etap安装教程(ets安装教程)

etap安装教程(ets安装教程)nbsp nbsp nbsp nbsp key nbsp nbsp nbsp nbsp nbsp nbsp nbsp nbsp image nbsp nbsp nbsp nbsp nbsp nbsp nbsp nbsp value nbsp nbsp nbsp nbsp nbsp nbsp nbsp nbsp



{
    "key":[
        "image"
    ],
    "value":[
        "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"
    ]
}





















以人才引领产业发展,以技术驱动产业升级,成为激活创新动能的重要抓手。 1月9日,由海尔海纳云、百度飞桨、中科曙光、山东大学联合中国海洋大学、中国石油大学等27所高校共同发起的公益性创新平台——青岛智能物联网产才融合中心正式成立。 该中心以“大学与产业协同”为发展战略,致力于建设成为国际一流的“人才培养中心、产业培育中心、技术开发中心”。 将通过组织人工智能公益教育资源、全国智能物联网和数据科学系列年度赛事大赛、课题申报等多形式,形成产学研用一体化的人才培养和产业升级新机制。




人工智能成为新一轮科技革命和产业变革的重要驱动力量,人才储备是推进人工智能技术创新应用和产业发展的重要工作,各地政府、企业、高校等都在开启深度融合培养人才的创新行动。

青岛智能物联网产才融合中心面向27家合作高校,由山东大学结合百度飞桨AI Studio学习与实训社区、海尔海纳云智能物联网平台、中科曙光智能计算平台,正式推出涵盖线上线下六门课程的“人工智能微专业”,每年将惠及超过1万名高校学子。





据介绍,该“人工智能微专业”课程从零基础开始,通过大量实践教学案例和“无人驾驶”及“智慧城市”两个大型产业实训项目,让学生充分了解企业真实的开发场景,理论与动手实践相结合,完整掌握人工智能产业的数据采集、模型训练和工程部署技能。完成六门课程的学生能够获得山大与合作企业认证的微专业证书,并进入合作企业实习。高校与企业共同建起从课程学习、项目实践、专业认证到实习就业的人才培养新模式。

山东大学党委常委、副校长、威海校区校长刘建亚,青岛市委组织部副部长吴学新,百度深度学习技术平台部高级总监马艳军,曙光智算副总裁明立波,海纳云副总、CTO金岩,山东大学数据科学研究院副院长郭亮,海纳云IoT平台总经理马国庆 以及来自政府、企业、高校、科研院所等的各界代表出席青岛智能物联网产才融合中心成立仪式。





人工智能与产业发展场景的结合越来越多化和专业化,更需要跨专业、跨领域的新型复合型人才去支撑行业的发展。作为我国首个自主研发、功能丰富、开源开放的产业级深度学习平台,飞桨面向技术和产业发展需求,一方面保持核心技术创新突破,已具备开发便捷的深度学习框架、超大规模深度学习模型训练技术、多端多平台部署的高性能推理引擎以及产业级开源模型库四大技术优势;另一方面,飞桨开源开放,集开发者和产学研各界力量,建设和壮大生态,培养AI人才,实现技术、产业、人才和生态共享共创。

飞桨产业级深度学习开源开放平台由核心框架、基础模型库、开发套件、工具组件以及飞桨AI Studio学习与实训社区构成。飞桨AI Studio注册用户超过120万。数据显示,几乎全国每一所高校都有学生在飞桨AI Studio进行学习;每年提供价值过亿的免费算力支持,积累了240万+实训项目、6000+课时课程、110+场竞赛及8万+数据集。

截至2021年12月,百度飞桨已凝聚406万开发者,创建47.6万个模型,服务15.7万家企事业单位,在中国深度学习平台综合市场份额第一。飞桨坚持构建人工智能开源生态,与硬件芯片厂商广泛适配并融合创新,建设飞桨人工智能产业赋能中心推动区域人工智能产业发展,举办深度学习师资培训班、AICA首席AI架构师培养计划、AI快车道、AI私享会等系列活动,培养多层次、多领域的复合型AI人才,为我国AI产业发展提供人才保障。




27家合作高校名单:
中国海洋大学、中国石油大学、山东大学、山东理工大学、山东建筑大学、山东科技大学、山东财经大学、山东师范大学、临沂大学、济南大学、齐鲁工业大学、曲阜师范大学、山东交通学院、菏泽学院、青岛大学、青岛理工大学、青岛科技大学、哈尔滨工业大学(威海)、哈尔滨工程大学(青烟研究生院)、河北师范大学、河南师范大学、河南工程学院、贵州师范大学、贵州理工学院、珠海科技学院、杭州师范大学、北部湾大学。




关注公众号,获取更多技术内容~

2021年9月23日PaddlePaddle Hackathon飞桨黑客马拉松正式上线,历经100+天完美收官啦!本次活动有2300+位社区开发者参与到飞桨开源社区的贡献中,你加入了吗?







线上开发任务,有 1800+开发者 报名参与, 297支队伍 认领任务,完成了 477余次 报名&PR提交,为飞桨社区贡献 60余个PR 并被合入到飞桨框架中。




线下Coding Party于12月18日在全国7个线下会场:上海、北京、成都、海口、西南大学、中国矿业大学、武汉科技大学以及线上开展,500余位开发者报名,最终 34支队伍 100+开发者 两天一夜 ,同步展开 48小时不间断 Coding。

这34个项目也获得了来自大众评审的认可,经过14天的线上投票, 2021飞桨黑客马拉松Coding Party人气奖诞生啦,恭喜以下项目:





来看看飞桨黑客马拉松到底怎么玩:




PS:什么,你说没赶上这次活动,很可惜?不着急,2022年飞桨黑客马拉松即将上线,欢迎开发者们参与到飞桨开源社区中,与我们一起共建。




飞桨黑客马拉松:

PaddlePaddle Hackathon 2021飞桨黑客马拉松,由百度飞桨联合深度学习技术及应用国家工程实验室主办,MLflow、Kubeflow、TVM 等开源项目,及启智OpenI、上海白玉兰开源开放研究院、木兰开源社区、Datawhale开源社区等共同出品,是面向全球开发者的深度学习领域编程活动。
关注公众号,获取更多技术内容~

生命体中,大量的奇妙数据与人类未来息息相关,而人工智能技术的日渐成熟,使得诸多研究领域中数据处理、计算精度等传统问题得以解决,生命科学也正迎来数据驱动的新时代,计算机科学与生命科学的结合势不可挡。







那么,当生命科学遇上AI,会产生怎样1+1>2效果?




