[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-156":3,"related-156":93},{"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":24,"tags":26,"primary_topic":32,"secondary_topics":38,"classification":39,"search_phrases":88,"slug":91,"view_count":46,"doi":8,"paper":8,"created_at":92},156,"山东实施'人工智能+农业'行动：数据'种'进泥土，云端守护丰收","https:\u002F\u002Fnew.qq.com\u002Frain\u002Fa\u002F20260802A03UUS00","山东出台实施'人工智能+农业'行动方案。在寿光，新增大棚已迭代升级至第七代智能物联网'云棚'，物联网应用率超过85%，劳动生产效率较以往翻倍；省农科院发布'舜耕·粮安'农业遥感大模型；由中国农业科学院与山东极智生物联合研发的iWheat小麦智能设计育种平台整合3万余份小麦种质资源，结合快速繁育体系能将传统8年育种周期压缩2至3年。2025年全省应用水肥一体化技术的粮油作物面积达到630万亩。",null,"技术的生命力在于解决实际问题。放眼全国，农业面临着劳动力老龄化、气候变化风险加剧等压力，也存在特色产业附加值低、品牌力不强等问题。人工智能的介入推动农业从“靠经验”转向“靠数据”，为端牢“中国饭碗”、塑强农业产业注入确定性。在不少人看来，人工智能的应用推广，恰逢其时。\n\n“山东是农业大省，也是农业产业化发源地，产业规模大、业态全，生产者队伍素质高、接受新事物能力强。与此同时，山东工业基础雄厚，人工智能技术发展迅速。新技术在农业方面的转化推广具有显著的成本和效率优势。”省委党校（山东行政学院）教授、山东乡村振兴研究院院长刘岳认为，山东有着发展“人工智能+农业”的坚实基础，在完整的产业体系、可观市场规模、丰富应用场景等优势加持下，田野阡陌将释放出更多可能。\n\n夏日的鲁北平原，玉米拔节生长，一片生机盎然。在齐河县金穗粮食种植专业合作社的示范田里，一套黑色管网发挥着大作用。“这是我们铺设的水肥一体化设施。每30厘米有一个出水孔，能精准滴灌到玉米根底部位置。相比大水漫灌，它的灌溉速度快3倍，可省水30%，一人一天能管理200亩地。”合作社理事长袁本刚说。依托大数据平台，合作社还可在“云端”实施全程监管，动动手指就能依据检测数据制订分地块水肥管理方案，为各村玉米田定制个性化基肥配方，同步开展测土配方、作物叶绿素检测等服务。\n\n今年山东夏粮逆势丰收，连续4年实现面积、单产、总产“三增”，这离不开水肥一体化技术在大田的推广。数据显示，2025年，全省应用水肥一体化技术的粮油作物面积已达到630万亩。“水肥一体化与减垄增地、‘大小行’密植等结合，加速了良种、良法、良机、良田深度融合，有效提升了农田规模化、标准化种植水平。”袁本刚体会深切，水肥一体化的落地，让丰收有了更充实的底气。\n\n种子是农业的“芯片”，也是保障粮食安全的关键，智能育种已成为育种科研人员聚焦的热点领域。\n\n今年4月，由中国农业科学院作物科学研究所与山东极智生物科技有限公司联合研发的“iWheat小麦智能设计育种平台”发布。对青岛农业大学小麦遗传育种团队来说，iWheat的应用正在帮助其攻克小麦育种“周期长、效率低、成本高”等行业难题。\n\n这一平台整合3万余份小麦种质资源的基因型、442个功能基因、6000余个功能位点及40余万条表型数据，为每一份种质建立起“数字身份证”，实现资源的精准检索与高效利用。“结合快速繁育体系，能将传统8年的育种周期压缩2至3年。”团队负责人张玉梅说，加速新品种的培育进程、降低育种成本，意味着更能提升整个种业行业的创新能力和竞争力，进而赋能小麦产业链上下游实现高质量发展。","新华社 2026-08-02 \u002F 腾讯网转载","2026-08-02T00:00:00Z","报道",10,false,85,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},24,22,18,13,8,1,"报道山东实施'人工智能+农业'行动，含具体案例与数据，省级政策影响，央媒权威，时效性稍差。",[25],{"name":10,"url":6},[27,28,29,30,31],"智慧农业","山东","人工智能+农业","寿光云棚","舜耕粮安",{"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},[45,52,59,65,71,77,83],{"code":33,"score":46,"supported":47,"title":48,"abstract":49,"tags":51},4,true,[],[50],"大模型",[29],{"code":53,"score":54,"supported":47,"title":55,"abstract":56,"tags":58},"A2",3,[],[57],"小麦",[],{"code":60,"score":54,"supported":47,"title":61,"abstract":62,"tags":64},"F1",[],[63],"种质资源",[],{"code":66,"score":54,"supported":47,"title":67,"abstract":68,"tags":70},"F3",[],[69],"水肥一体化",[],{"code":72,"score":54,"supported":47,"title":73,"abstract":74,"tags":76},"I3",[],[75],"遥感",[],{"code":78,"score":54,"supported":47,"title":79,"abstract":80,"tags":82},"I5",[],[81],"物联网",[],{"code":84,"score":22,"supported":14,"title":85,"abstract":86,"tags":87},"I1",[],[],[27],[89,90],"人工智能+农业 