[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2618":3,"related-2618":102},{"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":23,"tags":25,"primary_topic":31,"secondary_topics":37,"classification":38,"search_phrases":64,"slug":67,"view_count":21,"doi":68,"paper":69,"created_at":101},2618,"Mastitis detection in dairy cows with varying postures using infrared thermography based on YOLOv8n-improved and random forest models","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.compag.2026.112430","Current automated mastitis detection methods typically identify dairy cows only in a fixed standing posture, which limits their scope and practical flexibility. To address potential occlusion issues across different postures, we divided the key anatomical regions into three areas (the eye, the back surface of the udder (BSU) and the lower surface of the udder (LSU)) and developed an infrared thermography (IRT)-based automated diagnostic system for cows in various postures within a lactation barn. First, a three-stage image enhancement method was applied to extract contour and texture features from the IRT images. Next, the You Only Look Once v8 Nano (YOLOv8n) model was improved by integrating Dynamic Snake Convolution to better capture subtle and complex texture patterns. We further optimised the weight distribution of contour and texture features using an efficient multi-scale attention module to reduce the loss of critical information in the deep network. Structural improvements included adding a P2 detection head to focus on contour features in the target regions. Finally, we built three machine learning models to diagnose mastitis using the maximum body temperatures of these critical regions. Results showed that the three-stage image enhancement effectively enriched IRT details, strengthened contour and texture features and improved detection confidence for the eye, BSU and LSU by 0.025, 0.05 and 0.04, respectively. The improved YOLOv8n model achieved top performance, with precision (P) of 94.2%, 97.8% and 96.1%; recall (R) of 96.6%, 94.1% and 89.7%; and average precision at an intersection-over-union of 50% (AP@0.5) of 94.3%, 93.7% and 94.2% for the eye, BSU and LSU, respectively. Compared with the baseline YOLOv8n model, the enhanced version improved P, R and AP@0.5 metrics by 2%∼5.4% across the three regions of interest. Among the diagnostic models, random forest achieved the highest accuracy at 92.31%. This method broadens the application of automated mastitis detection and provides a reference framework for building automatic monitoring systems for mastitis in feeder barns.","目前的自动乳腺炎检测方法通常仅能在固定站立姿势下识别奶牛，这限制了其适用范围和实践灵活性。为解决不同姿势下可能出现的遮挡问题，我们将关键解剖区域划分为三个区域（眼部、乳房背面（BSU）和乳房下面（LSU）），并开发了一套基于红外热成像（IRT）的自动化诊断系统，用于泌乳牛舍中不同姿势的奶牛。首先，应用三阶段图像增强方法从IRT图像中提取轮廓和纹理特征。接着，通过集成动态蛇形卷积对YOLOv8n模型进行改进，以更好地捕捉细微复杂的纹理模式。我们进一步利用高效多尺度注意力模块优化轮廓和纹理特征的权重分配，以减少深层网络中关键信息的损失。结构改进包括添加P2检测头以聚焦目标区域的轮廓特征。最后，我们构建了三个机器学习模型，利用这些关键区域的最大体表温度来诊断乳腺炎。结果表明，三阶段图像增强有效丰富了IRT细节，增强了轮廓和纹理特征，并将眼部、BSU和LSU的检测置信度分别提高了0.025、0.05和0.04。改进后的YOLOv8n模型取得了最佳性能，眼部、BSU和LSU的精确率（P）分别为94.2%、97.8%和96.1%；召回率（R）分别为96.6%、94.1%和89.7%；交并比为50%时的平均精度（AP@0.5）分别为94.3%、93.7%和94.2%。