1月15日, 飞桨联合百图生科 带来主题为 “人工智能助力生命科学新发展” 的技术交流会,本次Meetup邀请了三位行业专家,围绕人工智能技术在细胞图像数据处理、基因组数据挖掘、蛋白质结构研究、药物合成等专题,分享人工智能技术在生物、制药等领域的前沿应用,探讨人工智能如何赋能生命科学,助力生命科学新发展。




活动介绍

活动时间:1月15日 14:30





活动地点:

北京 | 中关村创业大街百度大脑创新中心





报名要求

从事计算机/人工智能或相关交叉方向研究2年以上,且拥有博士学位(或博士在读);飞桨博士会成员可免审核参加。





飞桨博士会简介

飞桨博士会是由百度开源深度学习平台飞桨(PaddlePaddle)发起的中国深度学习技术俱乐部,旨在打造深度学习核心开发者交流圈,成员皆为博士,且具备深度学习多年研究和实践经验。此前飞桨博士会已举办多期线下沙龙,组织会员共同研讨自然语言处理(NLP)、计算机视觉(CV)、AutoDL自动深度学习建模技术、AI+科学计算等前沿技术。




分享主题介绍

分享一:《人工智能赋能新药研发》

宋乐
百图生科首席AI科学家
讲师介绍
曾任美国佐治亚理工学院计算机学院终身教授、机器学习中心副主任,阿联酋MBZUAI机器学习系主任,蚂蚁金服深度学习团队负责人(P10)、阿里巴巴达摩院研究员,国际机器学习大会董事会成员。自2008年起,宋乐博士在CMU从事生物计算相关的研究,利用机器学习技术对靶点挖掘、药物设计取得了一系列突破性成果,获得NeurIPS、ICML、AISTATS等主要机器学习会议的最佳论文奖。曾担任NeurIPS、ICML、ICLR、AAAI、IJCAI等AI顶会的领域主席,并将出任ICML 2022的大会主席,同行评议期刊JMLR、IEEE TPAMI的副主编。




议题介绍
本次分享将探讨AI如何在制药领域发挥作用,分享包括生物制药行业面临的现状和问题、产业解决方案、生物计算领域的研究进展、以及AI+生物制药的前沿探索等话题。

分享二:《用计算机视觉技术理解细胞生命》

杨戈
中国科学院自动化研究所研究员




讲师介绍
杨戈博士,现任中国科学院自动化研究所模式识别国家重点实验室研究员,中国科学院大学人工智能学院长聘教授。毕业于清华大学、中国科学院自动化研究所,是美国明尼苏达大学双城分校机器人学博士、美国斯克利普斯研究所计算细胞生物学博士后。曾担任美国卡内基梅隆大学生物医学工程系和计算生物学系副教授。曾获美国国家科学基金早期职业奖,国际生物医学图像会议最佳论文奖等。主要研究方向为计算生物学与人工智能。




议题介绍
细胞是构成地球生物体的基本结构和功能单。例如每个人体由约37万亿个细胞组成。理解细胞生命过程对于理解生命的本质和研发战胜人类疾病的药物至关重要。细胞的生命过程由成千上万生物大分子之间的相互作用驱动,具有复杂的时空动力学行为。计算机视觉技术在理解细胞生命过程的时空动力学行为方面发挥了关键作用。通过介绍相关的代表性研究工作,本报告将展示如何创新计算机视觉技术理解复杂细胞生命过程的内在时空规律以及如何由此驱动创新药物的研发。




分享三:《长读长测序及其在生物医疗中的应用》

高欣
百图生科CEO生物计算顾问
兼蛋白质AI主任科学家
讲师介绍
世界著名研究型大学沙特阿卜杜拉国王科技大学(KAUST)计算机科学系终身正教授、计算生物学中心副主任、智慧医疗中心副主任、结构和功能生物信息学课题组负责人,生物计算领域的著名学者。在生物信息及机器学习的顶级期刊和会议上发表论文270多篇,是超过50个美国及国际专利的第一发明人,担任领域5个期刊的副主编、4个国际特刊的特邀总编,担任了13个国际会议的主席或共同主席、45个国际会议的(资深)程序委员会委员,并受邀成为多国的基金评审专家。他在蛋白质AI领域完成了一系列开创性的研究,涉及蛋白质结构预测、靶点表位分析、抗体构型等关键问题。




议题介绍
纳米孔测序作为长读长测序的代表性技术,具有便携性、长读长、及不需要PCR扩增等众多优势。但是,纳米孔测序的数据处理含有一系列的技术难题,因此严重阻碍了纳米孔技术的广泛应用。在这次分享中,我将介绍我们团队研发的端到端的纳米孔测序数据处理的计算分析平台,以及如何用这个平台解决各个相关科学领域的关键性问题。




欢迎大家扫描海报中的二维码或阅读原文,进行报名。

关注公众号,获取更多技术内容~

每当夜深人静时,你打开网易云音乐,或听歌刷乐评,或看直播闲聊,享受着以心交心的放松愉悦。在这背后,有一群技术人员苦思冥想地探索着,只为让“云村”越来越懂你。




“不同于一般的聊天文本或图片,音乐本身是跨域数据,具备若干特征,数据维度非常多。而1.8亿月活海量用户的音乐相关数据,带来的计算量、推荐量、参数规模都巨大无比。”网易云音乐机器学习平台技术团队意识到,在这样复杂问题面前,传统机器学习方法渐渐无力招架。




此外,云音乐的直播业务兴起,商业化表现良好,团队的担子更重了,“直播行为与音乐行为差异甚大,这意味着计算量与难度进一步增加。”压力之下,该团队将目光瞄向“图神经网络”,并最终选择应用百度飞桨PGL图神经网络技术来迭代升级云音乐的推荐系统。


作为全球知名音乐社区,网易云音乐在繁荣发展的同时,其推荐系统面临三大难题:囊括音乐、歌单、Mlog、直播、云圈、动态等的多域数据;海量用户产出的超大规模数据;超30万音乐人发布歌曲,超28亿用户产生歌单,27%用户交流/生产内容构成的动态数据标签。




