[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-1229":3,"related-1229":89},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":8,"source_name":9,"source_url":6,"published_at":10,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":15,"sources":22,"tags":24,"primary_topic":8,"secondary_topics":30,"classification":31,"search_phrases":67,"slug":70,"view_count":71,"doi":72,"paper":73,"created_at":88},1229,"AI-integrated smart biosensors for real-time multi-contaminant detection in food systems","https:\u002F\u002Fdoi.org\u002F10.3389\u002Ffsufs.2026.1905105","Food safety systems are under growing pressure and threat due to the presence of multiple chemical and biological contaminants across complex food networks. Conventional detection techniques remain time-consuming, lab-intensive, and single-target-oriented, with limited reliable, real-time data for rapid decision-making. Recent biosensor advancements, driven by artificial intelligence (AI), machine learning (ML), and the integration of the Internet of Things (IoT), have enabled smarter on-field deployment and greater accuracy. The current mini-review summarises recent advancements in AI-integrated biosensors, such as multi-analyte sensing designs, paper-based platforms, smartphone-assisted detection, and nanozyme-based sensors. It critically examines their potential to enable rapid, cost-effective, and scalable monitoring with an emphasis on AI-enabled signal resolution and representative detection performance across food matrices. The challenges associated with multi-contaminant detection, data integration, model reliability, and practical implementation in resource-constrained environments are given special attention. The review proposes a conceptual framework for intelligent biosensing systems that integrate sensing, data analytics, and decision support. This work demonstrates the transition from prototype-level sensors to reliable, real-time food safety monitoring systems by integrating technological innovation with sustainability and policy considerations.",null,"Frontiers in Sustainable Food Systems","2026-08-27T00:00:00Z","论文",10,false,73,{"impact":16,"substance":17,"depth":16,"authority":18,"freshness":19,"relevant":20,"comment":21},18,20,12,5,1,"综述AI集成生物传感器在食品多污染物实时检测中的进展，提出智能传感框架，对农业食品安全监测有参考价值。",[23],{"name":9,"url":6},[25,26,27,28,29],"智慧农业","农业人工智能","实时监测","食品安全","生物传感器",[],{"status":32,"locked":13,"version":33,"method":34,"evidence":35},"needs_review","2026-10-22.2","rules",{"candidates":36,"agricultural_evidence":13,"reason":66},[37,46,54,61],{"code":38,"score":12,"supported":39,"title":40,"abstract":41,"tags":45},"I2",true,[],[42,43,44],"machine learning","artificial intelligence","decision support",[26],{"code":47,"score":48,"supported":39,"title":49,"abstract":50,"tags":53},"I5",6,[],[51,52],"internet of things","IoT",[],{"code":55,"score":56,"supported":39,"title":57,"abstract":58,"tags":60},"E4",4,[],[59],"food safety",[28],{"code":62,"score":20,"supported":13,"title":63,"abstract":64,"tags":65},"I1",[],[],[25],"农业相关性需要复核",[68,69],"农业人工智能 生物传感器 实时监测 智慧农业","农业人工智能 生物传感器","农业人工智能生物传感器实时监测智慧农业-1229",3,"10.3389\u002Ffsufs.2026.1905105",{"doi":72,"openalex_id":74,"authors":75,"venue":9,"cited_by_count":79,"oa_url":80,"card":81,"direction":85,"ingested_from":87},"W7204445129",[76],{"name":77,"orcid":78},"Mehak