与基线YOLOv8n模型相比，增强版本在三个感兴趣区域的P、R和AP@0.5指标上提升了2%∼5.4%。在诊断模型中，随机森林达到了最高准确率92.31%。该方法拓宽了自动乳腺炎检测的应用范围，并为构建饲养牛舍乳腺炎自动监测系统提供了参考框架。",null,"Computers and Electronics in Agriculture","2026-09-15T00:00:00Z","论文",10,false,81,{"impact":17,"substance":18,"depth":17,"authority":19,"freshness":20,"relevant":21,"comment":22},18,22,14,9,1,"该论文提出基于改进YOLOv8n与随机森林的红外热成像奶牛乳腺炎检测方法，支持多姿态识别，方法新颖、数据详实，对智慧牧场疫病监测有参考价值，值得进入每日精选。",[24],{"name":10,"url":6},[26,27,28,29,30],"智慧农业","农业人工智能","奶牛养殖","红外热成像","疫病智能诊断",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":39,"locked":14,"version":40,"method":41,"evidence":42},"classified","2026-10-22.2","rules",{"candidates":43,"agricultural_evidence":46},[44,52,59],{"code":32,"score":45,"supported":46,"title":47,"abstract":48,"tags":51},7,true,[],[49,50],"机器学习","machine learning",[27],{"code":53,"score":54,"supported":46,"title":55,"abstract":56,"tags":58},"C2",4,[],[57],"奶牛",[28],{"code":60,"score":21,"supported":14,"title":61,"abstract":62,"tags":63},"I1",[],[],[26],[65,66],"农业人工智能 疫病智能诊断 红外热成像 奶牛养殖","农业人工智能 疫病智能诊断","农业人工智能疫病智能诊断红外热成像奶牛养殖-2618","10.1016\u002Fj.compag.2026.112430",{"doi":68,"openalex_id":70,"authors":71,"venue":10,"cited_by_count":93,"oa_url":6,"card":94,"direction":98,"ingested_from":100},"W7213340298",[72,75,77,79,81,84,86,88,90],{"name":73,"orcid":74},"Hang Song","https:\u002F\u002Forcid.org\u002F0009-0008-7352-0855",{"name":76,"orcid":9},"Longwei Guo",{"name":78,"orcid":9},"Hang Shi",{"name":80,"orcid":9},"Miao Wu",{"name":82,"orcid":83},"Hang Xue","https:\u002F\u002Forcid.org\u002F0009-0008-4953-8748",{"name":85,"orcid":9},"Hui Zhang",{"name":87,"orcid":9},"Huize Lv",{"name":89,"orcid":9},"Qiuju Xie",{"name":91,"orcid":92},"Jun Hu","https:\u002F\u002Forcid.org\u002F0009-0003-2743-0641",0,{"tldr":95,"method":96,"finding":97,"direction":98,"opportunity":99},"用改进YOLOv8n与随机森林，通过红外热成像实现奶牛不同姿态下的乳腺炎自动检测。","三阶段图像增强、动态蛇形卷积改进YOLOv8n、多尺度注意力与P2检测头，随机森","改进模型对眼、乳房后表面、乳房下表面检测AP@0.5达94%左右，随机森林诊断准确率92.31%。","农业人工智能与决策模型","可探索多模态融合与轻量化边缘部署，实现牛舍内多姿态、多疾病的实时连续监测。","openalex","2026-09-16T23:30:02.203663Z",{"total":103,"page":21,"page_size":103,"items":104},6,[105,177,244,306,369,421],{"id":106,"title":107,"url":108,"summary":109,"summary_zh":110,"content":9,"source_name":111,"source_url":108,"published_at":112,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":113,"score_detail":114,"sources":117,"tags":119,"primary_topic":122,"secondary_topics":127,"classification":129,"search_phrases":149,"slug":152,"view_count":21,"doi":153,"paper":154,"created_at":176},4244,"Automated dairy cattle lameness assessment using a 38-keypoint whole-body YOLOv8 pose framework","https:\u002F\u002Fdoi.org\u002F10.14202\u002Fvetworld.2026.4286-4306","Background and Aim: Lameness is a major health, welfare, and production concern in dairy cattle, whereas conventional visual locomotion scoring is intermittent and susceptible to observer-dependent variation. Markerless computer vision may enable objective and repeatable locomotion monitoring under routine farm conditions. This study aimed to develop a 38-keypoint whole-body pose-estimation framework for automated dairy cattle lameness assessment and evaluate its agreement with blinded Sprecher locomotion scoring. Materials