寿光云棚 智慧农业 舜耕粮安","人工智能+农业 寿光云棚","人工智能+农业寿光云棚智慧农业舜耕粮安-156","2026-08-01T23:53:42.061928Z",{"total":94,"page":22,"page_size":94,"items":95},6,[96,167,220,273,341,406],{"id":97,"title":98,"url":99,"summary":100,"summary_zh":8,"content":101,"source_name":102,"source_url":8,"published_at":103,"category":104,"cover_url":8,"hotness":13,"is_selected":14,"score":105,"score_detail":106,"sources":109,"tags":111,"primary_topic":115,"secondary_topics":116,"classification":117,"search_phrases":162,"slug":165,"view_count":94,"doi":8,"paper":8,"created_at":166},277,"山东省委十二届十一次全会明确实施「人工智能农业」行动——构建农业大模型+智能体落地体系,推动农业从「靠经验」转向「靠数据」","https:\u002F\u002Fweibo.com\u002F7469265487\u002F5327769942885036","山东省委十二届十一次全会着眼把农业建成现代化大产业,推动农业数智化转型升级,明确提出实施「人工智能农业」行动。多家行业机构预测,2026年中国智慧农业市场规模将突破1340亿元,年复合增长率保持在15%以上。山东行动重点:水肥一体化方面,2025年全省应用水肥一体化技术的粮油作物面积已达630万亩,相比大水漫灌可省水30%;智能育种方面,4月由中国农业科学院作物科学研究所与山东极智生物科技有限公司联合研发的「iWheat小麦智能设计育种平台」发布,整合3万余份小麦种质资源基因型、442个功能基因、6000余个功能位点及40余万条表型数据,结合快速繁育体系将传统8年育种周期压缩2-3年;农业大模型方面,山东省农科院7月发布「舜耕·粮安」农业遥感大模型,基于20年积累的2000万份人工样本,覆盖山东全域全地貌、全物候周期,构建「耕地智能提取—肥力墒情预测—作物精准识别—长势动态监测—病虫草害预警—作物智能估产」全链条智能监测体系,精度高达95%,效率提升超100%;智能装备方面,寿光新建大棚已迭代升级至第七代智能物联网「云棚」,物联网应用率超过85%,劳动生产效率较以往翻倍;下游加工方面,浪潮云洲联合中国农业大学计算中心、烟台葡乡农业开发公司在莱山区共同打造「莱山葡萄种植和酿造大模型」,实现种植端气候适应性调控与酿造端AI模拟发酵动态优化。","从“靠天吃饭”到“知天而作”，我省实施“人工智能+农业”行动 数据“种”进泥土，云端守护丰收\n\n□ 本报记者 于新悦 刘兵\n\n 麦田上空，无人机低空掠过，哪片麦苗缺氮、哪处有蚜虫，屏幕上即时标红；大棚内，人工智能（AI）算法值守，缺水、缺肥时，系统立刻报警；分拣车间里，AI视觉系统每秒“端详”上百颗苹果，按色泽、果径、瑕疵精准分级，误判率不到百分之一……\n\n 从“靠天吃饭”到“知天而作”，当前人工智能正成为助力农业稳产提质、带动农民增收致富的“好帮手”，为农业农村现代化提供全新范式。\n\n 2026年中央一号文件明确提出“因地制宜发展农业新质生产力，促进人工智能与农业发展相结合”，将AI技术从单点应用升级为贯穿农业全产业链的战略引擎。多家行业机构预测，2026年中国智慧农业市场规模将突破1340亿元，年复合增长率保持在15%以上。\n\n 刚刚召开的省委十二届十一次全会，着眼把农业建成现代化大产业，推动农业数智化转型升级，明确提出实施“人工智能+农业”行动。何谓“人工智能+农业”？简言之，就是以人工智能技术为核心，驱动传统农业向智能化、高效化转型的新型生产体系。与会同志认为，从更深层次看，“人工智能+农业”不是简单的技术叠加，而是农业生产方式、组织形式和价值链条的系统性重塑。\n\n AI种地，田野阡陌释放出更多可能\n\n 技术的生命力在于解决实际问题。放眼全国，农业面临着劳动力老龄化、气候变化风险加剧等压力，也存在特色产业附加值低、品牌力不强等问题。人工智能的介入推动农业从“靠经验”转向“靠数据”，为端牢“中国饭碗”、塑强农业产业注入确定性。在不少人看来，人工智能的应用推广，恰逢其时。\n\n “山东是农业大省，也是农业产业化发源地，产业规模大、业态全，生产者队伍素质高、接受新事物能力强。与此同时，山东工业基础雄厚，人工智能技术发展迅速。新技术在农业方面的转化推广具有显著的成本和效率优势。”省委党校（山东行政学院）教授、山东乡村振兴研究院院长刘岳认为，山东有着发展“人工智能+农业”的坚实基础，在完整的产业体系、可观市场规模、丰富应用场景等优势加持下，田野阡陌将释放出更多可能。