[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-4860":3,"related-4860":74},{"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":8,"secondary_topics":32,"classification":33,"search_phrases":69,"slug":72,"view_count":22,"doi":8,"paper":8,"created_at":73},4860,"逛一场农交会,读懂山东田野里的新活力——2026山东(国际)农产品交易会侧记","https:\u002F\u002Fwww.toutiao.com\u002Farticle\u002F7695552857415729700","10月10日,2026山东(国际)农产品交易会在山东国际会展中心拉开帷幕,3万平方米展厅、1000多家参展商带来近万种特色农产品。山东省农业科学院展台前围满观众,科研人员现场讲解新品种培育技术,展示耐储存的果蔬新品、低耗高效的种植方案。一台小型智慧水肥一体机引得种植大户驻足问询,可依据土壤墒情自动调配水和肥料。来自湖南、四川、海南等28个省份的特色物产同台陈列,电商直播专区同步带货,把直播间的订单不停弹出。",null,"金秋十月，硕果飘香。\n\n10月10日，2026山东(国际)农产品交易会，在山东国际会展中心正式拉开帷幕。3万平方米的展厅之内，1000多家参展商带来近万种来自全国各地乃至海外的特色农产品，人流穿梭、香气弥漫，一场带着泥土芬芳的丰收盛会火热开启。\n\n从带着晨露的鲜果蔬菜，到科技赋能的智慧农业展品;从市民悠闲品鉴采购，到海内外客商洽谈签约，热气腾腾的展会现场，既是一次物产的集中亮相，更是观察山东农业高质量发展的一扇窗口。\n\n![Image 1](https:\u002F\u002Fp3-sign.toutiaoimg.com\u002Ftos-cn-i-axegupay5k\u002F0935b6a30d4a45128e0f4850574db74f~tplv-tt-origin-web:gif.jpeg?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1792541240&x-signature=lqpKSDtZTnuBqXz1eZp1QZg3hCc%3D)\n**乡土好物齐登场**\n\n走进4、5号展厅，浓郁的果蔬香气扑面而来，琳琅满目的展品铺满长长的展线。沂源红富士苹果果皮鲜亮，整齐摆放在展台上;章丘大葱挺拔修长，莱芜生姜带着湿润的泥土印记，平阴玫瑰制品散发着幽幽花香，一件件“好品山东”地标产品格外醒目。来自湖南、四川、海南、贵州等28个省份的特色物产同台陈列，腊肉、莲藕、热带鲜果、特色粮油，天南地北的风味在此相逢。\n\n不少展位设置免费品鉴台，游客排起不长的队伍，尝一块鲜果，品一杯茶汤，感受来自千里之外的味道。\n\n“就是奔着各地土特产过来的，在这里一次就能买到很多平时不好找的好物。”一位正在挑选礼盒的市民笑着说道。除了生鲜农产品，玫瑰鲜花饼、姜酒、果干、即食水产等深加工产品比比皆是，小小的展位，折射出从“卖原产品”到“卖加工品、卖品牌”的转变，让土里长出的物产拥有了更高的附加值。\n\n![Image 2](https:\u002F\u002Fp3-sign.toutiaoimg.com\u002Ftos-cn-i-tjoges91tu\u002F9d97d0a47f57365843ccf5f826013aad~tplv-tt-origin-web:gif.jpeg?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1792541240&x-signature=7JlZDO1KV7nBv6XDcN53WiyUhvE%3D)\n**科技元素亮眼出圈**\n\n如果说丰富的农品唤醒了人们的味蕾，那么新质生产力展区，则刷新着大家对于农业的固有印象。\n\n在这里，看不到传统的锄头与箩筐，取而代之的是智慧种养设备、农业物联网监测终端、节水灌溉系统、农产品快速检测仪器。山东省农科院的展台前围满观众，科研人员现场讲解新品种培育技术，展示耐储存的果蔬新品、低耗高效的种植方案。\n\n一台小型智慧水肥一体机引得不少种植大户驻足问询。工作人员介绍，这套设备可以依据土壤墒情自动调配水和肥料，大幅节约人力、提升产出。从前靠经验种地，如今靠数据种田。越来越多的山东农户，正在把实验室的技术搬到田间地头。\n\n展厅的一角，电商直播专区灯火通明，主播们对着镜头热情推介，线上线下同步带货，直播间的订单不停弹出。新技术不仅改变了耕种方式，也打通了销售链路，为乡土产品插上数字化的翅膀，让丰收不再愁销路。\n\n![Image 3](https:\u002F\u002Fp3-sign.toutiaoimg.com\u002Ftos-cn-i-tjoges91tu\u002Fd7b6f30032c9cf0f9adf2d712db2dc0e~tplv-tt-origin-web:gif.jpeg?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1792541240&x-signature=55IR96EeMUCB9hzk1umddDen%2Bn0%3D)\n**搭起供需对接桥梁**\n\n热闹的展会，不只是市民的“赶集日”，更是产业对接的“交易会”。