and Methods: A You Only Look Once version 8 (YOLOv8)-pose framework was developed using 12,000 annotated images obtained from 449 Holstein cows. Thirty-eight anatomical keypoints represented the head, spine, pelvis, forelimbs, hindlimbs, and hooves. Four model and input resolution configurations were evaluated. Joint-angle time series, coefficients of variation, and whole-body asymmetry features were derived from detected keypoints. For independent clinical validation, 120 cattle videos, obtained from animals with no overlap at the animal level with the 449-cow model-development cohort, were evaluated blindly by three assessors using the five-point Sprecher scale. Inter-observer reliability, consensus scores, rank associations, group differences, and exploratory receiver operating characteristic analyses were determined. Results: YOLOv8l-pose at 2560 × 2560 pixels showed the best performance, achieving pose mean average precision at an intersection over union threshold of 0.5 of 83.11%, mean average precision across thresholds of 0.5–0.95 of 37.39%, and recall of 91.74%. All three assessors assigned identical Sprecher scores to 87.5% of videos. Pairwise quadratic-weighted Cohen κ ranged from 0.843 to 0.926, Fleiss κ was 0.779, and intraclass correlation coefficients were 0.883 and 0.958 for single and average measurements, respectively. The continuous artificial intelligence (AI) score correlated significantly with consensus Sprecher scores (Spearman ρ = 0.234, p = 0.010), and score distributions differed among consensus groups (Kruskal–Wallis H = 13.09, p = 0.0014). Exploratory discrimination of consensus Sprecher scores ≥3, using a threshold internally derived from and evaluated within the same stratified validation cohort (not externally validated), yielded an area under the curve of 0.875; at an AI threshold of 2.9, sensitivity and specificity were 62.5% and 98.2%, respectively. Conclusion: The 38-keypoint whole-body model demonstrates technical feasibility as an anatomically comprehensive, sensor-free approach for quantifying dairy cattle locomotion. Because the clinical validation cohort was deliberately stratified and diagnostic thresholds were internally derived, its agreement with independent visual assessment should be interpreted as evidence of potential decision-support utility rather than established clinical performance. Prospective external validation and independent calibration of diagnostic thresholds in an unselected cohort are warranted before clinical or commercial deployment. Keywords: artificial intelligence, bovine lameness, computer vision, dairy cattle, locomotion assessment, pose-estimation, Sprecher score, whole-body keypoints.","背景与目的：跛行是奶牛重要的健康、福利和生产问题，而传统的视觉运动评分具有间歇性且易受观察者差异影响。无标记计算机视觉可能有助于在常规农场条件下实现客观且可重复的运动监测。本研究旨在开发一个38关键点全身姿态估计框架，用于自动化奶牛跛行评估，并评估其与盲法Sprecher运动评分的一致性。材料与方法：使用来自449头荷斯坦奶牛的12,000张标注图像，开发了You Only Look Once第8版（YOLOv8）-pose框架。38个解剖关键点代表头部、脊柱、骨盆、前肢、后肢和蹄部。评估了四种模型和输入分辨率配置。从检测到的关键点中提取关节角度时间序列、变异系数和全身不对称特征。为进行独立临床验证，由三名评估者使用五分制Sprecher量表对120头牛的视频进行盲法评估，这些动物在个体层面与449头模型开发队列无重叠。确定了观察者间可靠性、共识评分、秩相关、组间差异和探索性受试者工作特征分析。结果：2560×2560像素下的YOLOv8l-pose表现最佳，在交并比阈值为0.5时姿态平均精度为83.11%，在0.5–0.95阈值范围内的平均精度为37.39%，召回率为91.74%。