传统的机器学习方法需要严格制定一套规范来提取样本,逐项指定样本的各个特征。但云音乐用户产生的多域数据,可能会有若干个特征,加上近2亿的用户规模以及高频率的动态更新,必须进行巨量的计算,机器学习方法的训练效率因此大受限制,变得十分低效。




而图神经网络技术的约束性较小,把每个用户当做点,用户的标签作为边,不同用户之间基于点和边的关联形成网,在此基础上建模分析,因此能更高效地表征、筛选某一类用户。比如,当两位素不相识的宝妈,同样爱听某些亲子歌曲时,她们在“图”中就有可触达的连接,模型会根据这些连接关系学习出合适的表征,并把这些亲子歌曲推荐给相似的用户群体。




事实上,图神经网络已经成为目前互联网企业高效表征用户与内容结构的关键技术。既能基于用户在歌曲、歌单、动态、Mlog等各方面的跨域行为联合建模;又能支持多种行为子图,如深挖用户在歌曲方面的播放、点赞行为;并支持载入用户节点的画像特征与内容节点的类型特征;还支持灵活扩展,如适用音乐业务场景的图神经网络应用能很方便地迁移改造用来支持直播业务场景。





飞桨PGL图神经网络的

三大领先能力









市面上提供图神经网络技术的厂商不少,说起选择百度飞桨PGL的原因,网易云音乐机器学习平台技术团队总结了三点:飞桨PGL支持超大规模数据的全图存储、子图检索、高效图学习三大领先能力。




团队曾经尝试过多家国内外顶级厂商的图神经网络技术,其中两家国际大厂的产品没有现成的分布式编程范式,无法高效地处理超大规模图模型训练当中遇到的图存储、分布式训练等问题,在单机层面顶多支持到千万级别或亿级别,而到了百亿甚至千亿级别,只有飞桨PGL挺住了。




据介绍,云音乐的数据规模非常庞大,数据关系即使经过裁剪也高达千亿级别以上。而飞桨PGL技术,原生支持分布式图存储和分布式采样,可将图的特征存储在不同的Server上,也支持将不同子图的采样分布式处理,并基于PaddlePaddle Fleet API来完成分布式训练,实现在分布式的“瘦计算节点”上加速计算,因而能够为云音乐处理高达百亿级别的大规模数据。




不仅如此,飞桨PGL 实现了极低成本的大规模图存储 ,这让网易云音乐技术团队非常认可。“飞桨PGL的分布式图存储方案比较灵活,适合云音乐,能快速搭起若干个分布式网络,无需专业数据库存储底层能力,存储成本降低70%+。”在4亿节点与400亿边数据这样的场景下,飞桨PGL的分布式图引擎资源,以60弹性节点(4CPU,16GB)的配置,可提供比中心化数据库更简单、更灵活的存储服务。




再者,他们团队还体验到飞桨PGL的另一个优点,即 灵活的子图检索模式 。飞桨PGL不仅预置常用模式,同时联动分布式图存储引擎,支持自定义子图检索模式,更符合业务实际需求,使用起来更顺手更高效。




飞桨PGL给网易云音乐技术团队印象最深的一次是,用不到30多台闲置老旧CPU机器在1天内训练完100个epoch数百亿边的LightGCN模型。这在业内人士听来可能会有些不可思议。“要是换成过去那种单机方案很难实现,因为内存早已爆掉了,无法存储这么巨大的图。”团队成员介绍道,也许还有其他方案能实现,但飞桨PGL的方案,性价比极高,适合大规模应用。云音乐的推荐系统采用飞桨PGL技术后,在冷门歌曲分发、云村广场、陌生人一起听等多个细分业务场景的效率都有不同程度的显著提升,最高甚至提升了近一倍。




可以说,飞桨PGL所提供的 支持超大规模数据的极低成本全图存储、灵活子图检索、高效图学习等能力 ,在云音乐的工业实践中真正用下来,发现都是能够满足实际需要的。这正是飞桨平台源于产业实践,更适合产业应用的最好证明。








飞桨PGL图神经网络

打开应用新空间









基于超大规模复杂数据的用户与内容理解是许多互联网内容企业所面临的共同课题。而飞桨PGL图神经网络技术在网易云音乐的成功落地,佐证了自身作为企业可用的高性价比超大规模图神经网络方案的强大实力,将助力这些企业高效、低成本地表征用户与内容,创建完善精准推荐机制,做用户的“知心人”,进而催生新形态新模式,从中获取商业收益。





接下来网易云音乐机器学习技术团队还将立足云音乐的实践,探索图神经网络技术与AI的深度融合创新,如构建音乐社区的用户和内容理解中台,以及基于知识图谱的图神经网络落地应用;并计划与飞桨一起反哺开源社区,助推图神经网络技术在产业界广泛落地。




关注公众号,获取更多技术内容~

1月8日、9日,飞桨领航团新年首次发车,分别在重庆和深圳(线上)举办了开发者Meetup活动。探讨话题从AI在工业视觉领域的技术与应用,到智能图像标注工具,再到如何用AI玩转手游......实用有趣的内容吸引了一众AI开发者热情参与。





1月8日 重庆

智能工业时代,AI技术可以应用到哪些场景?重庆飞桨领航团联合重庆AI+工业互联网产业基地共同举办本期Meetup,以飞桨如何赋能轨道交通巡检与工业材料质检为例,着重探讨了AI在工业视觉领域的技术与应用。








成都国铁研发总监、飞桨高级开发者技术专家胡文锐带来了基于飞桨开发套件的轨道交通基础设施智能巡检开发案例。这一检测项目对于轨道交通安全有重要作用,举例来说,基于PaddleSeg的道床积水和裂缝检测可实现数十项一体化智能巡检功能,如:接触网零部件异常识别、隧道表征病害识别、轨道部件缺陷识别等等。








为了更好地让开发者了解并上手深度学习项目,飞桨开发者技术专家康洪菠介绍了PaddleX与目标检测技术与应用,并详细讲解了基于PaddleX的钢材缺陷检测项目。这是一个将深度学习应用于传统工业材料检测的典型案例。关于该分享的更多干货内容,欢迎移步AI Studio项目主页:

https://aistudio.baidu.com/aistudio/projectdetail/?ad-from=mtup1

在实操环节,开发者们在讲师带领下亲自上手AI开发平台EasyDL,从模型创建、数据集准备、再到模型训练,体验了AI物体检测项目的全流程实践。





1月9日 深圳

由深圳飞桨领航团和OpenI启智社区共同举办的AI开发者Meetup因深圳疫情临时改为线上活动,开发者们热情不减,嘉宾分享内容精彩纷呈,将高效工作与畅快游戏的方法用AI一网打尽。




