为一个互联且可扩展的农业辅助系统提供了基础，旨在支持更明智且更可持续的农耕决策。关键词：农业与农户数字助手，AFDA，人工智能，机器学习，作物推荐，智慧农业，农业决策支持，天气指导，基于角色的访问控制。",54,{"impact":20,"substance":46,"depth":17,"authority":13,"freshness":13,"relevant":22,"comment":265},"面向农户的AI决策支持平台论文，架构与需求设计较完整，但仅做功能验证、无精度或性能数据，属早期原型，参考价值有限。",[267],{"name":10,"url":260},[269,27,28,270,31],"数字农业","作物推荐",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":274},{"candidates":275,"agricultural_evidence":47},[276,282,286,290],{"code":33,"score":18,"supported":47,"title":277,"abstract":278,"tags":281},[49],[279,280,188,49],"机器学习","machine learning",[28],{"code":53,"score":54,"supported":47,"title":283,"abstract":284,"tags":285},[],[57,58],[],{"code":61,"score":185,"supported":47,"title":287,"abstract":288,"tags":289},[],[27],[269,27],{"code":291,"score":292,"supported":47,"title":293,"abstract":294,"tags":296},"I6",3,[],[295],"agricultural data",[],[298,299],"AFDA 农业数字助手","农业人工智能 农业决策支持 作物推荐 数字农业","AFDA农业数字助手-4657","10.5281\u002Fzenodo.23278187",{"doi":301,"openalex_id":303,"authors":304,"venue":10,"cited_by_count":83,"oa_url":260,"card":307,"direction":159,"ingested_from":90},"W7221331787",[305],{"name":306,"orcid":9},"B. Siddhu, G. Naveen, S. Guru Charan, T. Ajay, C. Naresh",{"tldr":308,"method":309,"finding":310,"direction":88,"opportunity":311},"开发了基于AI的农业数字决策支持平台AFDA，为农民提供作物推荐和农事指导。","采用分层架构的Web平台，集成AI处理、数据库和推荐服务，支持多语言。","平台实现了用户认证、数据输入、作物推荐和天气指导等核心功能，但未提供精度或响应时间基准。","可补充传感器实时数据集成与模型精度评估，提升决策支持的可靠性和实时性。","2026-10-11T23:30:07.538409Z",{"id":314,"title":315,"url":316,"summary":317,"summary_zh":9,"content":9,"source_name":318,"source_url":9,"published_at":319,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":320,"score_detail":321,"sources":324,"tags":326,"primary_topic":327,"secondary_topics":328,"classification":329,"search_phrases":342,"slug":345,"view_count":346,"doi":9,"paper":347,"created_at":355},3439,"《Agentic Artificial Intelligence in Agriculture: A Systematic Mapping Review of Reported Architectures, Applications, Challenges, and Future Directions》","https:\u002F\u002Fwww.mdpi.com\u002F2227-7080\u002F14\u002F9\u002F591","作者按PRISMA 2020方案从4111项记录筛选至181项研究，对2020—2026年农业代理式人工智能文献做系统映射综述：领域跨度30多年但LLM子集非常年轻（2024年才出现、46\u002F47项发表于2025—2026年）；82%研究报告合作能力，计划和推理分别仅31%、记忆6%、反思4%；只有33项研究报告现场或实际部署，68项仍停留在概念性阶段，仅11项报告了一个季度以上的评估。","《Technologies》2026, 14(9), 591 \u002F MDPI","2026-09-22T00:00:00Z",80,{"impact":18,"substance":322,"depth":18,"authority":220,"freshness":20,"relevant":22,"comment":323},23,"基于PRISMA的农业代理式AI系统映射综述，量化揭示LLM应用年轻化与落地不足，信息增量与专业深度突出，值得进入每日精选。",[325],{"name":318,"url":316},[269,27,28,212,29],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":330},{"candidates":331,"agricultural_evidence":47},[332,337],{"code":33,"score":333,"supported":47,"title":334,"abstract":335,"tags":336},11,[188],[],[28,212,29],{"code":61,"score":338,"supported":14,"title":339,"abstract":340,"tags":341},2,[],[],[269,27],[343,344],"农业代理式人工智能 系统映射综述","农业人工智能 农业大模型 数字农业 智慧农业","农业代理式人工智能系统映射综述-3439",4,{"doi":9,"openalex_id":9,"authors":348,"venue":9,"cited_by_count":83,"oa_url":9,"card":349,"direction":88,"ingested_from":354},[],{"tldr":350,"method":351,"finding":352,"direction":88,"opportunity":353},"系统映射181项研究，梳理农业代理式AI的架构、应用、挑战与未来方向。","按PRISMA 2020筛选4111项记录至181项，做系统映射综述。","LLM代理2024年才出现，多具合作能力但规划、记忆、反思薄弱，实际部署少。","农业LLM代理的长期田间部署、记忆与反思机制及跨季度评估仍是明显空白。","agent","2026-09-25T00:09:33.995675Z",{"id":357,"title":358,"url":359,"summary":360,"summary_zh":361,"content":9,"source_name":10,"source_url":359,"published_at":362,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":363,"score_detail":364,"sources":366,"tags":368,"primary_topic":372,"secondary_topics":378,"classification":380,"search_phrases":397,"slug":400,"view_count":83,"doi":401,"paper":402,"created_at":416},5592,"Evaluation of Concept-Rank Collapse Audits for Computer