Puri","https:\u002F\u002Forcid.org\u002F0000-0002-5267-9217",0,"https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fsustainable-food-systems\u002Farticles\u002F10.3389\u002Ffsufs.2026.1905105\u002Fpdf",{"tldr":82,"method":83,"finding":84,"direction":85,"opportunity":86},"综述AI集成智能生物传感器在食品多污染物实时检测中的进展与挑战。","综述多分析物传感、纸基平台、智能手机辅助及纳米酶传感器，结合AI\u002FML\u002FIoT。","AI集成生物传感器可实现快速、经济、可扩展的多污染物检测，但需解决数据集成和模型可靠性问题。","农业人工智能与决策模型","可探索AI生物传感器在农业现场多残留检测的实时决策支持系统，结合边缘计算与轻量化模型。","openalex","2026-09-01T04:03:08.761098Z",{"total":48,"page":20,"page_size":48,"items":90},[91,189,251,313,375,438],{"id":92,"title":93,"url":94,"summary":95,"summary_zh":96,"content":8,"source_name":97,"source_url":94,"published_at":98,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":99,"score_detail":100,"sources":105,"tags":107,"primary_topic":110,"secondary_topics":115,"classification":121,"search_phrases":162,"slug":165,"view_count":79,"doi":166,"paper":167,"created_at":188},5400,"Artificial intelligence in food safety: a bibliometric and empirical synthesis of emerging technologies and research frontiers","https:\u002F\u002Fdoi.org\u002F10.1108\u002Fbfj-06-2025-0819","Purpose This study aimed to identify recent research trends in the integration of artificial intelligence (AI) with food safety and security through a comprehensive bibliometric analysis. Design\u002Fmethodology\u002Fapproach A total of 2,782 documents published between 2021 and 2025 were retrieved from the Web of Science Core Collection. The dataset was refined by removing duplicates, filtering document types and excluding irrelevant records. VOSviewer was then used to conduct co-authorship, keyword co-occurrence, citation, co-citation and bibliographic coupling analyses. Findings Machine learning and deep learning emerged as core technologies associated with food security, contamination detection and quality prediction. Keyword clustering demonstrated interdisciplinary connections with agriculture, climate change and public health. Citation and co-citation analyses identified three major intellectual foundations: traditional machine learning, computer vision-based deep learning and time-series modeling. China, India and the United States were the most influential contributors in terms of research output and technological development. Bibliographic coupling and source co-citation analyses further indicated increasing convergence between food science and computational science. Originality\u002Fvalue This study provides a comprehensive overview of the intellectual structure, collaborative networks and thematic evolution of AI-driven food safety research. Unlike previous studies focusing on specific technologies or applications, it maps global research patterns and interdisciplinary linkages. The findings provide strategic insights for researchers, policymakers and industry stakeholders. Future studies should prioritize explainable AI integration into regulatory frameworks and the development of larger, more diverse datasets to improve model generalizability across food systems.","目的 本研究旨在通过全面的文献计量分析，识别人工智能（AI）与食品安全及保障融合领域的最新研究趋势。设计与方法 从Web of Science核心合集中检索了2021年至2025年间发表的2，782篇文献。