先来说说高效工作。深度学习的关键是需要大量的数据训练,而在数据训练之前,又必须先对大量的数据进行标注,作为机器学习的先导经验。本期Meetup上,百度研发工程师玉莹介绍了开源标注软件在分割领域的发展现状和交互式分割背后的算法原理,并带来了 CVer的福音:让标注效率提升10倍的方案——智能标注工具EISeg,以及医疗和遥感这两个垂类领域中的定制化智能标注实现方法。





除了提高工作效率,AI还能让你游戏玩得更畅快。飞桨开发者技术专家欧阳世雄为大家分享了如何利用飞桨深度强化学习框架PARL玩转《明日方舟》:从深度强化学习算法的发展历史、到明日方舟交互环境的构建、再到PARL快速并行训练的实现,从实战的角度讲解了如何将AI模型在手机游戏领域落地实践。关于该分享的更多干货内容,欢迎移步AI Studio项目主页:

https://aistudio.baidu.com/aistudio/projectdetail/?ad-from=mtup2





2022年飞桨领航团的AI开发者活动将点亮更多城市,探索更多有趣又硬核的话题和玩法。欢迎开发者朋友持续关注和热情参与!




与此同时,为了更好地组织开发者喜爱的活动,飞桨领航团现诚邀志愿者加入。无论你在哪个城市、哪所高校,只要你对技术有热情、对开源有兴趣、认同我们的社区文化、愿意为社区贡献出时间/力量/知识/想法,欢迎报名成为飞桨领航团志愿者!

飞桨领航团志愿者招募

飞桨领航团是飞桨开发者兴趣社区,面向所有对人工智能及深度学习领域感兴趣的开发者开放,提供丰富的本地技术沙龙、Meetup、及线上交流平台。在各个城市/高校飞桨领航团团长及成员的热情支持下,飞桨领航团已在全球建立200+社群,覆盖全国30个省级行政区、160+高校,聚集超过16000名AI开发者。





我们期待你具有以下特长之一:




你将收获:




如何加入:
关注飞桨领航团微信公众号,回复“志愿者”报名。





加入飞桨领航团,结识更多本地技术同好,共建开源社区, 共享开源成果与快乐。

关注公众号,获取更多技术内容~





西北工业大学航天学院副院长秦飞 为大家带来的演讲主题是: AI+CFD,面向空天动力的科学机器学习新方法与新范式。

它主要分为三部分:

WAVE SUMMIT+2021深度学习开发者峰会

【交叉前沿,AI共拓】论坛

新一代信息技术的发展,推动各行各业向数字化、网络化、智能化的方向发展,不断催生全新的技术、业务形态和模式,同时也引起了更为激烈的技术竞争。国家“十四五”智能制造发展规划中,“数字孪生”这个词语被多次引用,“数字孪生”也成为新一轮科技和企业变革的驱动力量。“数字孪生”的概念是指在虚拟空间内真实事物的动态孪生体。借由传感器,本体的运行状态以及外部环境数据均可实时映射到孪生体当中。







另一方面,以航天技术发展角度来说,人们对空间资源利用和深空探测的需求越来越强烈,依靠传统运载火箭进行航天发射已经难以满足快速进出空间迫切需求。吸气式重复使用航天运输系统是全球快速达到,以及廉价、快速、可靠、便捷进出空间的核心技术,是各航空航天大国竞相抢占的技术制高点。





这个技术不但可以作为高超声速飞机动力,还能够在普通机场水平起降实现卫星发射,因此,吸气式重复使用航天运输系统的动力已经成为快速响应航天运输最潜在的动力形式之一。同时为了适应未来越发复杂和频繁的多任务需求,航天运输系统要具备分布式、无人化、弹性化以及智能化特征,其核心技术之一是便是“数字孪生”体系建立。





综上所述,对于下一代基于吸气式重复使用航天运输系统,AI技术将极大支撑空天动力数字孪生体系的构建。








空天动力全生命周期

数字孪生









空天动力全生命周期的数字孪生体系主要分为三个大的阶段:第一是设计阶段,第二是制造阶段,第三是服役阶段


智能设计阶段形成基于机器学习的仿真模拟和人工智能方案生成,实现智能方案的生成、优化与定型,大大加快了研发的速度;智能制造阶段基于虚拟映射技术,生产过程中的缺陷、公差都可以反映到数字模型当中,为后续服役的寿命监控等提供基础数据;智能服役阶段基于机器学习的仿真模型,融合虚实映射实现飞行物理状态、实现飞行状态的空天映射以及未来状态智能监测、智能风险预测。





通过上面的阐述我们明白了空天动力数字孪生的核心是如何将数字空间与现实空间进行实时连接,并且进行实时预测,实现上述能力的核心有两个模型:
第一个是 虚拟映射模型 ,通过物理空间感知技术和反演技术,实现实时的物理与数字空间链接; 第二个是 仿真计算模型 ,通过高精度、实时的性能计算模型,实现对物理空间的实时预测,最后才能实现智能化。




因此,模型是支撑数字孪生基石。而现在,模型也不再是单一的物理模型,将基于AI的知识与历史大数据的融合模型,是知识与数据的结晶。下面针对两个模型,把我们团队研究成果进行简要介绍:


1.虚实映射模型

主要用于发动机制造过程和飞行过程,通过植入发动机电子传感器制造过程当中感知发动机是否存在损伤或者变形以及飞行过程中感知发动机工作状态一些信息。对发动机来讲,传统传感器也可以获得发动机关键参数,但这些参数信息很零散,因此必须采取重构技术,将这些零散的信息转为连续的场信息,构建由物理空间的传感器到虚拟的数字镜像,最后对该状态进行预测。




这里我们以一个发动机作为例子:实现对发动机数据、流场的重建,通过传感器测量得到温度场是低分辨率,包含大量噪声信息的,如何通过Unet深度神经网络重构出高分辨率温度场,通过验证表明基于机器学习温度场重构的精度优于一般的插值方法,误差更低,峰值性噪比更高。






2.智能仿真模型

当前基于求解物理方程的仿真方法,计算精度与计算效率矛盾突出。比如一个大型发动机的精细化燃烧流场计算,在超算上可能要耗费几千天,而这样的计算量对于工程设计是无法接受的。







传统数据方法通过几十年发展已经进入了瓶颈,我们需要运用机器学习方法对仿真计算进行加速,机器学习的方法在以前多数用于图片识别、自动驾驶,那对于这种物理场的求解和上述问题有什么异同?