三名评估者对87.5%的视频给出了相同的Sprecher评分。两两二次加权Cohen κ范围为0.843至0.926，Fleiss κ为0.779，单次测量和平均测量的组内相关系数分别为0.883和0.958。连续人工智能（AI）评分与共识Sprecher评分显著相关（Spearman ρ = 0.234，p = 0.010），且各共识组之间的评分分布不同（Kruskal–Wallis H = 13.09，p = 0.0014）。使用在同一分层验证队列内部推导并评估的阈值（未经外部验证），对共识Sprecher评分≥3进行探索性判别时，曲线下面积为0.875；在AI阈值为2.9时，敏感性和特异性分别为62.5%和98.2%。结论：38关键点全身模型表现","Veterinary World","2026-10-03T00:00:00Z",80,{"impact":17,"substance":18,"depth":17,"authority":115,"freshness":20,"relevant":21,"comment":116},13,"基于12000张标注图像与120头独立验证视频的38关键点YOLOv8姿态估计框架，为奶牛跛行无接触客观评估提供技术可行性证据，方法新颖、数据扎实，但阈值内推、缺乏外部验证，属细分领域技术进展。",[118],{"name":111,"url":108},[26,27,120,121,28],"计算机视觉","动物健康",{"code":53,"slug":123,"name":124,"path":125,"parent_code":126},"ruminants","草食牲畜","\u002Ftopic\u002Flivestock\u002Fruminants","C",[128],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":39,"locked":14,"version":40,"method":41,"evidence":130},{"candidates":131,"agricultural_evidence":46},[132,138,145],{"code":53,"score":133,"supported":46,"title":134,"abstract":136,"tags":137},15,[135],"cattle",[57,135],[28],{"code":32,"score":139,"supported":46,"title":140,"abstract":141,"tags":144},11,[],[120,142,143],"artificial intelligence","computer vision",[27,120],{"code":60,"score":21,"supported":14,"title":146,"abstract":147,"tags":148},[],[],[26],[150,151],"YOLOv8 奶牛 跛行 姿态估计","奶牛跛行 自动评估 计算机视觉","YOLOv8奶牛跛行姿态估计-4244","10.14202\u002Fvetworld.2026.4286-4306",{"doi":153,"openalex_id":155,"authors":156,"venue":111,"cited_by_count":93,"oa_url":169,"card":170,"direction":175,"ingested_from":100},"W7218963807",[157,160,163,166],{"name":158,"orcid":159},"А В Греченева","https:\u002F\u002Forcid.org\u002F0000-0002-7341-5237",{"name":161,"orcid":162},"Maxim Baknin","https:\u002F\u002Forcid.org\u002F0000-0002-3124-6252",{"name":164,"orcid":165},"Sergey Vladimirovich Akchurin","https:\u002F\u002Forcid.org\u002F0000-0002-6822-0013",{"name":167,"orcid":168},"Ирина Владимировна Акчурина","https:\u002F\u002Forcid.org\u002F0000-0002-1975-2929","https:\u002F\u002Fveterinaryworld.org\u002FVol.19\u002FOctober-2026\u002F3.pdf",{"tldr":171,"method":172,"finding":173,"direction":98,"opportunity":174},"构建38关键点全身YOLOv8姿态框架，实现奶牛跛行自动评估并与Sprecher评分比对。","YOLOv8l-pose，12000张449头荷斯坦牛图像，120视频盲评验证。","模型mAP@0.5达83.11%，AI评分与共识Sprecher评分显著相关，AUC达0.875。","可扩展至多牧场外部验证，并结合时序行为与多模态数据提升跛行早期预警泛化性。","数字乡村与农业信息化","2026-10-05T23:30:33.217120Z",{"id":178,"title":179,"url":180,"summary":181,"summary_zh":182,"content":9,"source_name":183,"source_url":180,"published_at":184,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":185,"score_detail":186,"sources":191,"tags":193,"primary_topic":195,"secondary_topics":196,"classification":197,"search_phrases":219,"slug":222,"view_count":223,"doi":224,"paper":225,"created_at":243},1909,"Prediction of lactation milk yields using machine learning based on milk production data and reproduction performance indicators","https:\u002F\u002Fdoi.org\u002F10.1038\u002Fs41598-026-66738-0","The purpose of this research is prediction of total lactation milk yields with the application of machine learning techniques on the basis of data set containing production and reproductive performance indicators for Brown Swiss cows