通过去除重复记录、筛选文献类型并排除不相关记录对数据集进行精炼。随后使用VOSviewer进行合著、关键词共现、引文、共被引及文献耦合分析。研究发现 机器学习和深度学习成为与食品安全、污染检测和质量预测相关的核心技术。关键词聚类显示了与农业、气候变化和公共卫生的跨学科联系。引文和共被引分析识别出三大主要知识基础：传统机器学习、基于计算机视觉的深度学习和时间序列建模。中国、印度和美国在研究产出和技术发展方面是最具影响力的贡献者。文献耦合和来源共被引分析进一步表明食品科学与计算科学之间的融合日益加深。原创性\u002F价值 本研究全面概述了AI驱动的食品安全研究的知识结构、合作网络和主题演变。与以往聚焦特定技术或应用的研究不同，本研究绘制了全球研究模式和跨学科联系。研究结果为研究人员、政策制定者和行业利益相关者提供了战略性见解。未来研究应优先关注可解释AI融入监管框架以及开发更大、更多样化的数据集，以提高模型在食品系统中的泛化能力。","British Food Journal","2026-10-19T00:00:00Z",80,{"impact":16,"substance":101,"depth":16,"authority":102,"freshness":103,"relevant":20,"comment":104},22,14,8,"基于2782篇文献的AI食品安全计量分析，方法规范、结论有参考价值，但属综述性研究，产业落地影响有限。",[106],{"name":97,"url":94},[25,26,108,28,109],"机器学习","文献计量",{"code":38,"slug":111,"name":112,"path":113,"parent_code":114},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[116],{"code":55,"slug":117,"name":118,"path":119,"parent_code":120},"food-safety","质量安全与溯源","\u002Ftopic\u002Ffood-supply\u002Ffood-safety","E",{"status":122,"locked":13,"version":33,"method":34,"evidence":123},"classified",{"candidates":124,"agricultural_evidence":39},[125,134,139,145,152,158],{"code":38,"score":126,"supported":39,"title":127,"abstract":128,"tags":133},19,[43],[129,108,130,42,131,43,132],"深度学习","计算机视觉","deep learning","computer vision",[26,108],{"code":55,"score":135,"supported":39,"title":136,"abstract":137,"tags":138},15,[59],[28,59],[28],{"code":140,"score":135,"supported":39,"title":141,"abstract":143,"tags":144},"I7",[142],"bibliometric",[109,142],[109],{"code":146,"score":48,"supported":39,"title":147,"abstract":148,"tags":151},"H7",[],[149,150],"气候变化","climate change",[],{"code":153,"score":71,"supported":39,"title":154,"abstract":155,"tags":157},"A8",[],[156],"food security",[],{"code":62,"score":20,"supported":13,"title":159,"abstract":160,"tags":161},[],[],[25],[163,164],"AI 食品安全 文献计量","机器学习 食品质量预测","AI食品安全文献计量-5400","10.1108\u002Fbfj-06-2025-0819",{"doi":166,"openalex_id":168,"authors":169,"venue":97,"cited_by_count":79,"oa_url":8,"card":183,"direction":85,"ingested_from":87},"W7222920742",[170,172,174,177,180],{"name":171,"orcid":8},"Harang Lee",{"name":173,"orcid":8},"지은서",{"name":175,"orcid":176},"Yoonbin Kim","https:\u002F\u002Forcid.org\u002F0000-0002-5625-2560",{"name":178,"orcid":179},"Joon-Goo Lee","https:\u002F\u002Forcid.org\u002F0000-0003-3617-5518",{"name":181,"orcid":182},"Woo‐Ju Kim","https:\u002F\u002Forcid.org\u002F0000-0002-6954-1580",{"tldr":184,"method":185,"finding":186,"direction":85,"opportunity":187},"对2021-2025年2782篇文献做文献计量，梳理AI与食品安全交叉研究趋势。","Web of Science文献计量，VOSviewer共被引、关键词共现与耦合","机器学习与深度学习是核心，形成传统ML、计算机视觉、时序建模三大知识基础。","可解释AI嵌入监管框架、构建更大更多样数据集以提升跨食物系统泛化能力。","2026-10-20T23:30:31.567802Z",{"id":190,"title":191,"url":192,"summary":193,"summary_zh":8,"content":8,"source_name":194,"source_url":192,"published_at":195,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":196,"score_detail":197,"sources":201,"tags":203,"primary_topic":206,"secondary_topics":211,"classification":212,"search_phrases":230,"slug":233,"view_count":71,"doi":234,"paper":235,"created_at":250},3876,"Agrivlm: integrating vision-language models into real-time remote sensing for smart agriculture","https:\u002F\u002Fdoi.org\u002F10.1007\u002Fs11276-026-04228-x","Agrivlm: integrating vision-language models into real-time remote sensing for smart agriculture。