首先相同点是数值仿真的可以认为是时间序列的数据集,即有监督学习。但是又有不同点,就是样本相对几十万的大数据来讲,样本量是比较小的。另外,最关键的区别在于其有物理背景,预测值需要符合物理规律,比如质量守恒等。鉴于以上特点,我们对机器学习在物理问题中的应用,采用了三个不同的思路,三种思路应用对象也不同。

①第一个层级是,我们仅仅对仿真过程中求解的物理方程中计算量最大的项进行机器学习建模,也就是这个偏微分方程中的S项,主要反应的燃烧过程。
②第二个层级就是端到端,即将物理方程的求解结果,作为时间序列数据来进行学习,获得一个大致可以预测时间序列的模型。
③然而,我们如果直接利用现有的机器学习方法进行时间序列学习,发现会存在很多非物理的预测结果,因而,发展了第三个层级,基于物理机器学习的方法。

第一个层级基于传统学习方法,将方程中S项数据单独导出,通过机器学习进行聚类,每个子类利用神经网络进行学习,形成基于机器学习的燃烧计算模块。 以发动机大涡模拟燃烧为例子,通过机器学习利用神经网络进行学习与传统的动态自适应化学方法的计算精度是相当的,计算效率方面机器学习有5倍加速,可见机器学习方法还是比较有效的。

第二个层级,对于整个仿真过程进行端对端建模, 首先通过改变台阶高度和位置参数,构建台阶非稳态数值模拟数据库,输入一百个样本对,使用一个包含时间序列Unet结构进行训练,完成训练后,即可获得基于机器学习的快速求解器,输入台阶的参数即可获得非稳态数值模拟结果。





结果来看,我们可获得非稳态的计算结果,对结果的预测时间大幅度提高。但是,我们可以发现其中存在非物理的现象。本质是因为Unet网络将其视为特征,进行了图像的特征识别与匹配。




上述结果表明直接通过机器学习方法求解物理问题存在非物理结果,2018年,Raissi等人引入了“物理信息神经网络”(Physics-informed neural networks, PINN), 将偏微分方程及其边界条件放在损失函数中对预测值进行约束,使神经网络的预测值满足偏微分方程组,通过这种方法,可以将预测值约束于物理方程,从而实现物理量的守恒。




物理机器学习的核心就在于损失函数的构建,考虑将损失函数定义为加权求和的L2范数方程和边界条件的残差。比如对于一维瞬态对流扩散方程,他的损失函数Lf中就包含了该方程,训练过程中会保证Lf趋于0,即保证了物理方程。




利用该方法构建纯流动交互式模型可以实现实时的数值模拟,获得基于机器学习的求解器无需划分网格,构型、边界实时可变,可以实时改变形状和位置,对于将来工程优化非常有利。

以前面提到的超音速台阶为例,构造包含控制方程的损失函数,超音速的控制方程的复杂度较上文复杂了很多,非线性也更强。

下面给出我们损失函数当中包含物理方程形式,右上角给出物理方程得到的结果,把传统计算结果作为真值输入使用。在网络结构训练方面,输入集合了边界条件、网络坐标以及故事留长,训练使用过程当中用自动微分方法对损失函数高阶函数进行求解,预测过程只需要物理机器学习网络,即可自行求解,获得输出的结果。

这里给出了模型计算的结果(如下图),最上为传统数值模拟CFD的计算结果,作为真值,中间结果为包含物理约束的机器学习仿真结果,最下为不含物理约束的机器学习仿真结果。可以看到不含物理约束的机器学习仿真结果对于该问题,无法不断自行迭代获得正确的结果,而物理机器学习仿真结果和CFD计算结果基本一致,能够实现复杂物理过程的守恒预测。同时,预测计算效率较CFD提高了500余倍。如果我们将其部署于部署于APU,可实现实时的仿真与预测。






结果和展望









综上所述,传统的建模方法无法实现又快又好的模型建立,基于机器学习的模型将是推动数字转型的核心,。但与此同时,如何构建工程可信的机器学习方法与体系又是我们需要解决的问题,我们相信人工智能的方法在空天动力的应用将大有可为。

除了AI+CFD,本周六下午,飞桨博士会第11期线下交流会将带来AI+Science的另一个领域的前沿分享AI+生命科学。





此次交流会由飞桨与百图生科联合出品,邀请了三位行业专家,围绕人工智能技术在细胞图像数据处理、基因组数据挖掘、蛋白质结构研究、药物合成等领域的应用,探讨人工智能如何赋能生命科学。




下方图片了解活动详情












相关推荐





‍‍‍‍‍‍‍‍‍‍‍

语义检索相比传统基于字面关键词的检索有诸多优势,广泛应用于问答、搜索系统中。 今天小编就手把手带大家完成一个基于领域预训练和对比学习SimCSE的语义检索小系统。







所谓语义检索(也称基于向量的检索),是指检索系统不再拘泥于用户Query字面本身(例如BM25检索),而是能精准捕捉到用户Query背后的真正意图并以此来搜索,从而向用户返回更准确的结果。




最终可视化demo如下,一方面 可以获取文本的向量表示 ;另一方面可以做 文本检索 ,即得到输入Query的top-K相关文档!





语义检索 ,底层技术是语义匹配,是NLP最基础常见的任务之一。从广度上看,语义匹配可以应用到 QA、搜索、推荐、广告 等各大方向;从技术深度上看,语义匹配需要融合各种SOTA模型、双塔和交互两种常用框架的魔改、以及样本处理的艺术和各种工程tricks。

比较有趣的是,在查相关资料的时候,发现百度飞桨PaddleNLP最近刚开源了类似的功能,可谓国货之光!之前使用过PaddleNLP,基本覆盖了NLP的各种应用和SOTA模型,调用起来也非常方便,强烈推荐大家试试!