reared on a commercial dairy farm. The dataset includes variables age, number of lactations, number of days in milk (DIM), peak milk yield, 7-day average milk yield (DMY7), dry period, number of inseminations and calving interval, as well as the reproductive status as a categorical variable. Total lactation milk yield is used as a target variable. Performance of various machine learning algorithms was compared with 5-fold cross-validation and the separate test set using R², RMSE, MAE and MAPE criteria. Test results showed that the best prediction performance was provided by Extra Trees algorithm (R² = 0.924, RMSE = 510.27 kg, MAPE = 7.63%). The results showed that ensemble-based tree models allow to better capture non-linear dependencies between milk yield and production\u002Freproductive performance indicators. The interpretation of the model with SHAP analysis showed that day of milking and short-term milk production indicators have the highest importance for prediction. The results suggest that machine learning-based approaches may serve as a reliable tool for decision support systems in milk yield prediction.","本研究旨在利用机器学习技术，基于商业奶牛场饲养的瑞士褐牛的生产与繁殖性能指标数据集，预测全泌乳期产奶量。数据集包含年龄、泌乳次数、泌乳天数（DIM）、峰值产奶量、7天平均产奶量（DMY7）、干奶期、输精次数和产犊间隔等变量，以及作为分类变量的繁殖状态。全泌乳期产奶量作为目标变量。通过5折交叉验证和独立测试集，采用R²、RMSE、MAE和MAPE指标比较了多种机器学习算法的性能。测试结果表明，Extra Trees算法提供了最佳的预测性能（R² = 0.924，RMSE = 510.27 kg，MAPE = 7.63%）。结果显示，基于集成的树模型能够更好地捕捉产奶量与生产\u002F繁殖性能指标之间的非线性依赖关系。通过SHAP分析对模型进行解释表明，泌乳天数和短期产奶指标对预测具有最高的重要性。研究结果表明，基于机器学习的方法可作为产奶量预测决策支持系统的可靠工具。","Scientific Reports","2026-09-07T00:00:00Z",69,{"impact":187,"substance":188,"depth":189,"authority":115,"freshness":45,"relevant":21,"comment":190},12,20,17,"基于机器学习预测奶牛泌乳量，方法新颖且数据可靠，对智慧养殖有参考价值。",[192],{"name":183,"url":180},[26,27,194,49,28],"产量预测",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[],{"status":39,"locked":14,"version":40,"method":41,"evidence":198},{"candidates":199,"agricultural_evidence":46},[200,206,210,214],{"code":32,"score":201,"supported":46,"title":202,"abstract":203,"tags":205},19,[50],[49,50,204],"decision support",[27,49],{"code":53,"score":54,"supported":46,"title":207,"abstract":208,"tags":209},[],[57],[28],{"code":60,"score":21,"supported":14,"title":211,"abstract":212,"tags":213},[],[],[26],{"code":215,"score":21,"supported":14,"title":216,"abstract":217,"tags":218},"I3",[],[],[194],[220,221],"农业人工智能 产量预测 奶牛养殖 智慧农业","农业人工智能 产量预测","农业人工智能产量预测奶牛养殖智慧农业-1909",2,"10.1038\u002Fs41598-026-66738-0",{"doi":224,"openalex_id":226,"authors":227,"venue":183,"cited_by_count":93,"oa_url":237,"card":238,"direction":98,"ingested_from":100},"W7210294025",[228,231,234],{"name":229,"orcid":230},"Zeynep Sönmez","https:\u002F\u002Forcid.org\u002F0000-0003-2696-9138",{"name":232,"orcid":233},"Tuba Adar","https:\u002F\u002Forcid.org\u002F0000-0003-4749-5226",{"name":235,"orcid":236},"İremnur AYDIN","https:\u002F\u002Forcid.org\u002F0000-0003-3374-4586","https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41598-026-66738-0_reference.pdf",{"tldr":239,"method":240,"finding":241,"direction":98,"opportunity":242},"用机器学习基于生产和繁殖数据预测奶牛总泌乳量，Extra Trees表现最佳。","比较多种机器学习算法，5折交叉验证，SHAP分析特征重要性。","Extra