Wireless Networks","Wireless Networks","2026-09-30T00:00:00Z",78,{"impact":16,"substance":17,"depth":16,"authority":198,"freshness":199,"relevant":20,"comment":200},13,9,"将视觉语言模型引入农业实时遥感，方法新颖且具应用潜力，值得关注。",[202],{"name":194,"url":192},[25,26,27,204,205],"视觉语言模型","遥感",{"code":207,"slug":208,"name":209,"path":210,"parent_code":114},"I3","remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing",[],{"status":122,"locked":13,"version":33,"method":34,"evidence":213},{"candidates":214,"agricultural_evidence":39},[215,220,225],{"code":62,"score":18,"supported":39,"title":216,"abstract":218,"tags":219},[217],"smart agriculture",[217],[25],{"code":207,"score":18,"supported":39,"title":221,"abstract":223,"tags":224},[222],"remote sensing",[222],[205],{"code":38,"score":226,"supported":13,"title":227,"abstract":228,"tags":229},2,[],[],[26,204],[231,232],"Agrivlm 视觉语言模型 遥感","智慧农业 实时遥感","Agrivlm视觉语言模型遥感-3876","10.1007\u002Fs11276-026-04228-x",{"doi":234,"openalex_id":236,"authors":237,"venue":194,"cited_by_count":79,"oa_url":8,"card":8,"direction":249,"ingested_from":87},"W7214888751",[238,240,243,245,247],{"name":239,"orcid":8},"Sasmita Padhy",{"name":241,"orcid":242},"Preetam Suman","https:\u002F\u002Forcid.org\u002F0000-0001-7607-1372",{"name":244,"orcid":8},"Naween Kumar",{"name":246,"orcid":8},"Akansha Singh",{"name":248,"orcid":8},"Krishna Kant Singh","智慧农业 \u002F 农业物联网","2026-10-01T23:30:08.990045Z",{"id":252,"title":253,"url":254,"summary":255,"summary_zh":256,"content":8,"source_name":257,"source_url":254,"published_at":258,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":259,"score_detail":260,"sources":263,"tags":265,"primary_topic":8,"secondary_topics":267,"classification":268,"search_phrases":284,"slug":287,"view_count":226,"doi":288,"paper":289,"created_at":312},1536,"A Non-Invasive Approach for Detecting Water Adulteration in Orange Juice Using Computer Vision and Deep Learning","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.jfca.2026.109484","Food fraud due to water adulteration in orange juice is a significant concern for the beverage industry. This study aimed to develop a rapid and non-invasive computer-vision method based on deep learning to detect and classify water-adulteration levels ranging from 1% to 15% in three orange juice products representing freshly squeezed juice, juice from concentrate, and orange nectar. High-resolution images were acquired under two shutter-speed-based exposure conditions: the higher-exposure acquisition condition (1\u002F30 s) and the lower-exposure acquisition condition (1\u002F250 s), while maintaining constant scene illumination, and were analyzed using ResNet50 convolutional neural networks. In the image-level hold-out test, the model trained with images acquired at 1\u002F250 s achieved an accuracy of 88.3%, whereas the model trained at 1\u002F30 s reached 83.1%. To provide a more stringent assessment of sample-level transferability, the previously trained and fixed 1\u002F250 s model was subsequently evaluated using 240 independently prepared and blindly coded samples obtained from subsequent purchases of the same commercial products. This independent blind