\n\n 夏日的鲁北平原，玉米拔节生长，一片生机盎然。在齐河县金穗粮食种植专业合作社的示范田里，一套黑色管网发挥着大作用。“这是我们铺设的水肥一体化设施。每30厘米有一个出水孔，能精准滴灌到玉米根底部位置。相比大水漫灌，它的灌溉速度快3倍，可省水30%，一人一天能管理200亩地。”合作社理事长袁本刚说。依托大数据平台，合作社还可在“云端”实施全程监管，动动手指就能依据检测数据制订分地块水肥管理方案，为各村玉米田定制个性化基肥配方，同步开展测土配方、作物叶绿素检测等服务。\n\n 今年山东夏粮逆势丰收，连续4年实现面积、单产、总产“三增”，这离不开水肥一体化技术在大田的推广。数据显示，2025年，全省应用水肥一体化技术的粮油作物面积已达到630万亩。“水肥一体化与减垄增地、‘大小行’密植等结合，加速了良种、良法、良机、良田深度融合，有效提升了农田规模化、标准化种植水平。”袁本刚体会深切，水肥一体化的落地，让丰收有了更充实的底气。\n\n 种子是农业的“芯片”，也是保障粮食安全的关键，智能育种已成为育种科研人员聚焦的热点领域。\n\n 今年4月，由中国农业科学院作物科学研究所与山东极智生物科技有限公司联合研发的“iWheat小麦智能设计育种平台”发布。对青岛农业大学小麦遗传育种团队来说，iWheat的应用正在帮助其攻克小麦育种“周期长、效率低、成本高”等行业难题。\n\n 这一平台整合3万余份小麦种质资源的基因型、442个功能基因、6000余个功能位点及40余万条表型数据，为每一份种质建立起“数字身份证”，实现资源的精准检索与高效利用。“结合快速繁育体系，能将传统8年的育种周期压缩2至3年。”团队负责人张玉梅说，加速新品种的培育进程、降低育种成本，意味着更能提升整个种业行业的创新能力和竞争力，进而赋能小麦产业链上下游实现高质量发展。\n\n 让“技术盆景”真正落地为“产业风景”\n\n 农业大模型能够整合气象、土壤、作物生长等多维度数据，显著提升农业生产的精准性和预见性。我省明确，要加强农业大模型与人工智能体落地应用。\n\n 在“中国蔬菜之乡”寿光，蔬菜大棚迎来了一批特殊“员工”。一群脚步灵活的机器狗，沿着田垄灵活穿梭，搭载的高清摄像头与传感器如同“火眼金睛”，精准识别病虫害迹象。巡检结束，数据实时传回后台，人工智能模型迅速给出种植决策。如今，寿光新建大棚已迭代升级至第七代智能物联网“云棚”，物联网应用率超过85%，劳动生产效率较以往翻倍。\n\n 就在一个多月前，省农科院发布“舜耕·粮安”农业遥感大模型。省农科院党委书记梁金光介绍，该模型基于20年来积累的2000万份人工样本的海量数据，覆盖山东全域全地貌、全物候周期，构建起“耕地智能提取—肥力墒情预测—作物精准识别—长势动态监测—病虫草害预警—作物智能估产”全链条智能监测体系，精度高达95%，效率提升超100%。\n\n 在中游加工环节，大模型也能有效提升产品附加值与标准化水平。在烟台莱山区，由浪潮云洲旗下数农云（山东）信息科技有限公司联合中国农业大学计算中心、烟台葡乡农业开发有限公司共同打造的“莱山葡萄种植和酿造大模型”，一方面在种植端自动生成气候适应性调控方案，实现水肥精准配施、病虫害智能预警与防治，葡萄产量与品质同步提升；另一方面在酿造端用AI模拟发酵、陈酿全过程，动态优化关键参数，完成原料分级、发酵控制和酒体风味数字化建模。以此，让每一批葡萄酒的口感更稳定、品质更可控。\n\n 当然，让人工智能与广袤乡村实现更深层次的双向赋能，还有很长的路要走——\n\n 一套AI系统，传感终端、边缘计算、决策模型全算上，初始投入比较大。普通农户能不能用得起、用得好？\n\n 农业生产周期长、环境复杂、区域差异大，怎样才能收集利用一批高质量农业数据，推动AI模型迭代？\n\n 下沉到村、懂技术懂农业的青年人才较为短缺，农业人工智能技术人才怎样引育？\n\n 让“技术盆景”真正落地为“产业风景”，刘岳认为，关键是鼓励龙头企业、高校院所与新型农业经营主体协同攻关，围绕粮食、果蔬、畜牧、水产等优势产业，打造一批“听得懂农事、算得清账本、管得了全程”的行业大模型。\n\n 此外，刘岳还建议，要更加重视构建与农村新质生产力相适应的农村新型生产关系，盘活整合农村各类资产资源，为充分发挥数字化、智能化技术进步作用提供保障。“推动农村数字化从‘局域网’向‘互联网’转化，要破除长期存在的‘信息孤岛’，构建乡村振兴的数据底座。”","大众日报\u002F微博 2026-08-03","2026-08-03T00:00:00Z","政策",81,{"impact":107,"substance":18,"depth":19,"authority":13,"freshness":94,"relevant":22,"comment":108},25,"省级政策明确实施人工智能+农业行动，内容详实，案例丰富，具有较高参考价值。",[110],{"name":102,"url":99},[27,28,112,31,113,114],"农业大模型","人工智能农业","iWheat",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":118},{"candidates":119,"agricultural_evidence":47},[120,125,130,134,138,144,148,152,158],{"code":33,"score":121,"supported":47,"title":122,"abstract":123,"tags":124},12,[50],[50],[112],{"code":72,"score":94,"supported":47