接下来基于PaddleNLP提供的轮子一步步搭建语义检索系统。整体框架如下,由于计算量与资源的限制,一般工业界的搜索系统都会设计成多阶段级联结构,主要有召回、排序(粗排、精排、重排)等模块,各司其职。
语义检索技术框架图








1.1 数据

数据来源于某文献检索系统,分为有监督(少量)和无监督(大量)两种。






1.2 代码

首先clone代码:



 
   

运行环境是:



还有一些依赖包可以参考requirements.txt。








离线建库









从上面的语义检索技术框架图中可以看出,首先我们需要一个语义模型对输入的Query/Doc文本提取向量,这里选用基于对比学习的SimCSE,核心思想是使语义相近的句子在向量空间中临近,语义不同的互相远离。





那么,如何训练才能充分利用好模型,达到更高的精度呢?对于预训练模型,一般常用的训练范式已经从 『通用预训练->领域微调』 的两阶段范式变成了 『通用预训练->领域预训练->领域微调』 三阶段范式。




具体地,在这里我们的模型训练分为几步(代码和相应数据在下一节介绍):
1.在无监督的领域数据集上对通用ERNIE 1.0 进一步领域预训练,得到 领域ERNIE
2.以领域ERNIE为热启,在无监督的文献数据集上对 SimCSE 做预训练;
3.在有监督的文献数据集上结合In-Batch Negatives策略微调步骤2模型,得到最终的模型,用于抽取文本向量表示,即我们所需的语义模型,用于建库和召回。

由于召回模块需要从千万量级数据中快速召回候选集合,通用的做法是借助向量搜索引擎实现高效 ANN,从而实现候选集召回。这里采用Milvus开源工具,关于Milvus的搭建教程可以参考官方教程
https://milvus.io/cn/docs/v1.1.1/

Milvus是一款国产高性能检索库, 和Facebook开源的Faiss功能类似。
离线建库的代码位于PaddleNLP/applications/neural_search/recall/milvus
 
   





2.1 抽取向量

依照Milvus教程搭建完向量引擎后,就可以利用预训练语义模型提取文本向量了。运行feature_extract.py即可,注意修改需要建库的数据源路径。



运行结束会生成1000万条的文本数据,保存为corpus_embedding.npy。





2.2 插入向量

接下来,修改config.py中的Milvus ip等配置,将上一步生成的向量导入到Milvus库中。



 
   

抽取和插入向量两步,如果机器资源不是很"富裕"的话,可能会花费很长时间。这里建议可以先用一小部分数据进行功能测试,快速感知,等真实部署的阶段再进行全库的操作。

插入完成后,我们就可以通过Milvus提供的可视化工具[1]查看向量数据,分别是文档对应的ID和向量。








文档召回









召回阶段的目的是从海量的资源库中,快速地检索出符合Query要求的相关文档Doc。出于计算量和对线上延迟的要求,一般的召回模型都会设计成双塔形式,Doc塔离线建库,Query塔实时处理线上请求。





召回模型采用Domain-adaptive Pretraining + SimCSE + In-batch Negatives方案。

另外,如果只是想快速测试或部署, PaddleNLP也贴心地开源了训练好的模型文件,下载即可用 ,这里直接贴出模型链接:





3.1 领域预训练

Domain-adaptive Pretraining的优势在之前文章已有具体介绍,不再赘述。直接给代码,具体功能都标注在后面。



 
   


3.2 SimCSE无监督预训练

双塔模型,采用ERNIE 1.0热启,引入 SimCSE 策略。训练数据示例如下代码结构如下,各个文件的功能都有备注在后面,清晰明了。



 
   

对于训练、评估和预测分别运行scripts目录下对应的脚本即可。训练得到模型,我们一方面可以用于提取文本的语义向量表示,另一方面也可以用于计算文本对的语义相似度,只需要调整下数据输入格式即可。





3.3 有监督微调

对上一步的模型进行有监督数据微调,训练数据示例如下,每行由一对语义相似的文本对组成,tab分割,负样本来源于引入 In-batch Negatives 采样策略。







关于In-batch Negatives 的细节,可以参考文章:
大规模搜索+预训练,百度是如何落地的?
https://mp.weixin..com/s/MyVK6iKTiI-VpP1LKf4LIA




整体代码结构如下:
 
    

训练、评估、预测的步骤和上一步无监督的类似,聪明的你肯定一看就懂了!





3.4 语义模型效果

前面说了那么多,来看看几个模型的效果到底怎么样?对于匹配或者检索模型,常用的评价指标是Recall@K,即前TOP-K个结果检索出的正确结果数与全库中所有正确结果数的比值。




对比可以发现,首先利用ERNIE 1.0做Domain-adaptive Pretraining,然后把训练好的模型加载到SimCSE上进行无监督训练,最后利用In-batch Negatives 在有监督数据上进行训练能获得最佳的性能。





3.5 向量召回

终于到了召回,回顾一下,在这之前我们已经训练好了语义模型、搭建完了召回库,接下来只需要去库中检索即可。代码位于
PaddleNLP/applications/neural_search/recall/milvus/inference.py
 
   




以输入 国有企业引入非国有资本对创新绩效的影响——基于制造业国有上市公司的经验证据 为例,检索返回效果如下




返回结果的最后一列为相似度,Milvus默认使用的是欧式距离,如果想换成余弦相似度,可以在Milvus的配置文件中修改。








文档排序









不同于召回,排序阶段由于面向的打分集合相对小很多,一般只有几千级别,所以可以使用更复杂的模型,这里采用 ERNIE-Gram 预训练模型,loss选用 margin_ranking_loss。





训练数据示例如下,三列,分别为(query,title,neg_title),tab分割。对于真实搜索场景,训练数据通常来源业务线上的日志,构造出正样本和强负样本。




代码结构如下
 
   




训练运行sh scripts/train_pairwise.sh即可。




同样,PaddleNLP也开源了排序模型:
https://bj.bcebos.com/v1/paddlenlp/models/ernie_gram_sort.zip




对于预测,准备数据为每行一个文本对,最终预测返回文本对的语义相似度。
 
  








总结









本文基于PaddleNLP提供的Neural Search功能自己快速搭建了一套语义检索系统。相对于自己从零开始,PaddleNLP非常好地提供了一套轮子。如果直接下载PaddleNLP开源训练好的模型文件,对于语义相似度任务,调用现成的脚本几分钟即可搞定!对于语义检索任务,需要将全量数据导入Milvus构建索引,除训练和建库时间外,整个流程预计30-50分钟即可完成。





在训练的间隙还研究了下,发现GitHub上的文档也很清晰详细啊,对于小白入门同学,做到了一键运行,不至于被繁杂的流程步骤困住而逐渐失去兴趣; 模型全部开源 ,拿来即用;对于想要深入研究的同学,PaddleNLP也 开源了数据和代码 ,可以进一步学习,赞!照着跑下来,发现PaddleNLP太香了!赶紧Star收藏一下,持续跟进最新能力吧,也表示对开源社区的一点支持~
https://github.com/PaddlePaddle/PaddleNLP




另外我们还可以基于这些功能进行自己额外的开发,譬如开篇的动图,搭建一个更直观的语义向量生成和检索服务。Have Fun!