Trees预测最优（R²=0.924），泌乳天数和短期产奶量最重要。","可探索将繁殖性能指标与生产数据结合，优化模型泛化性，并开发实时决策支持系统。","2026-09-08T23:30:31.298607Z",{"id":245,"title":246,"url":247,"summary":248,"summary_zh":249,"content":9,"source_name":250,"source_url":247,"published_at":251,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":185,"score_detail":252,"sources":256,"tags":258,"primary_topic":262,"secondary_topics":268,"classification":270,"search_phrases":286,"slug":289,"view_count":93,"doi":290,"paper":291,"created_at":305},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",{"impact":253,"substance":254,"depth":17,"authority":187,"freshness":13,"relevant":21,"comment":255},8,21,"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。",[257],{"name":250,"url":247},[26,27,259,260,261],"小农户","植物病害识别","可解释性",{"code":263,"slug":264,"name":265,"path":266,"parent_code":267},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services","G",[269],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":39,"locked":14,"version":40,"method":41,"evidence":271},{"candidates":272,"agricultural_evidence":46},[273,278,282],{"code":263,"score":187,"supported":46,"title":274,"abstract":276,"tags":277},[275],"smallholder",[275],[259],{"code":32,"score":13,"supported":46,"title":279,"abstract":280,"tags":281},[143],[],[27,260],{"code":60,"score":21,"supported":14,"title":283,"abstract":284,"tags":285},[],[],[26],[287,288],"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":290,"openalex_id":292,"authors":293,"venue":250,"cited_by_count":93,"oa_url":247,"card":300,"direction":98,"ingested_from":100},"W7223137323",[294,296,298],{"name":295,"orcid":9},"Akshay Ramgoolam",{"name":297,"orcid":9},"Priya Bundhoo Appadoo",{"name":299,"orcid":9},"Jean-Noël LEspérance",{"tldr":301,"method":302,"finding":303,"direction":98,"opportunity":304},"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":307,"title":308,"url":309,"summary":310,"summary_zh":311,"content":9,"source_name":250,"source_url":309,"published_at":251,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":185,"score_detail":312,"sources":314,"tags":316,"primary_topic":320,"secondary_topics":326,"classification":328,"search_phrases":351,"slug":354,"view_count":93,"doi":355,"paper":356,"created_at":368},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":253,"substance":18,"depth":17,"authority":187,"freshness":20,"relevant":21,"comment":313},"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。",[315],{"name":250,"url":309},[26,27,317,318,319],"灌溉决策","安全保证","试验设计",{"code":321,"slug":322,"name":323,"path":324,"parent_code":325},"H3","irrigation","水资源与灌溉","\u002Ftopic\u002Fenvironment\u002Firrigation","H",[327],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":39,"locked":14,"version":40,"method":41,"evidence":329},{"candidates":330,"agricultural_evidence":46},[331,336,340,347],{"code":321,"score":133,"supported":46,"title":332,"abstract":333,"tags":335},[322],[334,322],"灌溉",[317],{"code":32,"score":187,"supported":46,"title":337,"abstract":338,"tags":339},[142],[204],[27],{"code":341,"score":103,"supported":46,"title":342,"abstract":343,"tags":346},"H2",[],[344,345],"土壤","soil",[],{"code":60,"score":21,"supported":14,"title":348,"abstract":349,"tags":350},[],[],[26],[352,353],"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":355,"openalex_id":357,"authors":358,"venue":250,"cited_by_count":93,"oa_url":309,"card":363,"direction":98,"ingested_from":100},"W7223131205",[359,361],{"name":360,"orcid":9},"Nassima