,"title":126,"abstract":127,"tags":129},[],[75,128],"估产",[],{"code":84,"score":46,"supported":47,"title":131,"abstract":132,"tags":133},[],[27],[27],{"code":53,"score":54,"supported":47,"title":135,"abstract":136,"tags":137},[],[57],[],{"code":139,"score":54,"supported":47,"title":140,"abstract":141,"tags":143},"B2",[],[142],"葡萄",[],{"code":60,"score":54,"supported":47,"title":145,"abstract":146,"tags":147},[],[63],[],{"code":66,"score":54,"supported":47,"title":149,"abstract":150,"tags":151},[],[69],[],{"code":153,"score":54,"supported":47,"title":154,"abstract":155,"tags":157},"H7",[],[156],"气候适应",[],{"code":78,"score":54,"supported":47,"title":159,"abstract":160,"tags":161},[],[81],[],[163,164],"人工智能农业 农业大模型 智慧农业 舜耕粮安","人工智能农业 农业大模型","人工智能农业农业大模型智慧农业舜耕粮安-277","2026-08-07T23:57:47.758140Z",{"id":168,"title":169,"url":170,"summary":171,"summary_zh":8,"content":9,"source_name":172,"source_url":8,"published_at":11,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":173,"score_detail":174,"sources":180,"tags":182,"primary_topic":185,"secondary_topics":189,"classification":191,"search_phrases":214,"slug":217,"view_count":218,"doi":8,"paper":8,"created_at":219},234,"舜耕粮安农业遥感大模型发布——山东省农科院党委书记梁金光介绍全链条智能监测体系","https:\u002F\u002Fnews.cau.edu.cn\u002Fmtndnew\u002F37a295bf8e3c4e5b92cb8c665bdc8f4f.htm","山东省农业科学院发布舜耕粮安农业遥感大模型。山东省农科院党委书记梁金光介绍,该模型基于20年来积累的2000万份人工样本的海量数据,覆盖山东全域全地貌、全物候周期,构建起耕地智能提取—肥力墒情预测—作物精准识别—长势动态监测—病虫草害预警—作物智能估产全链条智能监测体系,精度高达95%,效率提升超100%。山东省同步推动人工智能+农业行动实施,在寿光推广第七代智能物联网云棚,物联网应用率超85%,劳动生产效率较以往翻倍;在烟台莱山区,浪潮云洲联合中国农业大学打造莱山葡萄种植和酿造大模型,实现水肥精准配施、病虫害智能预警、AI模拟发酵陈酿。","大众日报 2026-08-02",64,{"impact":175,"substance":176,"depth":177,"authority":13,"freshness":178,"relevant":22,"comment":179},20,15,14,5,"报道山东农业遥感大模型发布，结合水肥一体化与智能育种案例，省级产业进展，信息较充实，但时效稍旧。",[181],{"name":172,"url":170},[27,183,29,31,184],"山东省农科院","农业遥感大模型",{"code":72,"slug":186,"name":187,"path":188,"parent_code":37},"remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing",[190],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":192},{"candidates":193,"agricultural_evidence":47},[194,198,202,206,210],{"code":72,"score":177,"supported":47,"title":195,"abstract":196,"tags":197},[75],[75,128],[],{"code":33,"score":121,"supported":47,"title":199,"abstract":200,"tags":201},[50],[50],[29],{"code":139,"score":54,"supported":47,"title":203,"abstract":204,"tags":205},[],[142],[],{"code":78,"score":54,"supported":47,"title":207,"abstract":208,"tags":209},[],[81],[],{"code":84,"score":22,"supported":14,"title":211,"abstract":212,"tags":213},[],[],[27],[215,216],"农业遥感大模型 