validation achieved a 24-class accuracy of 86.7%. Misclassifications occurred predominantly between adjacent or closely related water-adulteration levels within the same juice product. When the independent-validation predictions were collapsed into a binary pure-versus-adulterated screening task, the model achieved 100.0% sensitivity, 93.3% specificity, 99.2% accuracy, and 96.7% balanced accuracy. These findings support the feasibility of the method as a rapid pre-screening tool for detecting visible-image patterns associated with controlled water dilution and show that detecting the presence of adulteration is more reliable than assigning an exact value to closely spaced adulteration levels. The independent validation provides evidence of transferability to newly prepared samples from subsequent purchases of the same products. However, further validation across additional brands, production batches, orange origins, seasons, and acquisition environments is required before broader applicability can be established.","橙汁中因掺水导致的食品欺诈是饮料行业关注的重要问题。本研究旨在开发一种基于深度学习的快速、非侵入性计算机视觉方法，用于检测和分类三种橙汁产品（鲜榨汁、浓缩还原汁和橙汁饮料）中1%至15%的掺水水平。在保持场景照明恒定的条件下，分别以两种基于快门速度的曝光条件采集高分辨率图像：高曝光采集条件（1\u002F30秒）和低曝光采集条件（1\u002F250秒），并使用ResNet50卷积神经网络进行分析。在图像级留出测试中，以1\u002F250秒采集图像训练的模型准确率达到88.3%，而以1\u002F30秒训练的模型准确率为83.1%。为对样本级可迁移性进行更严格的评估，先前训练并固定的1\u002F250秒模型随后被用于评估240个独立制备、盲法编码的样本，这些样本来自后续购买的同款商业产品。该独立盲法验证实现了24类别86.7%的准确率。误分类主要发生在同一果汁产品内相邻或相近的掺水水平之间。当独立验证预测结果被合并为二元纯正与掺假筛查任务时，模型实现了100.0%的灵敏度、93.3%的特异度、99.2%的准确率和96.7%的平衡准确率。这些发现支持该方法作为快速预筛查工具的可行性，用于检测与受控水稀释相关的可见图像模式，并表明检测掺假的存在比精确赋值于间隔较小的掺水水平更为可靠。独立验证提供了该方法对后续购买同款产品新制备样本具有可迁移性的证据。然而，在确立更广泛的适用性之前，仍需在更多品牌、生产批次、橙子产地、季节和采集环境中进行进一步验证。","Journal of Food Composition and Analysis","2026-09-01T00:00:00Z",71,{"impact":18,"substance":17,"depth":16,"authority":102,"freshness":261,"relevant":20,"comment":262},7,"研究提出基于深度学习的橙汁掺水无损检测方法，独立验证准确率高，对食品安全监管有参考价值。",[264],{"name":257,"url":254},[25,26,266,28],"无损检测",[],{"status":32,"locked":13,"version":33,"method":34,"evidence":269},{"candidates":270,"agricultural_evidence":13,"reason":66},[271,276,280],{"code":38,"score":272,"supported":39,"title":273,"abstract":274,"tags":275},26,[131,132],[129,130,131],[26],{"code":55,"score":226,"supported":13,"title":277,"abstract":278,"tags":279},[],[],[266,28],{"code":62,"score":20,"supported":13,"title":281,"abstract":282,"tags":283},[],[],[25],[285,286],"农业人工智能 无损检测 智慧农业 食品安全","农业人工智能 无损检测","农业人工智能无损检测智慧农业食品安全-1536","10.1016\u002Fj.jfca.2026.109484",{"doi":288,"openalex_id":290,"authors":291,"venue":257,"cited_by_count":79,"oa_url":306,"card":307,"direction":85,"ingested_from":87},"W7204957864",[292,294,296,299,301,303],{"name":293,"orcid":8},"Ana M. Pérez-Calabuig",{"name":295,"orcid":8},"Sandra Pradana‐López",{"name":297,"orcid":298},"John C. Cancilla","https:\u002F\u002Forcid.org\u002F0000-0003-3645-7224",{"name":300,"orcid":8},"María Luz Mena",{"name":302,"orcid":8},"Carlos López-Pingarrón",{"name":304,"orcid":305},"José