\n\n专门设置的大买家对接专区，成了展馆里另一处火热的地方。来自印尼、马来西亚、泰国以及国内各大商超、批发市场、电商平台的约500名采购商，和参展企业面对面洽谈，核对产品参数、商量供货价格，一笔笔合作意向在这里萌生。\n\n“我们一直想把山东的苹果推向东南亚市场，这次专门过来寻找稳定的供货伙伴。”一位境外采购商告诉记者。展会还配套举办产业论坛、进出口政策解读、乡村电商培训等一系列活动，不止做一次性的买卖，更着眼长期稳定的产业链接。\n\n东西部协作专区内，来自帮扶地区的农产品集中展销，以市场的力量助力协作共赢。政府指导、市场化运营的办展模式，打破地域的壁垒，让资源流动起来，把流量转化成实实在在的订单，把产地和市场紧紧联系在一起。\n\n![Image 4](https:\u002F\u002Fp3-sign.toutiaoimg.com\u002Ftos-cn-i-tjoges91tu\u002Fba26bb8033ecd4aa79c602959b375d11~tplv-tt-origin-web:gif.jpeg?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1792541240&x-signature=Tjy5QrBTSfilOjWUlzl3Q8kM99o%3D)\n一场农交会，一端连着广阔的田野，一端连着万家的餐桌。它不仅展示着琳琅满目的物产，更呈现出齐鲁大地上产业升级、科技助农、品牌崛起的蓬勃图景。从一粒种子到一桌美味，从偏僻乡村走向广阔的国内国际市场，一个个蓬勃生长的产业，正在绘就乡村振兴的壮美画卷。","舜网发布 今日头条 2026 年 10 月 12 日","2026-10-12T00:00:00Z","报道",10,false,57,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},18,14,12,6,7,1,"省级农交会侧记，展现智慧农业与产销对接亮点，但属综合报道、信息增量有限，可作主题聚合素材而非每日精选。",[25],{"name":10,"url":6},[27,28,29,30,31],"智慧农业","电商直播","东西部协作","农产品品牌","山东农交会",[],{"status":34,"locked":14,"version":35,"method":36,"evidence":37},"needs_review","2026-10-22.2","rules",{"candidates":38,"agricultural_evidence":42,"reason":68},[39,47,53,58,63],{"code":40,"score":41,"supported":42,"title":43,"abstract":44,"tags":46},"F3",3,true,[],[45],"肥料",[],{"code":48,"score":41,"supported":42,"title":49,"abstract":50,"tags":52},"H2",[],[51],"土壤",[],{"code":54,"score":22,"supported":14,"title":55,"abstract":56,"tags":57},"E5",[],[],[30],{"code":59,"score":22,"supported":14,"title":60,"abstract":61,"tags":62},"G6",[],[],[29],{"code":64,"score":22,"supported":14,"title":65,"abstract":66,"tags":67},"I1",[],[],[27],"多个产业证据相当，需要确认主归属",[70,71],"智慧水肥一体机 山东省农科院","东西部协作 农产品品牌 山东农交会 智慧农业","智慧水肥一体机山东省农科院-4860","2026-10-14T00:04:08.269127Z",{"total":20,"page":22,"page_size":20,"items":75},[76,146,197,268,332,384],{"id":77,"title":78,"url":79,"summary":80,"summary_zh":8,"content":81,"source_name":82,"source_url":8,"published_at":83,"category":12,"cover_url":8,"hotness":84,"is_selected":14,"score":85,"score_detail":86,"sources":91,"tags":102,"primary_topic":107,"secondary_topics":113,"classification":114,"search_phrases":141,"slug":144,"view_count":22,"doi":8,"paper":8,"created_at":145},4255,"AI赋能农业如何走出建得起养不起怪圈","https:\u002F\u002Fwww.farmer.com.cn\u002F2026\u002F09\u002F30\u002F991030157.html","农民日报评论：我国农业科技进步贡献率超过64%、农业生产信息化率超过30%、30余万架植保无人机年作业面积突破4.6亿亩、116个国家智慧农业创新应用项目分品种落地、16个人工智能+行动典型场景。