人工智能+农业 山东省农科院 智慧农业","农业遥感大模型 人工智能+农业","农业遥感大模型人工智能+农业山东省农科院智慧农业-234",9,"2026-08-05T23:53:01.179071Z",{"id":221,"title":222,"url":223,"summary":224,"summary_zh":8,"content":8,"source_name":225,"source_url":8,"published_at":226,"category":104,"cover_url":8,"hotness":13,"is_selected":14,"score":227,"score_detail":228,"sources":231,"tags":233,"primary_topic":238,"secondary_topics":239,"classification":240,"search_phrases":268,"slug":271,"view_count":54,"doi":8,"paper":8,"created_at":272},880,"北京\"十五五\"将打造具有全球影响力的农业科技创新中心 建设农业中关村","https:\u002F\u002Fbj.people.com.cn\u002FBIG5\u002Fn2\u002F2026\u002F0821\u002Fc14540-41674434.html","8月21日，北京市政府新闻办举行2026世界农业科技创新大会新闻发布会。北京\"十五五\"时期将立足京津冀国际科技创新中心定位，推进农业中关村建设，打造具有区域带动力、全国引领力和全球影响力的农业科技创新中心。重点聚焦生物育种、\"人工智能+\"农业、农业生物制造领域，实施关键核心技术攻关，培育自主原创品种；布局农业大模型、专用芯片、智能农机装备研发，打造智慧农业应用场景；推进\"前沿攻关—验证熟化—典型示范—规模推广\"一体化路径。","人民网北京 2026-08-21","2026-08-21T00:00:00Z",79,{"impact":107,"substance":175,"depth":176,"authority":121,"freshness":229,"relevant":22,"comment":230},7,"北京十五五农业科技创新中心规划，政策层级高，内容具体，央媒发布，时效性稍差。",[232],{"name":225,"url":223},[234,27,235,29,236,237],"十五五","北京","农业中关村","世界农业科技创新大会",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":241},{"candidates":242,"agricultural_evidence":47},[243,247,251,257,263],{"code":84,"score":46,"supported":47,"title":244,"abstract":245,"tags":246},[],[27],[27],{"code":33,"score":46,"supported":47,"title":248,"abstract":249,"tags":250},[],[50],[29],{"code":252,"score":54,"supported":47,"title":253,"abstract":254,"tags":256},"F2",[],[255],"生物育种",[],{"code":258,"score":54,"supported":47,"title":259,"abstract":260,"tags":262},"F5",[],[261],"农机",[],{"code":264,"score":22,"supported":14,"title":265,"abstract":266,"tags":267},"J1",[],[],[234],[269,270],"世界农业科技创新大会 人工智能+农业 农业中关村 智慧农业","世界农业科技创新大会 人工智能+农业","世界农业科技创新大会人工智能+农业农业中关村智慧农业-880","2026-08-24T00:06:08.645033Z",{"id":274,"title":275,"url":276,"summary":277,"summary_zh":278,"content":8,"source_name":279,"source_url":276,"published_at":280,"category":281,"cover_url":8,"hotness":13,"is_selected":14,"score":282,"score_detail":283,"sources":286,"tags":288,"primary_topic":293,"secondary_topics":299,"classification":301,"search_phrases":318,"slug":321,"view_count":322,"doi":323,"paper":324,"created_at":340},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","论文",69,{"