Vision Agriculture Disease Classifiers Under Low-Label Regimes: An Interpretability Capacity Bound for Smallholder Fungicide Selectivity","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23545193","Concept-based interpretability methods for plant-disease classifiers are typically evaluated by whether their concept activations correlate with human-annotated labels, but this criterion is silent on a prior question: whether the classifier's internal representation retains enough concept-rank structure for any audit to be informative. the present analysis formalises that prior question as an interpretability capacity bound. A classifier is modelled as a concept-rank operator whose singular values govern how much concept-level signal survives the projection from pixel space to class logits, and a collapse audit is defined as the ratio of retained concept-rank energy to total concept energy. Under a low-label regime, where concept annotations are scarce and the audit must be estimated from few labelled examples, the bound predicts that audit reliability degrades as a function of the concept-rank spectrum and the label budget. The analysis is a theoretical and methodological contribution: it derives the bound, states the conditions under which a collapse audit is identifiable, and characterises the operating characteristics of the audit as a decision procedure under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64). Power was estimated across 2000 design cells for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points; 52 clusters were required for 5 percentage points at 80% power. The bound reframes concept audits as capacity tests rather than fidelity tests, with consequences for how smallholder fungicide-selectivity decisions should be supported by interpretable disease classifiers.","面向植物病害分类器的基于概念的可解释性方法，通常以其概念激活是否与人工标注标签相关作为评价标准，但该标准对一个更先前的问题保持沉默：分类器的内部表征是否保留了足够的概念秩结构，从而使任何审计具有信息量。本分析将该先前问题形式化为一个可解释性容量上界。分类器被建模为概念秩算子，其奇异值支配着从像素空间到类别logits的投影中概念级信号得以保留的程度，并将坍缩审计定义为保留的概念秩能量与总概念能量之比。在低标签情形下，概念标注稀缺且审计必须由少量标注样本估计，该上界预测审计可靠性会随概念秩谱和标签预算的函数而退化。本分析是一项理论和方法论贡献：推导出该上界，陈述坍缩审计可识别的条件，并在声明的方差结构（簇间SD = 6.0，残差SD = 4.5，组内相关系数 = 0.64）下刻画审计作为决策程序的操作特征。在2000个设计单元中，针对簇数为8、16、24和32以及效应量为3、5和8个百分点的情形估计了功效；在80%功效下达到5个百分点需要52个簇。该上界将概念审计重新框定为容量检验而非保真度检验，并对可解释病害分类器应如何支持小农杀菌剂选择性决策具有启示意义。","2026-10-21T00:00:00Z",69,{"impact":20,"substance":171,"depth":18,"authority":17,"freshness":13,"relevant":22,"comment":365},"理论方法论文，提出可解释性容量上界并给出审计功效分析，对智慧农业病害识别可信度有参考价值，但属小众方法学研究、缺乏田间实证，暂不宜进入每日精选。",[367],{"name":10,"url":359},[27,28,369,370,371],"小农户","植物病害识别","可解释性",{"code":373,"slug":374,"name":375,"path":376,"parent_code":377},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services","G",[379],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":381},{"candidates":382,"agricultural_evidence":47},[383,388,393],{"code":373,"score":17,"supported":47,"title":384,"abstract":386,"tags":387},[385],"smallholder",[385],[369],{"code":33,"score":13,"supported":47,"title":389,"abstract":391,"tags":392},[390],"computer vision",[],[28,370],{"code":61,"score":22,"supported":14,"title":394,"abstract":395,"tags":396},[],[],[27],[398,399],"农业人工智能 植物病害识别 可解释性 智慧农业","农业人工智能 植物病害识别","农业人工智能植物病害识别可解释性智慧农业-5592","10.5281\u002Fzenodo.23545193",{"doi":401,"openalex_id":403,"authors":404,"venue":10,"cited_by_count":83,"oa_url":359,"card":411,"direction":88,"ingested_from":90},"W7223137323",[405,407,409],{"name":406,"orcid":9},"Akshay Ramgoolam",{"name":408,"orcid":9},"Priya Bundhoo Appadoo",{"name":410,"orcid":9},"Jean-Noël LEspérance",{"tldr":412,"method":413,"finding":414,"direction":88,"opportunity":415},"提出可解释性容量上界，评估低标签下概念秩崩溃审计的可靠性。","将分类器建模为概念秩算子，推导容量上界并做2000格功效模拟。","审计可靠性随概念秩谱与标签预算下降；5个百分点效应需52个聚类达80%功效。","可将容量上界用于小农户杀菌剂选择性决策的可解释模型审计与标签预算设计。","2026-10-22T11:21:10.425290Z"]