2026年智慧农业市场规模将突破1340亿元。AI农业存在高前期投入、长回报周期、强公共属性，建议健全多元主体参与机制，以财政撬动替代财政兜底。","一、健全多元主体参与机制，以“财政撬动”替代“财政兜底”，是引导社会资本参与涉农数智化建设的核心路径。由于涉农数智化建设具有多元受益属性，必须在强化政府引领作用的基础上构建“政府主导+社会协同”机制，通过投入资金多源化有效降低政府财政负担。一方面，激发内生动力，有效组织农业生产者仍是首要任务。组织化程度直接决定着涉农数智化建设成效——通过提升新型农业经营主体的协同能力，不仅能优化生产效率，更有助于搭建科学的利益分配与风险共担长效机制，从而充分激发农业生产实际出资主体对涉农数智化建设的长远规划意识与持续投资意愿，构建起稳定的建设主体架构。另一方面，要把着力点从“买设备”转向“买服务”。当前，“农业即服务”模式快速兴起，农户无需一次性投入巨资购置智能装备，而是通过按亩付费、以租代售、分时租赁获得植保飞防、变量施肥、智能巡检等服务。地方政府应顺势而为，把有限的补贴资金从“补建设”转向“补作业、补成效”，用作业量和服务农户数说话，让财政资金四两拨千斤。与此同时，要用好信贷、担保、保险等金融工具，并抓住公共数据资源开发利用的机遇，探索涉农数据确权、估值与质押融资，把田间数据从“沉睡资源”转化为可变现资产，为轻资产的涉农AI服务商打开融资通道。在试点选择上，应优先在农业基础好、市场区位优的县市先行先试，重点在智能农机协作平台、农业低空经济、AI智能体托管种植、智慧农（牧、渔）场等市场化程度高的项目上，探索创新收益分配与风险分担模式，切实增强运营商、互联网企业及金融机构等社会资本的投资意愿。\n\n二、推广标准化生产是AI赋能农业的基础，更是降低财政投入强度的“省钱之道”。各级政府若期望AI赋能农业经济发展，必须提前在辖区内开展农业标准化体系建设。首先，只有标准化才能为涉农AI模型训练提供高质量数据。大模型竞争归根结底是数据竞争，国务院《加快农业农村现代化“十五五”规划》明确要求“建设大规模、多模态、高质量数据集，完善数据标准和规范”。通过政策引导使农业生产各环节执行统一操作规范，才能系统性生成结构化、可机读的数据；否则，分散经营造成的数据碎片化会成倍推高模型训练成本，财政投入也将事倍功半。其次，标准化可有效降低智能终端设备的落地门槛。统一技术接口与应用场景，才能让无人机、机器人、传感器、大模型真正“即插即用”，避免各搞一套“土标准”带来的重复建设和适配成本。2026年6月，农业农村部归口管理的8项农业人工智能行业标准获立项，覆盖智能装备、作物智能监测、农业农村大模型等领域，地方应主动对标国家标准、行业标准，让一次投入、多方复用成为现实。最后，标准化是分级奖补的依据。《数字乡村高质量发展行动计划（2026—2030年）》首次提出建立智慧农（牧、渔）场数字化场景认证体系，地方可据此把“感知—决策—执行—数据闭环”的能力分级量化，按级奖补、精准滴灌，推动农业生产模式从“个体经验”向“数据驱动”转型，以边际成本递减效应加速农业现代化进程。\n\n三、打造差异化品牌是利用智能营销体系实现农产品溢价的前提。良好品牌能为智能营销体系提供信任背书与内容抓手，同时为AI驱动的各类新业态提供坚实的价值支撑。例如，高市场认可度的农副产品在利用AI进行精准推送或开展虚拟主播营销时，更易获得消费者信任与青睐，从而显著提升营销转化效率。需要看到的是，品牌建设已不再是“吆喝”层面的事。农业农村部印发的《全国农产品产地市场体系发展“十五五”规划》明确提出“推进‘人工智能+’农产品流通发展”，加强农产品消费大数据分析与生产端反馈应用，打造面向市场、数据驱动、高效协同的智能供应链。这意味着，AI可以把消费端的偏好实时反馈给生产端，实现“以销定产”，既减少滞销带来的资源浪费和财政兜底压力，又让好产品卖出好价钱。2025年全国农村网络零售额已突破3万亿元，较“十三五”末增长67.6%，线上渠道为品牌溢价提供了广阔舞台。因此，尚有余力的地方政府应前瞻性布局，积极引领农企、合作社等经营主体，将农副产品品牌建设与AI技术深度融合，探索“AI+品牌溯源”“AI+内容生成”等应用，并积极参与区块链溯源认证和智慧农（牧、渔）场场景认证，用可验证的数据为品牌背书。上述举措不仅能实现本地优质农副产品的“跨区量价齐升”，更能系统性培育出具有市场号召力的农业品牌，并最终凭借品牌价值与流通效率的双重提升，加速实现农产品溢价。\n\n需要指出的是，破解财政瓶颈，归根结底要处理好“政府有为”与“市场有效”的关系。政府把钱花在标准、数据集、基础设施和联农带农机制上，把经营性环节交给市场，把运营成本摊薄到服务之中，才能真正走出“财政背包袱、设备睡大觉”的怪圈。同时也要守住一条底线：在引导社会资本和数据要素下乡的过程中，应通过明晰的数据权属和利益联结机制，确保农民不仅是数据的提供者，更是增值收益的分享者，防止“算法”成为新的中间商。