S. Torrecilla","https:\u002F\u002Forcid.org\u002F0000-0003-1209-203X","https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS0889157526006277\u002Fpdf",{"tldr":308,"method":309,"finding":310,"direction":85,"opportunity":311},"用深度学习计算机视觉检测橙汁掺水，实现非侵入式快速筛查。","采集不同曝光图像，用ResNet50分类掺水等级，独立盲样验证。","掺水检测准确率高，区分精确等级较难，模型可迁移至新样本。","可扩展至其他食品掺假检测，需跨品牌、批次、季节等验证，或开发便携设备。","2026-09-03T23:30:50.596809Z",{"id":314,"title":315,"url":316,"summary":317,"summary_zh":318,"content":8,"source_name":319,"source_url":316,"published_at":320,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":321,"score_detail":322,"sources":325,"tags":327,"primary_topic":331,"secondary_topics":337,"classification":339,"search_phrases":355,"slug":358,"view_count":79,"doi":359,"paper":360,"created_at":374},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。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-21T00:00:00Z",69,{"impact":103,"substance":323,"depth":16,"authority":18,"freshness":12,"relevant":20,"comment":324},21,"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。",[326],{"name":319,"url":316},[25,26,328,329,330],"小农户","植物病害识别","可解释性",{"code":332,"slug":333,"name":334,"path":335,"parent_code":336},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services","G",[338],{"code":38,"slug":111,"name":112,"path":113,"parent_code":114},{"status":122,"locked":13,"version":33,"method":34,"evidence":340},{"candidates":341,"agricultural_evidence":39},[342,347,351],{"code":332,"score":18,"supported":39,"title":343,"abstract":345,"tags":346},[344],"smallholder",[344],[328],{"code":38,"score":12,"supported":39,"title":348,"abstract":349,"tags":350},[132],[],[26,329],{"code":62,"score":20,"supported":13,"title":352,"abstract":353,"tags":354},[],[],[25],[356,357],"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":359,"openalex_id":361,"authors":362,"venue":319,"cited_by_count":79,"oa_url":316,"card":369,"direction":85,"ingested_from":87},"W7223137323",[363,365,367],{"name":364,"orcid":8},"Akshay Ramgoolam",{"name":366,"orcid":8},"Priya Bundhoo Appadoo",{"name":368,"orcid":8},"Jean-Noël LEspérance",{"tldr":370,"method":371,"finding":372,"direction":85,"opportunity":373},"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",{"id":376,"title":377,"url":378,"summary":379,"summary_zh":380,"content":8,"source_name":319,"source_url":378,"published_at":320,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":321,"score_detail":381,"sources":383,"tags":385,"primary_topic":389,"secondary_topics":395,"classification":397,"search_phrases":420,"slug":423,"view_count":79,"doi":424,"paper":425,"created_at":437},5591,"Hierarchical Conformal Risk Control for Artificial Intelligence Agriculture Irrigation Scheduling Under Unmodeled Infiltration: A Computer Safety-Guarantee Evaluation","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23543665","Irrigation scheduling algorithms that rely on soil-moisture forecasts are vulnerable to unmodeled infiltration, the process by which water bypasses the root zone through preferential pathways that resistance-based soil models cannot represent. The framework is evaluated through a replication-based design study that characterises the operating characteristics of the before-after-control-impact estimator used to audit scheduler performance, under a declared variance structure with between-cluster standard deviation 6.0, residual standard deviation 4.5 and intraclass correlation 0.64. Power was estimated across 2000 design cells spanning cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The analysis indicates that 54 clusters are required to detect a 5-percentage-point effect at 80% power, and that the minimum detectable effect declines steeply as cluster count rises from 8 to 32. The contribution is a formal safety guarantee for irrigation decision support that remains valid when the infiltration model is wrong, together with an audit design that makes the guarantee empirically checkable.","依赖土壤湿度预报的灌溉调度算法容易受到未建模入渗的影响，即水分通过优先路径绕过根区的过程，而基于阻力的土壤模型无法表征这一过程。