[{"data":1,"prerenderedAt":57},["ShallowReactive",2],{"content-query-cxOzhC4hUH":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":51,"_id":52,"_source":53,"_file":54,"_stem":55,"_extension":56},"/activities/zh/3707","zh",false,"","直播 | 带你看如何在任意几何下高效求解物理方程","4月25日19:00，昇思MindSpore Flow&Earth SIG技术分享会启动！本期我们特邀 中国人民大学 曾博成老师 为大家讲解《用于时空PDE系统的物理编码消息传递图神经网络》。","2025-04-23","https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2025/04/25/dad777beae5745a3b92116bbf663c492.png","activities",1,{"type":15,"children":16,"toc":48},"root",[17,25,30,35,40],{"type":18,"tag":19,"props":20,"children":22},"element","h1",{"id":21},"直播-带你看如何在任意几何下高效求解物理方程",[23],{"type":24,"value":8},"text",{"type":18,"tag":26,"props":27,"children":28},"p",{},[29],{"type":24,"value":9},{"type":18,"tag":26,"props":31,"children":32},{},[33],{"type":24,"value":34},"本期内容中的研究成果基于2025ICLR SpotLight的稿件。目前现有的大多数神经模型依赖于丰富的训练数据，外推和泛化能力有限，并且在复杂的条件下难以产生精确或可靠的物理预测。为此，我们提出了一种新的图学习方法，即物理编码的消息传递图网络（PhyMPGN），给定小的训练数据集，在不规则网格上对时空PDE系统建模。我们将带大家了解物理编码的消息传递图网络，看它如何利用较小数据，更准确地模拟复杂系统的动态变化。",{"type":18,"tag":26,"props":36,"children":37},{},[38],{"type":24,"value":39},"本期技术分享会在腾讯会议和B站直播间【MindSpore官方】直播，欢迎各位参与。",{"type":18,"tag":26,"props":41,"children":42},{},[43],{"type":18,"tag":44,"props":45,"children":47},"img",{"alt":7,"src":46},"https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2025/04/25/d27f981b08794bbcbd1dfc0dd59474c2.png",[],{"title":7,"searchDepth":49,"depth":49,"links":50},4,[],"markdown","content:activities:zh:3707.md","content","activities/zh/3707.md","activities/zh/3707","md",1776506030095]