唯其如此，智能化发展红利才能更多更公平惠及广大农民群众，为农业强国建设和中国式现代化提供坚实支撑。","中国农网 农民日报 2026-09-30","2026-09-30T00:00:00Z",55,82,{"impact":87,"substance":88,"depth":17,"authority":18,"freshness":89,"relevant":22,"comment":90},24,22,4,"央媒深度分析AI赋能农业的财政可持续路径，提出财政撬动、标准化、品牌溢价三大抓手，政策与数据细节扎实，但发布已逾一周，时效性偏弱。",[92,93,96,99],{"name":82,"url":79},{"name":94,"url":95},"新疆维吾尔自治区农业农村厅 2026-10-12","https:\u002F\u002Fnynct.xinjiang.gov.cn\u002Fxjnynct\u002Fc113576\u002F202610\u002F24505848b5c346999ef97ecec7734c21.shtml",{"name":97,"url":98},"农民日报电子版 2026 年 10 月 13 日","https:\u002F\u002Fszb.farmer.com.cn\u002Fnmrb\u002Fhtml\u002F2026\u002F20261013\u002F20261013_6\u002Fnmrb_20261013_13424_6_2109757807733608534.html",{"name":100,"url":101},"农民日报 转载定西市畜牧兽医局 2026 年 10 月 13 日","http:\u002F\u002Fxmj.dingxi.gov.cn\u002Fart\u002F2026\u002F10\u002F13\u002Fart_8322_1916899.html",[103,27,104,30,105,106],"数字乡村","农业人工智能","财政撬动","农业标准化",{"code":108,"slug":109,"name":110,"path":111,"parent_code":112},"F4","crop-protection","农药与绿色防控","\u002Ftopic\u002Finputs-equipment\u002Fcrop-protection","F",[],{"status":115,"locked":14,"version":35,"method":36,"evidence":116},"classified",{"candidates":117,"agricultural_evidence":42},[118,122,127,131,136],{"code":64,"score":89,"supported":42,"title":119,"abstract":120,"tags":121},[],[27],[27],{"code":108,"score":41,"supported":42,"title":123,"abstract":124,"tags":126},[],[125],"植保",[],{"code":54,"score":22,"supported":14,"title":128,"abstract":129,"tags":130},[],[],[30],{"code":132,"score":22,"supported":14,"title":133,"abstract":134,"tags":135},"I2",[],[],[104],{"code":137,"score":22,"supported":14,"title":138,"abstract":139,"tags":140},"I6",[],[],[103],[142,143],"农业人工智能 行业标准 立项","智慧农场 数字化场景认证","农业人工智能行业标准立项-4255","2026-10-06T00:04:59.914874Z",{"id":147,"title":148,"url":149,"summary":150,"summary_zh":8,"content":151,"source_name":152,"source_url":8,"published_at":153,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":154,"score_detail":155,"sources":160,"tags":162,"primary_topic":165,"secondary_topics":170,"classification":171,"search_phrases":192,"slug":195,"view_count":22,"doi":8,"paper":8,"created_at":196},1785,"从\"靠天吃饭\"到\"数智种田\"——高校以\"电子信息+\"赋能凉山乡村振兴","http:\u002F\u002Fwww.jyb.cn\u002Frmtzgjyb\u002F202609\u002Ft20260902_2111516023.html","电子科技大学经管学院依托数字供应链金融平台，使参与合作的农户农资采购成本下降10%以上、农资销售企业呆账坏账率下降60%以上。今年初夏学院副院长第四次走进凉山农户罗流芬家中孵化\"新农人\"网红，将传统\"摆摊叫卖\"升级为\"社群营销+电商直供\"现代商业模式，中草药有机肥培育的水果销售价格比普通水果高30%以上，6000亩农产品亩均产值增加20%以上。","夏末的四川凉山，安宁河谷温暖静谧。放眼望去，在一片绿色的海洋中，油桃的红格外醒目。\n\n“今年的油桃挂果饱满，口感更好了。”