该框架通过一项基于重复的设计研究进行评估，该研究刻画了用于审计调度器性能的前后对照影响估计量的运行特性，其所依据的声明方差结构为：组间标准差6.0、残差标准差4.5、组内相关系数0.64。功效估计覆盖2000个设计单元，涵盖聚类数8、16、24和32以及效应量3、5和8个百分点。分析表明，在80%功效下检测5个百分点效应需要54个聚类，且最小可检测效应随聚类数从8增至32而急剧下降。本文的贡献在于为灌溉决策支持提供了一种正式的安全保证，该保证在入渗模型错误时仍然有效，同时提出了一种使该保证可经实证检验的审计设计。",{"impact":103,"substance":101,"depth":16,"authority":18,"freshness":199,"relevant":20,"comment":382},"该研究提出在入渗模型失配条件下仍有效的灌溉决策安全保证框架，并给出可审计的试验设计，方法新颖、结论可靠，但属细分方法学研究，产业影响有限，适合作为专业精选而非头条。",[384],{"name":319,"url":378},[25,26,386,387,388],"灌溉决策","安全保证","试验设计",{"code":390,"slug":391,"name":392,"path":393,"parent_code":394},"H3","irrigation","水资源与灌溉","\u002Ftopic\u002Fenvironment\u002Firrigation","H",[396],{"code":38,"slug":111,"name":112,"path":113,"parent_code":114},{"status":122,"locked":13,"version":33,"method":34,"evidence":398},{"candidates":399,"agricultural_evidence":39},[400,405,409,416],{"code":390,"score":135,"supported":39,"title":401,"abstract":402,"tags":404},[391],[403,391],"灌溉",[386],{"code":38,"score":18,"supported":39,"title":406,"abstract":407,"tags":408},[43],[44],[26],{"code":410,"score":48,"supported":39,"title":411,"abstract":412,"tags":415},"H2",[],[413,414],"土壤","soil",[],{"code":62,"score":20,"supported":13,"title":417,"abstract":418,"tags":419},[],[],[25],[421,422],"灌溉调度 安全保证","农业人工智能 安全保证 智慧农业 灌溉决策","灌溉调度安全保证-5591","10.5281\u002Fzenodo.23543665",{"doi":424,"openalex_id":426,"authors":427,"venue":319,"cited_by_count":79,"oa_url":378,"card":432,"direction":85,"ingested_from":87},"W7223131205",[428,430],{"name":429,"orcid":8},"Nassima Benkhelifa",{"name":431,"orcid":8},"Yacine Aït Ahmed",{"tldr":433,"method":434,"finding":435,"direction":85,"opportunity":436},"提出分层保形风险控制，为未建模入渗下的灌溉调度提供安全保证。","分层保形风险控制与基于重复设计的BACI审计，2000个设计单元功效分析。","检测5个百分点效应需54个集群，最小可检测效应随集群数增加陡降。","可将保形风险控制扩展到其他未建模扰动下的农业决策安全保证与审计设计。","2026-10-22T11:21:10.283561Z",{"id":439,"title":440,"url":441,"summary":442,"summary_zh":443,"content":8,"source_name":319,"source_url":441,"published_at":320,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":444,"score_detail":445,"sources":447,"tags":449,"primary_topic":452,"secondary_topics":453,"classification":455,"search_phrases":470,"slug":473,"view_count":79,"doi":474,"paper":475,"created_at":489},5590,"Evaluation of Concept-Probe Temporal-Stability Audits for Computer Vision Agriculture Disease Classifiers Under Seasonal Phenology Drift: An Interpretability Decay Bound for Smallholder Fungicide Timing","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23545234","Concept-probe audits are increasingly used to certify that a crop-disease classifier's internal representations track the pathological concepts a smallholder would act on, but a probe certified at one point in the season may silently decouple from those concepts as phenology shifts the