在跑代码过程中也遇到一些问题,非常感谢飞桨同学的耐心解答。并且得知针对这个项目还有一节视频课程已经公开,链接即可观看课程:
https://aistudio.baidu.com/aistudio/course/introduce/24902




最后附上本次实践项目的代码:
https://github.com/PaddlePaddle/PaddleNLP/tree/develop/applications/neural_search

如有疑问,欢迎添加飞桨小助手微信进用户交流群:

注:作者cafedo @NewBeeNLP

  



关注公众号,获取更多技术内容~

随着短视频的快速发展以及安全管理的需求不断增多,视频领域的相关技术应用包括视频智能标签、智能教练、智能剪辑、智能安全管理、文本视频检索、视频精彩片段提取、视频智能封面正逐渐成为人们生活中的重要部分。




以视频相关业务为例,短视频网站希望能迅速给每个新作品打上标签并推送给合适的用户,剪辑人员希望从比赛视频中便捷地提取精彩比赛片段集锦,教练员希望系统分析运动员的动作并进行技术统计和分析,安全管理部门也希望能精准地进行视频内容审核比如实时识别违规行为,编辑人员希望通过文本检索相关的视频片段作为新闻素材,广告或推荐网站希望为视频生成更加美观的封面提升转化率。这些业务对传统的人工处理方式是很大的挑战。

视频理解是通过AI技术让机器理解视频内容,如今在短视频、推荐、搜索、广告,安全管理等领域有着广泛的应用和研究价值,像动作定位与识别、视频打标签、文本视频检索、视频内容分析之类的任务都可以通过视频理解技术搞定。
PaddleVideo是百度自主研发的产业级深度学习开源开放平台飞桨的视频开发套件,包含视频领域众多模型算法和产业案例,本次开源主要升级点如下:






十大视频场景化应用

工具详解









飞桨PaddleVideo基于体育行业中足球/篮球/乒乓球/花样滑冰等场景,开源出一套通用的体育类动作识别框架;针对互联网和媒体场景开源了基于知识增强的大规模多模态分类打标签、智能剪辑和视频拆条等解决方案;针对安全、教育、医疗等场景开源了多种检测识别案例。百度智能云结合飞桨深度学习技术也形成了一系列深度打磨的产业级多场景动作识别、视频智能分析和生产以及医疗分析等解决方案。


FootballAction基于行为识别PP-TSM模型、视频动作定位BMN模型和序列模型AttentionLSTM组合得到,不仅能准确识别出动作的类型,而且能精确定位出该动作发生的起止时间。目前能识别的动作类别有8个,包含:背景、进球、角球、任意球、黄牌、红牌、换人、界外球。准确率超过90%。




篮球案例BasketballAction整体框架与FootballAction类似,共包含7个动作类别,分别为:背景、进球-三分球、进球-两分球、进球-扣篮、罚球、跳球。准确率超过90%。

在百度Create 2021(百度AI开发者大会)上,PaddleVideo联合北京大学一同发布的乒乓球动作进行识别模型,基于超过500G的比赛视频构建了标准的训练数据集,标签涵盖发球、拉、摆短等8个大类动作。其中起止回合准确率达到了97%以上,动作识别也达到了80%以上。


使用姿态估计算法提取关节点数据,最后将关节点数据输入时空图卷积网络ST-GCN模型中进行动作分类,可以实现30种动作的分类。飞桨联合CCF(中国计算机学会)举办了花样滑冰动作识别大赛,吸引了300家高校与200家企业超过3800人参赛,冠军方案比基线方案精度提升了12个点,比赛top3方案已经开源。





在视频内容分析方向,飞桨开源了基础的VideoTag和多模态的MultimodalVideoTag。VideoTag支持3000个源于产业实践的实用标签,具有良好的泛化能力,非常适用于国内大规模短视频分类场景的应用,标签准确率达到89%。

MultimodalVideoTag模型基于真实短视频业务数据,融合文本、视频图像、音频三种模态进行视频多模标签分类,相比纯视频图像特征,能显著提升高层语义标签效果。模型提供一级标签25个,二级标签200+个,标签准确率超过85%。




在视频智能生产方向,主要目标是辅助内容创作者对视频进行二次编辑。飞桨开源了基于PP-TSM的视频质量分析模型,可以实现新闻视频拆条和视频智能封面两大生产应用解决方案,其中新闻拆条是广电媒体行业的编辑们的重要素材来源;智能封面在直播、互娱等泛互联网行业的率和推荐效果方面发挥重要作用。





飞桨开源了基于MA-Net的交互式视频分割(interactive VOS)工具,提供少量的人工监督信号来实现较好的分割结果,可以仅靠标注简单几帧完成全视频标注,之后可通过多次和视频交互而不断提升视频分割质量,直至对分割质量满意。





飞桨基于时空动作检测模型实现了识别多种人类行为的方案,利用视频多帧时序信息解决传统检测单帧效果差的问题,从数据处理、模型训练、模型测试到模型推理,可以实现AVA数据集中80个动作和自研的7个异常行为(挥棍、打架、踢东西、追逐、争吵、快速奔跑、摔倒)的识别。模型的效果远超目标检测方案。

‍‍‍‍‍

禁飞领域无人机检测有如下挑战:

(1)无人机目标微小,观测困难。
(2)无人机移动速度多变。
(3)无人机飞行环境复杂,可能被建筑、树木遮挡。




针对以上挑战,飞桨开源了无人机检测模型,以实现在众多复杂环境中对无人机进行检测。








基于公开的3D-MRI脑影像数据库,浙江大学医学院附属第二医院和百度研究院开源了帕金森3D-MRI脑影像的分类鉴别项目,数据集包括neurocon, taowu, PPMI和OASIS-1等公开数据集,囊括帕金森患者(PD)与正常(Con)共378个case。提供2D及3D基线模型和4种分类模型以及3D-MRI 脑影像的预训练模型。其中PP-TSN和PP-TSM取得了超过91%的准确度和超过97.5%的AUC,而TimeSformer实现了最高准确度也超过92.3%