在冕宁县漫水湾绿色油桃种植基地，农户曾国刚向记者口算起收成，“批发价3元，零售价6元，一亩地能有5000元左右的毛收入。”\n\n收成的底气从哪儿来？“高校为我们引入‘中草药微生物菌剂’，可以降解农残，既保证了果品的生态标准，又提升了售价。”曾国刚透露，近些年，电子科技大学经济与管理学院在凉山乡村为农户种田出谋划策，自己就是受益者之一。\n\n“不仅是油桃，现在安宁河谷的辣椒、葡萄等果蔬都开始享受数智技术支持的福利。”该校经管学院2016级校友唐朝平介绍，2020年底，学校与凉山彝族自治州政府签订政产学研校地合作协议，双方以共同成立的凉山州数字农业研究院为载体，推动凉山农业的数字化转型。\n\n没有农学的高校如何助农？最初，校地双方都犯过难，搞农业指导、育种栽培不是电子科大的赛道。唐朝平深度参与研究院工作，在一次次下田和农户交流、与农贸市场打交道的过程中，发现了“突破口”。\n\n“农户单打独斗卖‘散货’，口感参差、大小不一，难以卖出好价。”唐朝平解释，“我们要做的不是帮农民把桃子卖掉，而是用标准化、可视化、数字化的经验和模式告诉市场，这里的桃子为什么值这个价。”通过数字化手段解决农业生产标准化的问题，这条思路在研究院团队的脑海中逐渐清晰。\n\n经过探索，该校以数智技术和经管智慧，在农作物产前、产中和产后深入支持农户生产，开辟出一条金融服务农业、技术支持农产、模式赋能农销的新路。产前，为降低农资采购成本，团队引入银行授信解决农户生产经营流动资金问题，研发了数字供应链金融平台；产中，引入中草药有机肥，建设数字农业平台，通过人工智能大模型为农户提供适时技术指导，培训农户记录农事活动、作物生长情况，形成完整的农产品溯源数据；产后，搭建农产品电商供应链平台，使农产品保质保量地到达千家万户。\n\n截至目前，依托数字供应链金融平台，参与合作的农户农资采购成本下降了10%以上，农资销售企业的呆账坏账率下降了60%以上。\n\n同时，该校还以“品牌标准化”赋能凉山农产品价值提升。今年初夏，该校经管学院副院长肖延高一行，第四次走进农户罗流芬家中，在葡萄大棚里出谋划策——孵化“新农人”网红，将传统“摆摊叫卖”升级为“社群营销+电商直供”的现代商业模式。\n\n“过去我们只关心农产品产量，现在开始关注产业链上下游的联动，思考如何运用数智技术提升产品的附加值。”罗流芬说。\n\n唐朝平介绍，在新形态商业模式的加持下，中草药有机肥培育的水果销售价格比普通水果高出30%以上，6000亩农产品亩均产值增加了20%以上，团队还帮助农户、农业企业设计、培育了10余个新品牌，深受好评。\n\n近年来，该校经管学院致力于将数智供应链等优势学科方向与地方经济产业发展主战场结合，大力推进“电子信息+”战略。“2024年以来，在学校的推动和支持下，经管学院与6个和电子信息相关的工科学院建立了院际合作机制，共同开展管工交叉、产教融合的人才培养及区域经济发展服务工作，深耕成渝地区双城经济圈乡村振兴，赋能数据与人工智能、机器人与智能制造、ICT（信息与通信技术）与智慧服务、智慧医疗与健康、数字文创与新媒体等产业的高质量发展，让学科发展成果在祖国大地上开花结果。”该校经管学院院长陈旭说。\n\n有了高校作后盾，曾国刚把去年10亩的油桃试种田扩大到今年的200亩。“从前靠天吃饭，如今点击手机就能管理整片果林，销路也有了新模式。”这位“新农人”脸上挂着笑容。\n\n《中国教育报》2026年09月02日 第07版","中国教育报2026年9月2日 第07版","2026-09-02T00:00:00Z",71,{"impact":17,"substance":156,"depth":157,"authority":19,"freshness":158,"relevant":22,"comment":159},20,16,5,"高校赋能凉山农业数字化转型，产前产中产后全链条支持，数据详实，模式可复制，具有行业参考价值。",[161],{"name":152,"url":149},[103,27,163,164,30],"乡村振兴","数字供应链",{"code":166,"slug":167,"name":168,"path":169,"parent_code":8},"G","rural-services","农业服务与乡村发展","\u002Ftopic\u002Frural-services",[],{"status":172,"locked":14,"version":35,"method":36,"evidence":173},"partial",{"candidates":174,"agricultural_evidence":42},[175,180,184,188],{"code":166,"score":176,"supported":42,"title":177,"abstract":178,"tags":179},9,[163],[],[163],{"code":54,"score":22,"supported":14,"title":181,"abstract":182,"tags":183},[],[],[30],{"code":64,"score":22,"supported":14,"title":185,"abstract":186,"tags":187},[],[],[27],{"code":137,"score":22,"supported":14,"title":189,"abstract":190,"tags":191},[],[],[103],[193,194],"农产品品牌 数字供应链 