Benkhelifa",{"name":362,"orcid":9},"Yacine Aït Ahmed",{"tldr":364,"method":365,"finding":366,"direction":98,"opportunity":367},"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":370,"title":371,"url":372,"summary":373,"summary_zh":374,"content":9,"source_name":250,"source_url":372,"published_at":251,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":375,"score_detail":376,"sources":378,"tags":380,"primary_topic":383,"secondary_topics":384,"classification":386,"search_phrases":401,"slug":404,"view_count":93,"doi":405,"paper":406,"created_at":420},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":253,"substance":18,"depth":17,"authority":187,"freshness":13,"relevant":21,"comment":377},"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。",[379],{"name":250,"url":372},[26,27,259,381,382],"作物病害识别","模型审计",{"code":263,"slug":264,"name":265,"path":266,"parent_code":267},[385],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":39,"locked":14,"version":40,"method":41,"evidence":387},{"candidates":388,"agricultural_evidence":46},[389,393,397],{"code":263,"score":133,"supported":46,"title":390,"abstract":391,"tags":392},[275],[259,275],[259],{"code":32,"score":13,"supported":46,"title":394,"abstract":395,"tags":396},[143],[],[381,27],{"code":60,"score":21,"supported":14,"title":398,"abstract":399,"tags":400},[],[],[26],[402,403],"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":405,"openalex_id":407,"authors":408,"venue":250,"cited_by_count":93,"oa_url":372,"card":415,"direction":98,"ingested_from":100},"W7223088030",[409,411,413],{"name":410,"orcid":9},"Abdi-Basid Idriss Farah",{"name":412,"orcid":9},"Amina Abdillahi Guedi",{"name":414,"orcid":9},"Moussa Omar Youssouf",{"tldr":416,"method":417,"finding":418,"direction":98,"opportunity":419},"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",{"id":422,"title":423,"url":424,"summary":425,"summary_zh":426,"content":9,"source_name":250,"source_url":424,"published_at":251,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":427,"score_detail":428,"sources":430,"tags":432,"primary_topic":435,"secondary_topics":436,"classification":437,"search_phrases":455,"slug":458,"view_count":93,"doi":459,"paper":460,"created_at":470},5589,"Conformal-Risk Evaluation for Artificial Intelligence Agriculture Spray-Nozzle Fault Detection Under Label-Delay Feedback: A Computer Coverage Audit","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23544107","Spray-nozzle fault detection in precision agriculture is a classification problem in which the label that confirms a fault arrives after the spray pass, so the detector must act on unlabelled evidence and its error rate must be certified before the label exists. Conformal prediction offers distribution-free coverage guarantees, but the guarantee is stated for exchangeable calibration and test data, an assumption that label-delayed feedback violates because the calibration set is always older