impact":21,"substance":284,"depth":19,"authority":121,"freshness":13,"relevant":22,"comment":285},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。",[287],{"name":279,"url":276},[27,289,290,291,292],"农业人工智能","小农户","植物病害识别","可解释性",{"code":294,"slug":295,"name":296,"path":297,"parent_code":298},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services","G",[300],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":302},{"candidates":303,"agricultural_evidence":47},[304,309,314],{"code":294,"score":121,"supported":47,"title":305,"abstract":307,"tags":308},[306],"smallholder",[306],[290],{"code":33,"score":13,"supported":47,"title":310,"abstract":312,"tags":313},[311],"computer vision",[],[289,291],{"code":84,"score":22,"supported":14,"title":315,"abstract":316,"tags":317},[],[],[27],[319,320],"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",0,"10.5281\u002Fzenodo.23545193",{"doi":323,"openalex_id":325,"authors":326,"venue":279,"cited_by_count":322,"oa_url":276,"card":333,"direction":337,"ingested_from":339},"W7223137323",[327,329,331],{"name":328,"orcid":8},"Akshay Ramgoolam",{"name":330,"orcid":8},"Priya Bundhoo Appadoo",{"name":332,"orcid":8},"Jean-Noël LEspérance",{"tldr":334,"method":335,"finding":336,"direction":337,"opportunity":338},"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。","农业人工智能与决策模型","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。","openalex","2026-10-22T11:21:10.425290Z",{"id":342,"title":343,"url":344,"summary":345,"summary_zh":346,"content":8,"source_name":279,"source_url":344,"published_at":280,"category":281,"cover_url":8,"hotness":13,"is_selected":14,"score":282,"score_detail":347,"sources":349,"tags":351,"primary_topic":355,"secondary_topics":361,"classification":363,"search_phrases":388,"slug":391,"view_count":322,"doi":392,"paper":393,"created_at":405},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":21,"substance":18,"depth":19,"authority":121,"freshness":218,"relevant":22,"comment":348},"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。",[350],{"name":279,"url":344},[27,289,352,353,354],"灌溉决策","安全保证","试验设计",{"code":356,"slug":357,"name":358,"path":359,"parent_code":360},"H3","irrigation","水资源与灌溉","\u002Ftopic\u002Fenvironment\u002Firrigation","H",[362],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":364},{"candidates":365,"agricultural_evidence":47},[366,371,377,384],{"code":356,"score":176,"supported":47,"title":367,"abstract":368,"tags":370},[357],[369,357],"灌溉",[352],{"code":33,"score":121,"supported":47,"title":372,"abstract":374,"tags":376},[373],"artificial