乡村振兴 数字乡村","农产品品牌 数字供应链","农产品品牌数字供应链乡村振兴数字乡村-1785","2026-09-07T00:04:37.472436Z",{"id":198,"title":199,"url":200,"summary":201,"summary_zh":202,"content":8,"source_name":203,"source_url":200,"published_at":204,"category":205,"cover_url":8,"hotness":13,"is_selected":14,"score":206,"score_detail":207,"sources":211,"tags":213,"primary_topic":217,"secondary_topics":222,"classification":228,"search_phrases":245,"slug":248,"view_count":249,"doi":250,"paper":251,"created_at":267},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":208,"substance":209,"depth":17,"authority":19,"freshness":13,"relevant":22,"comment":210},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。",[212],{"name":203,"url":200},[27,104,214,215,216],"小农户","植物病害识别","可解释性",{"code":218,"slug":219,"name":220,"path":221,"parent_code":166},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services",[223],{"code":132,"slug":224,"name":225,"path":226,"parent_code":227},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",{"status":115,"locked":14,"version":35,"method":36,"evidence":229},{"candidates":230,"agricultural_evidence":42},[231,236,241],{"code":218,"score":19,"supported":42,"title":232,"abstract":234,"tags":235},[233],"smallholder",[233],[214],{"code":132,"score":13,"supported":42,"title":237,"abstract":239,"tags":240},[238],"computer vision",[],[104,215],{"code":64,"score":22,"supported":14,"title":242,"abstract":243,"tags":244},[],[],[27],[246,247],"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":250,"openalex_id":252,"authors":253,"venue":203,"cited_by_count":249,"oa_url":200,"card":260,"direction":264,"ingested_from":266},"W7223137323",[254,256,258],{"name":255,"orcid":8},"Akshay Ramgoolam",{"name":257,"orcid":8},"Priya Bundhoo Appadoo",{"name":259,"orcid":8},"Jean-Noël LEspérance",{"tldr":261,"method":262,"finding":263,"direction":264,"opportunity":265},"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":269,"title":270,"url":271,"summary":272,"summary_zh":273,"content":8,"source_name":203,"source_url":271,"published_at":204,"category":205,"cover_url":8,"hotness":13,"is_selected":14,"score":206,"score_detail":274,"sources":276,"tags":278,"primary_topic":282,"secondary_topics":288,"classification":290,"search_phrases":314,"slug":317,"view_count":249,"doi":318,"paper":319,"created_at":331},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