than the deployment set. the present analysis formalises label-delay feedback as a distribution shift in the conformity score and derives a coverage bound for split-conformal fault detectors under that shift. The bound separates the nominal miscoverage rate into an exchangeable component and a delay-induced component that scales with the total variation distance between the calibration and deployment score distributions. A replication-based evaluation of the ['Intercept', 'exposure', 'group', 'severity'] estimator was conducted: 600 data sets of N = 500 were drawn from a declared data-generating process (intercept 12.0, exposure 0.42, group -3.1, severity 1.35, error SD 4.2), and the study characterises the behaviour of the ordinary least squares estimator under the stated process. The estimator recovered the declared coefficients with estimated bias 0.01122 for the intercept and -0.0002 for exposure, root mean squared error 0.9217 and 0.01614 respectively, and empirical coverage of the nominal 95% intervals of 1.0 for both parameters across 600 replications. The coverage audit shows that the delay-induced term, not the exchangeable term, dominates the risk of a conformal spray-nozzle detector, and that the audit procedure can certify a detector before field labels return.","在精准农业中，喷雾喷嘴故障检测是一个分类问题，其中确认故障的标签在喷雾作业结束后才到达，因此检测器必须基于未标记的证据采取行动，并且其错误率必须在标签存在之前得到认证。保形预测（conformal prediction）提供了无分布覆盖保证，但该保证是针对可交换的校准数据和测试数据而言的，而标签延迟反馈违反了这一假设，因为校准集总是比部署集更旧。本分析将标签延迟反馈形式化为一致性分数中的分布偏移，并推导了在该偏移下分裂保形故障检测器的覆盖界。该界将名义错误覆盖 rate 分解为可交换分量和延迟诱导分量，后者随校准与部署分数分布之间的全变差距离而缩放。对 ['Intercept', 'exposure', 'group', 'severity'] 估计器进行了基于重复的评估：从声明的数据生成过程（截距 12.0，暴露 0.42，分组 -3.1，严重程度 1.35，误差标准差 4.2）中抽取了 600 个 N = 500 的数据集，研究刻画了普通最小二乘估计器在所述过程下的行为。该估计器恢复了声明的系数，截距的估计偏差为 0.01122，暴露为 -0.0002，均方根误差分别为 0.9217 和 0.01614，两个参数在 600 次重复中名义 95% 区间的经验覆盖率为 1.0。覆盖审计表明，延迟诱导项而非可交换项主导了保形喷雾喷嘴检测器的风险，并且该审计程序可以在田间标签返回之前对检测器进行认证。",62,{"impact":253,"substance":17,"depth":189,"authority":20,"freshness":13,"relevant":21,"comment":429},"论文提出标签延迟下共形预测覆盖界，方法有新意但属预印本、应用面窄，可作为专业补充而非每日精选头条。",[431],{"name":250,"url":424},[26,27,433,434],"精准施药","病虫害监测",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},[],{"status":39,"locked":14,"version":40,"method":41,"evidence":438},{"candidates":439,"agricultural_evidence":46},[440,444,450],{"code":32,"score":20,"supported":46,"title":441,"abstract":442,"tags":443},[142],[],[27],{"code":60,"score":45,"supported":46,"title":445,"abstract":446,"tags":449},[],[447,448],"精准农业","precision agriculture",[26],{"code":451,"score":21,"supported":14,"title":452,"abstract":453,"tags":454},"F4",[],[],[433],[456,457],"农业人工智能 病虫害监测 智慧农业 精准施药","农业人工智能 病虫害监测","农业人工智能病虫害监测智慧农业精准施药-5589","10.5281\u002Fzenodo.23544107",{"doi":459,"openalex_id":461,"authors":462,"venue":250,"cited_by_count":93,"oa_url":424,"card":465,"direction":98,"ingested_from":100},"W7223116087",[463],{"name":464,"orcid":9},"Achieng' Atim Okello",{"tldr":466,"method":467,"finding":468,"direction":98,"opportunity":469},"将标签延迟反馈建模为共形分数分布偏移，推导覆盖率界并审计喷雾喷嘴故障检测器。","分裂共形预测、总变差距离、600次重复模拟（N=500）评估OLS估计。","延迟诱导项而非可交换项主导检测器风险，审计可在田间标签返回前认证检测器。","可将该覆盖率审计框架扩展到其他标签延迟的农业传感器故障检测，并研究延迟分布估计误差对认证的影响。","2026-10-22T11:21:02.845718Z"]