intelligence",[375],"decision support",[289],{"code":378,"score":94,"supported":47,"title":379,"abstract":380,"tags":383},"H2",[],[381,382],"土壤","soil",[],{"code":84,"score":22,"supported":14,"title":385,"abstract":386,"tags":387},[],[],[27],[389,390],"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":392,"openalex_id":394,"authors":395,"venue":279,"cited_by_count":322,"oa_url":344,"card":400,"direction":337,"ingested_from":339},"W7223131205",[396,398],{"name":397,"orcid":8},"Nassima Benkhelifa",{"name":399,"orcid":8},"Yacine Aït Ahmed",{"tldr":401,"method":402,"finding":403,"direction":337,"opportunity":404},"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":407,"title":408,"url":409,"summary":410,"summary_zh":411,"content":8,"source_name":279,"source_url":409,"published_at":280,"category":281,"cover_url":8,"hotness":13,"is_selected":14,"score":412,"score_detail":413,"sources":415,"tags":417,"primary_topic":420,"secondary_topics":421,"classification":423,"search_phrases":438,"slug":441,"view_count":322,"doi":442,"paper":443,"created_at":457},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":21,"substance":18,"depth":19,"authority":121,"freshness":13,"relevant":22,"comment":414},"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。",[416],{"name":279,"url":409},[27,289,290,418,419],"作物病害识别","模型审计",{"code":294,"slug":295,"name":296,"path":297,"parent_code":298},[422],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":424},{"candidates":425,"agricultural_evidence":47},[426,430,434],{"code":294,"score":176,"supported":47,"title":427,"abstract":428,"tags":429},[306],[290,306],[290],{"code":33,"score":13,"supported":47,"title":431,"abstract":432,"tags":433},[311],[],[418,289],{"code":84,"score":22,"supported":14,"title":435,"abstract":436,"tags":437},[],[],[27],[439,440],"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":442,"openalex_id":444,"authors":445,"venue":279,"cited_by_count":322,"oa_url":409,"card":452,"direction":337,"ingested_from":339},"W7223088030",[446,448,450],{"name":447,"orcid":8},"Abdi-Basid Idriss Farah",{"name":449,"orcid":8},"Amina Abdillahi Guedi",{"name":451,"orcid":8},"Moussa Omar Youssouf",{"tldr":453,"method":454,"finding":455,"direction":337,"opportunity":456},"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"]