":208,"substance":88,"depth":17,"authority":19,"freshness":176,"relevant":22,"comment":275},"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。",[277],{"name":203,"url":271},[27,104,279,280,281],"灌溉决策","安全保证","试验设计",{"code":283,"slug":284,"name":285,"path":286,"parent_code":287},"H3","irrigation","水资源与灌溉","\u002Ftopic\u002Fenvironment\u002Firrigation","H",[289],{"code":132,"slug":224,"name":225,"path":226,"parent_code":227},{"status":115,"locked":14,"version":35,"method":36,"evidence":291},{"candidates":292,"agricultural_evidence":42},[293,299,305,310],{"code":283,"score":294,"supported":42,"title":295,"abstract":296,"tags":298},15,[284],[297,284],"灌溉",[279],{"code":132,"score":19,"supported":42,"title":300,"abstract":302,"tags":304},[301],"artificial intelligence",[303],"decision support",[104],{"code":48,"score":20,"supported":42,"title":306,"abstract":307,"tags":309},[],[51,308],"soil",[],{"code":64,"score":22,"supported":14,"title":311,"abstract":312,"tags":313},[],[],[27],[315,316],"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":318,"openalex_id":320,"authors":321,"venue":203,"cited_by_count":249,"oa_url":271,"card":326,"direction":264,"ingested_from":266},"W7223131205",[322,324],{"name":323,"orcid":8},"Nassima Benkhelifa",{"name":325,"orcid":8},"Yacine Aït Ahmed",{"tldr":327,"method":328,"finding":329,"direction":264,"opportunity":330},"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":333,"title":334,"url":335,"summary":336,"summary_zh":337,"content":8,"source_name":203,"source_url":335,"published_at":204,"category":205,"cover_url":8,"hotness":13,"is_selected":14,"score":338,"score_detail":339,"sources":341,"tags":343,"primary_topic":346,"secondary_topics":347,"classification":349,"search_phrases":364,"slug":367,"view_count":249,"doi":368,"paper":369,"created_at":383},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":208,"substance":88,"depth":17,"authority":19,"freshness":13,"relevant":22,"comment":340},"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。",[342],{"name":203,"url":335},[27,104,214,344,345],"作物病害识别","模型审计",{"code":218,"slug":219,"name":220,"path":221,"parent_code":166},[348],{"code":132,"slug":224,"name":225,"path":226,"parent_code":227},{"status":115,"locked":14,"version":35,"method":36,"evidence":350},{"candidates":351,"agricultural_evidence