24小时热门版块排行榜

>论坛更新日志 (3002)
>虫友互识 (235)
>导师招生 (183)
>文献求助 (180)
>论文投稿 (69)
>硕博家园 (55)
>休闲灌水 (55)
>公派出国 (48)
>考博 (42)
>招聘信息布告栏 (34)
>催化 (29)
>论文道贺祈福 (28)
>教师之家 (26)
>找工作 (26)
>考研 (23)
>博后之家 (22)

返回列表

【奖励】本帖被评价7次，作者huishouzhong增加金币 5.2 个

huishouzhong

铁杆木虫 (小有名气)

应助: 1 (幼儿园)
金币: 7648.6
帖子: 131
在线: 135.8小时
虫号: 2535754

[资源] Machine Learning in Modeling and Simulation - Methods Applications

分享计算力学大牛klaus-jürgen bathe和timon rabczuk在2023年新编的书：
machine learning in modeling and simulation methods and applications

两位大牛的引用次数：

Machine Learning in Modeling and Simulation - Methods Applications-1

---------

Machine Learning in Modeling and Simulation - Methods Applications-2

前言节选：
our objective in this book is to present ml techniques for computer-aided engineering with a focus on the fundamental theoretical ingredients and the exciting use of ml in the next generation of computational methods for modeling and simulation.

an impressive example of ml in engineering is the emergence and great potential of using digital twins for design and monitoring of structures, with the word “structures” interpreted to include not only traditional structures, like buildings, dams, and bridges, but also biological, offshore, electromagnetic, turbines, nuclear, and many other structures. we foresee here major and widely spread applications.

the book consists of 12 chapters focusing on machine learning in modeling and simulation, starting with an extensive overview of concepts and applications in chap. 1. the following two chapters provide a historical and theoretical overview—including implementation details—of two very popular machine learning approaches: artificial neural networks and gaussian processes. the remainder of the book focuses on the use of ml techniques to solve specific problems in engineering, physics or materials science starting with data-driven model discovery followed by physics-informed neural networks that may become a powerful alternative to classical discretization methods such as finite elements. namely, the networks allow the solution of partial differential equations while also incorporating experimentaldata. physics-informeddeepneuraloperatorscanbeseenasanimprovement and have the potential of solving not only one specific problem but any problem in a large class of problems. the next chapters focus on important topics regarding digital twins, reduced order modeling, regression models and the valuable use of ml in topology optimization methodologies. the last two chapters of the book focus on the design of new materials. the first approach is data-driven while the second approach takes advantage of interatomic potentials in the context of hierarchical multiscale modeling. these ml approaches have the potential to significantly widen the possibilities and accelerate the design of new materials.

the work on this book has been very exciting, and we greatly thank all authors and co-authors of the book chapters. their valuable contributions, dedicated work, and great cooperation made it possible to complete this book in a timely manner.

weimar, germany timon rabczuk
cambridge, usa klaus-jürgen bathe
december 2022

回复此楼

» 本帖附件资源列表

欢迎监督和反馈：小木虫仅提供交流平台，不对该内容负责。
本内容由用户自主发布，如果其内容涉及到知识产权问题，其责任在于用户本人，如对版权有异议，请联系邮箱：xiaomuchong@tal.com
附件 1 : 2023_Machine_Learning_in_Modeling_and_Simulation.pdf

2024-04-15 15:53:25, 12.76 M