[{"data":1,"prerenderedAt":281},["ShallowReactive",2],{"content-query-g1JZLNeTgS":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"date":10,"cover":11,"type":12,"state":13,"body":14,"_type":275,"_id":276,"_source":277,"_file":278,"_stem":279,"_extension":280},"/activities/zh/248","zh",false,"","MindSpore社区8月新社区群组介绍系列之二","MindSpore社区将持续优化各项功能提供最佳操作体验，今天向大家介绍旨在“无缝”融合深度学习模型和概率学习模型——MindSpore深度概率编程（MDP），相信在日常工作中有了MDP的协助，能更加高效的完成各项任务，让我们来一起了解MDP是什么吧~","2020-09-03","https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2020/09/21/b43c6ae1abed4444a80b1554def0c7bc.png","activities",1,{"type":15,"children":16,"toc":270},"root",[17,25,32,49,54,62,67,72,77,84,89,94,104,109,117,122,129,134,157,162,167,172,181,186,191,200,205,216,223,228,239,250,255,263],{"type":18,"tag":19,"props":20,"children":22},"element","h1",{"id":21},"mindspore社区8月新社区群组介绍系列之二",[23],{"type":24,"value":8},"text",{"type":18,"tag":26,"props":27,"children":29},"h2",{"id":28},"sig-dpp",[30],{"type":24,"value":31},"SIG-DPP",{"type":18,"tag":33,"props":34,"children":35},"p",{},[36,38,47],{"type":24,"value":37},"继昨天发布的",{"type":18,"tag":39,"props":40,"children":44},"a",{"href":41,"rel":42},"https://www.mindspore.cn/news/newschildren?id=247",[43],"nofollow",[45],{"type":24,"value":46},"MindSpore社区8月新社区群组介绍系列之一——SIG-MSLITE",{"type":24,"value":48},"（☚点击可回顾）向大家介绍了MindSpore推出的端云协同、轻量化、高性能AI推理框架MindSpore Lite，可以满足越来越多的端侧AI应用需求。",{"type":18,"tag":33,"props":50,"children":51},{},[52],{"type":24,"value":53},"MindSpore社区将持续优化各项功能提供最佳操作体验，今天向大家介绍旨在“无缝”融合深度学习模型和概率学习模型——MindSpore深度概率编程（MDP），相信在日常工作中有了MDP的协助，能更加高效的完成各项任务，接下来让我们来一起了解MDP是什么吧~",{"type":18,"tag":33,"props":55,"children":56},{},[57],{"type":18,"tag":58,"props":59,"children":61},"img",{"alt":7,"src":60},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2020/09/03/7414538185a34581abfa8e4f030decaf.gif",[],{"type":18,"tag":33,"props":63,"children":64},{},[65],{"type":24,"value":66},"为什么MindSpore社区要做深度概率编程库？",{"type":18,"tag":33,"props":68,"children":69},{},[70],{"type":24,"value":71},"相比于传统的深度神经网络模型，深度概率模型(Deep Probabilistic Programming Model )具有更好的鲁棒性和可解释性。但是深度概率模型需要用户理解概率模型的建模思维和编程方法，因此对于传统深度学习开发者而言有较高门槛。",{"type":18,"tag":33,"props":73,"children":74},{},[75],{"type":24,"value":76},"MindSpore深度概率编程库与清华大学的ZhuSuan相结合，提供与深度学习无缝融合的高效贝叶斯学习编程库，可以帮助开发者们更好的理解和学习。",{"type":18,"tag":33,"props":78,"children":79},{},[80],{"type":18,"tag":58,"props":81,"children":83},{"alt":7,"src":82},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2020/09/03/96ff4476c6b4450282400d616271421b.jpg",[],{"type":18,"tag":33,"props":85,"children":86},{},[87],{"type":24,"value":88},"MindSpore深度概率编程库特点",{"type":18,"tag":33,"props":90,"children":91},{},[92],{"type":24,"value":93},"MindSpore深度概率编程（MDP）旨在“无缝”融合深度学习模型和概率学习模型，满足不同用户的需求：",{"type":18,"tag":95,"props":96,"children":97},"ul",{},[98],{"type":18,"tag":99,"props":100,"children":101},"li",{},[102],{"type":24,"value":103},"对于深度学习的用户",{"type":18,"tag":33,"props":105,"children":106},{},[107],{"type":24,"value":108},"MDP支持深度神经网络一键转贝叶斯神经网络，方便上手，并提供不确定性估计等面向具体应用场景的工具箱；",{"type":18,"tag":95,"props":110,"children":111},{},[112],{"type":18,"tag":99,"props":113,"children":114},{},[115],{"type":24,"value":116},"对于专业的概率学习用户",{"type":18,"tag":33,"props":118,"children":119},{},[120],{"type":24,"value":121},"MDP提供通用、模块化的高性能贝叶斯建模和推断方法库，包括概率采样库、贝叶斯神经网络，概率推断算法等，帮助用户更加快速、简洁地进行概率编程。",{"type":18,"tag":33,"props":123,"children":124},{},[125],{"type":18,"tag":58,"props":126,"children":128},{"alt":7,"src":127},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2020/09/03/78a9c13cfe74432aa1f4a014516548d2.jpg",[],{"type":18,"tag":33,"props":130,"children":131},{},[132],{"type":24,"value":133},"MindSpore下一步工作计划",{"type":18,"tag":95,"props":135,"children":136},{},[137,142,147,152],{"type":18,"tag":99,"props":138,"children":139},{},[140],{"type":24,"value":141},"增加BNN