":42},[352,356,360],{"code":218,"score":294,"supported":42,"title":353,"abstract":354,"tags":355},[233],[214,233],[214],{"code":132,"score":13,"supported":42,"title":357,"abstract":358,"tags":359},[238],[],[344,104],{"code":64,"score":22,"supported":14,"title":361,"abstract":362,"tags":363},[],[],[27],[365,366],"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":368,"openalex_id":370,"authors":371,"venue":203,"cited_by_count":249,"oa_url":335,"card":378,"direction":264,"ingested_from":266},"W7223088030",[372,374,376],{"name":373,"orcid":8},"Abdi-Basid Idriss Farah",{"name":375,"orcid":8},"Amina Abdillahi Guedi",{"name":377,"orcid":8},"Moussa Omar Youssouf",{"tldr":379,"method":380,"finding":381,"direction":264,"opportunity":382},"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":385,"title":386,"url":387,"summary":388,"summary_zh":389,"content":8,"source_name":203,"source_url":387,"published_at":204,"category":205,"cover_url":8,"hotness":13,"is_selected":14,"score":390,"score_detail":391,"sources":394,"tags":396,"primary_topic":399,"secondary_topics":400,"classification":401,"search_phrases":418,"slug":421,"view_count":249,"doi":422,"paper":423,"created_at":433},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":208,"substance":17,"depth":392,"authority":176,"freshness":13,"relevant":22,"comment":393},17,"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。",[395],{"name":203,"url":387},[27,104,397,398],"精准施药","病虫害监测",{"code":132,"slug":224,"name":225,"path":226,"parent_code":227},[],{"status":115,"locked":14,"version":35,"method":36,"evidence":402},{"candidates":403,"agricultural_evidence":42},[404,408,414],{"code":132,"score":176,"supported":42,"title":405,"abstract":406,"tags":407},[301],[],[104],{"code":64,"score":21,"supported":42,"title":409,"abstract":410,"tags":413},[],[411,412],"精准农业","precision agriculture",[27],{"code":108,"score":22,"supported":14,"title":415,"abstract":416,"tags":417},[],[],[397],[419,420],"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":422,"openalex_id":424,"authors":425,"venue":203,"cited_by_count":249,"oa_url":387,"card":428,"direction":264,"ingested_from":266},"W7223116087",[426],{"name":427,"orcid":8},"Achieng' Atim Okello",{"tldr":429,"method":430,"finding":431,"direction":264,"opportunity":432},"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"]