五大冠军、顶会算法开源









百度研究院首次开源自冠军、顶会算法

ActBERT是融合了视频、图像和文本的多模态预训练模型,它使用一种全新的纠缠编码模块从三个来源进行多模态特征学习,以增强两个视觉输入和语言之间的互动功能。该纠缠编码模块,在全局动作信息的指导下,对语言模型注入了视觉信息,并将语言信息整合到视觉模型中。纠缠编码器动态选择合适的上下文以促进目标预测。简单来说,纠缠编码器利用动作信息催化局部区域与文字的相互关联。在文本视频检索、视频描述、视频问答等5个下游任务上,ActBERT均明显优于其他方法。下表展示了ActBERT模型在文本视频检索数据集MSR-VTT上的性能表现。





随着各种互联网视频尤其是短视频的火热,文本视频检索在近段时间获得了学术界和工业界的广泛关注。特别是在引入多模态视频信息后,如何精细化地配准局部视频特征和自然语言特征成为一大难点。T2VLAD采用一种高效的全局-局部的对齐方法,自动学习文本和视频信息共享的语义中心,并对聚类后的局部特征做对应匹配,避免了复杂的计算,同时赋予了模型精细化理解语言和视频局部信息的能力。


此外,T2VLAD直接将多模态的视频信息(声音、动作、场景、speech、OCR、人脸等)映射到同一空间,利用同一组语义中心来做聚类融合,计算同一中心的视频和文本特征的局部相似度,这在一定程度上解决了多模态信息难以综合利用的问题。T2VLAD在三个标准的Text-Video Retrieval Dataset上均取得了优异的性能。





视频目标分割(VOS)是计算机视觉领域的一个基础任务,有很多重要的应用场景,如视频编辑、场景理解及自动驾驶等。交互式视频目标分割由用户在视频的某一帧中给目标物体简单的标注(比如在目标物体上画几条简单的线),就能够通过算法获得整个视频中该目标物体的分割结果,用户可以通过多次和视频交互而不断提升视频分割质量,直到用户对分割质量满意。


由于交互式视频分割需要用户多次和视频交互,因此,需要兼顾算法的时效性和准确性。MA-Net 使用一个统一的框架进行交互和传播来生成分割结果,保证了算法的时效性。另外, MA-Net 通过记忆存储的方式,将用户多轮交互的信息存储并更新,提升了视频分割的准确性。下表展示了模型在DAVIS2017数据集上性能表现。

在视频目标分割领域中,半监督领域在今年来备受关注。给定视频中第一帧或多个参考帧中的目标标定,半监督方法需要精确跟踪并分割出目标物体在整个视频中的掩模。以往的视频目标分割方法都专注于提取给定的前景目标的鲁棒特征,但这在遮挡、尺度变化以及背景中存在相似物体的等等复杂场景下是十分困难的。基于此,我们重新思考了背景特征的重要性,并提出了前背景整合式的视频目标分割方法(CFBI)。


CFBI以对偶的形式同时提取目标的前景与背景特征,并通过隐式学习的方法提升前背景特征之间的对比度,以提高分割精度。基于CFBI,我们进一步将多尺度匹配和空洞匹配的策略引入视频目标中,并设计了更为鲁棒且高效的框架,CFBI+。

CFBI系列方法在视频目标分割领域上保持着单模型最高精度的记录。特别地,百度研究院的单模型性能优于旷视清华团队在CVPR2020视频目标分割国际竞赛上融合三个强力模型的结果。在今年刚刚结束的CVPR2021视频目标分割国际竞赛中,基于 CFBI设计的解决方案在两项任务上夺得了冠军。下表展示了CFBI模型在DAVIS-2017数据集上的表现。





ADDS是基于白天和夜晚图像的自监督单目深度估计模型,其利用了白天和夜晚的图像数据互补性质,减缓了昼夜图像较大的域偏移以及照明变化对深度估计的精度带来的影响,在具有挑战性的牛津RobotCar数据集上实现了全天图像的最先进的深度估计结果。下表展示了ADDS模型在白天和夜间数据集上的测试性能表现。






是不是干货满满,心动不如行动,大家可以直接前往Github地址获得完整开源项目代码,记得Star收藏支持一下哦:
https://github.com/PaddlePaddle/PaddleVideo





精彩课程预告









1.17~1.21日每晚20:15~21:30,飞桨联合百度智能云、百度研究院数十位高工为大家带来直播讲解,剖析行业痛点问题,深入解读体育、互联网、医疗、媒体等行业应用案例及产业级视频技术方案,并带来手把手项目实战。扫码或"阅读原文"进行报名,我们直播间不见不散~


扫码报名直播课,加入技术交流群





更多相关内容,请参阅以下内容
官网地址:
https://www.paddlepaddle.org.cn
项目地址:
GitHub:
https://github.com/PaddlePaddle/PaddleVideo






参考文献:
1.ActBERT: Learning Global-Local Video-Text Representations , Linchao Zhu, Yi Yang
2.T2VLAD: Global-Local Sequence Alignment for Text-Video Retrieval, Xiaohan Wang, Linchao Zhu, Yi Yang
3.Memory Aggregation Networks for Efficient Interactive Video Object Segmentation, Jiaxu Miao, Yunchao Wei, Yi Yang
4.Collaborative Video Object Segmentation by Foreground-Background Integration, Zongxin Yang, Yunchao Wei, Yi Yang
5.Self-supervised Monocular Depth Estimation for All Day Images using Domain Separation, Liu, Lina and Song, Xibin and Wang, Mengmeng and Liu, Yong and Zhang, Liangjun

关注公众号,获取更多技术内容~
今天的文章 etap安装教程(ets安装教程)分享到此就结束了,感谢您的阅读。
编程小号
上一篇 2025-01-28 16:21
下一篇 2026-02-13 21:57

相关推荐

版权声明:本文内容由互联网用户自发贡献,该文观点仅代表作者本人。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如发现本站有涉嫌侵权/违法违规的内容, 请发送邮件至 举报,一经查实,本站将立刻删除。
如需转载请保留出处:https://bianchenghao.cn/bian-cheng-ri-ji/43758.html