layers的实现方法，包括Local Reparameteration、Flipout等典型方法；",{"type":18,"tag":99,"props":143,"children":144},{},[145],{"type":24,"value":146},"丰富面向具体应用场景的工具箱，提供OoD（Out of Distribution）检测等功能；",{"type":18,"tag":99,"props":148,"children":149},{},[150],{"type":24,"value":151},"增加马尔可夫链蒙特卡洛(MCMC)方法，用于支持数据生成和推理等模型功能；",{"type":18,"tag":99,"props":153,"children":154},{},[155],{"type":24,"value":156},"增加贝叶斯上下文管理，用于支持构建复杂的贝叶斯模型。",{"type":18,"tag":33,"props":158,"children":159},{},[160],{"type":24,"value":161},"如何参与SIG-DPP",{"type":18,"tag":33,"props":163,"children":164},{},[165],{"type":24,"value":166},"参加社区开发非常简单：",{"type":18,"tag":33,"props":168,"children":169},{},[170],{"type":24,"value":171},"1. 请先查看SIG-DPP的立项文档：",{"type":18,"tag":33,"props":173,"children":174},{},[175],{"type":18,"tag":39,"props":176,"children":179},{"href":177,"rel":178},"https://gitee.com/mindspore/mindspore/issues/I1RG6C",[43],[180],{"type":24,"value":177},{"type":18,"tag":33,"props":182,"children":183},{},[184],{"type":24,"value":185},"如果你有任何问题，欢迎联系SIG Lead，来自清华大学的陈键飞博士(主要研究为贝叶斯推理，深度生成模型和图模型等)",{"type":18,"tag":33,"props":187,"children":188},{},[189],{"type":24,"value":190},"2. 查看MDP代码：",{"type":18,"tag":33,"props":192,"children":193},{},[194],{"type":18,"tag":39,"props":195,"children":198},{"href":196,"rel":197},"https://gitee.com/mindspore/mindspore/tree/r0.7/mindspore/nn/probability",[43],[199],{"type":24,"value":196},{"type":18,"tag":33,"props":201,"children":202},{},[203],{"type":24,"value":204},"3. 在社区提交ISSUE或者PR，正式开始提出你的问题或者做出你的贡献",{"type":18,"tag":33,"props":206,"children":207},{},[208,210],{"type":24,"value":209},"4. 微信添加好友搜索：",{"type":18,"tag":211,"props":212,"children":213},"strong",{},[214],{"type":24,"value":215},"mindspore0328",{"type":18,"tag":33,"props":217,"children":218},{},[219],{"type":18,"tag":58,"props":220,"children":222},{"alt":7,"src":221},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2020/09/03/f96f78e5a37847d082a92aa4f21f606e.jpg",[],{"type":18,"tag":33,"props":224,"children":225},{},[226],{"type":24,"value":227},"MindSpore官方资料",{"type":18,"tag":33,"props":229,"children":230},{},[231,233],{"type":24,"value":232},"GitHub:",{"type":18,"tag":39,"props":234,"children":237},{"href":235,"rel":236},"https://github.com/mindspore-ai/mindspore",[43],[238],{"type":24,"value":235},{"type":18,"tag":33,"props":240,"children":241},{},[242,244],{"type":24,"value":243},"Gitee:",{"type":18,"tag":39,"props":245,"children":248},{"href":246,"rel":247},"https://gitee.com/mindspore/mindspore",[43],[249],{"type":24,"value":246},{"type":18,"tag":33,"props":251,"children":252},{},[253],{"type":24,"value":254},"官方QQ群: 871543426",{"type":18,"tag":33,"props":256,"children":257},{},[258],{"type":18,"tag":211,"props":259,"children":260},{},[261],{"type":24,"value":262},"扫描下方二维码关注↓",{"type":18,"tag":33,"props":264,"children":265},{},[266],{"type":18,"tag":58,"props":267,"children":269},{"alt":7,"src":268},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2020/09/03/ba5fb35b9bc74ee9bd38b96a94867022.jpg",[],{"title":7,"searchDepth":271,"depth":271,"links":272},4,[273],{"id":28,"depth":274,"text":31},2,"markdown","content:activities:zh:248.md","content","activities/zh/248.md","activities/zh/248","md",1776506018271]