visual statistics of the field. the present analysis treats that decoupling as a formal object. It defines a temporal-stability audit as a sequence of concept-probe accuracies measured across defined phenological windows, and it derives an interpretability decay bound: the rate at which probe accuracy can fall before the fungicide-timing decision it is meant to support becomes unreliable. The bound is expressed as a function of the probe's initial margin, the drift in concept prevalence between windows, and the classifier's sensitivity to the drifted features, and it yields a certified audit interval rather than a single-point certificate. The paper then characterises the operating characteristics of the audit design under a declared variance structure, using a replication-based evaluation of the before-after-control-impact estimator with between-cluster SD = 6.0, residual SD = 4.5 and intraclass correlation = 0.64. Power was estimated across 2000 design cells over cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points; 54 clusters were required for 5 percentage points at 80% power. The contribution is a decision-theoretic account of when a concept-probe certificate expires, and a design rule for scheduling re-audits across the season.","概念探针审计（concept-probe audit）日益被用于验证作物病害分类器的内部表征是否追踪小农户会据以采取行动的病理学概念，但在生长季某一时点获得认证的探针，可能随着物候变化改变田间的视觉统计特征而悄然与这些概念脱钩。本分析将这种脱钩视为一个形式化对象。它将时间稳定性审计（temporal-stability audit）定义为在既定物候窗口内测量的一系列概念探针准确率，并推导出一个可解释性衰减界限（interpretability decay bound）：即探针准确率在其所支持的杀菌剂施用时机决策变得不可靠之前所能下降的速率。该界限表示为探针初始裕度、窗口间概念流行率漂移以及分类器对漂移特征敏感度的函数，并给出一个经认证的审计区间而非单点认证。随后，本文在声明的方差结构下刻画该审计设计的运行特征，采用基于重复的前后对照影响估计量评估，其中组间标准差 = 6.0、残差标准差 = 4.5、组内相关系数 = 0.64。功效在 2000 个设计单元上估计，聚类数取 8、16、24 和 32，效应量取 3、5 和 8 个百分点；在 80% 功效下达到 5 个百分点需要 54 个聚类。本文的贡献在于从决策论角度阐明概念探针认证何时失效，并给出在整个生长季内安排重新审计的设计规则。",70,{"impact":103,"substance":101,"depth":16,"authority":18,"freshness":12,"relevant":20,"comment":446},"该论文提出概念探针审计的时序稳定性框架与可解释性衰减界，方法新颖、统计设计扎实，但属小众方法论研究，产业影响有限，适合作为专业精选而非头条。",[448],{"name":319,"url":441},[25,26,328,450,451],"作物病害识别","模型审计",{"code":332,"slug":333,"name":334,"path":335,"parent_code":336},[454],{"code":38,"slug":111,"name":112,"path":113,"parent_code":114},{"status":122,"locked":13,"version":33,"method":34,"evidence":456},{"candidates":457,"agricultural_evidence":39},[458,462,466],{"code":332,"score":135,"supported":39,"title":459,"abstract":460,"tags":461},[344],[328,344],[328],{"code":38,"score":12,"supported":39,"title":463,"abstract":464,"tags":465},[132],[],[450,26],{"code":62,"score":20,"supported":13,"title":467,"abstract":468,"tags":469},[],[],[25],[471,472],"概念探针 时序稳定性 作物病害分类器","杀菌剂 施药时机 可解释性衰减","概念探针时序稳定性作物病害分类器-5590","10.5281\u002Fzenodo.23545234",{"doi":474,"openalex_id":476,"authors":477,"venue":319,"cited_by_count":79,"oa_url":441,"card":484,"direction":85,"ingested_from":87},"W7223088030",[478,480,482],{"name":479,"orcid":8},"Abdi-Basid Idriss Farah",{"name":481,"orcid":8},"Amina Abdillahi Guedi",{"name":483,"orcid":8},"Moussa Omar Youssouf",{"tldr":485,"method":486,"finding":487,"direction":85,"opportunity":488},"提出概念探针时间稳定性审计与可解释性衰减界，判断何时需重新审计。","定义跨物候窗口探针精度序列，推导衰减界，用BACI复制评估统计功效。","5个百分点效应需54个聚类达80%功效，审计证书会随物候漂移过期。","可将衰减界与再审计调度规则嵌入小农户杀菌剂时机决策支持系统，并验证跨作物泛化。","2026-10-22T11:21:08.978348Z"]