Adaptive momentum-based optimization to train deep neural network for simulating the static stability of the composite structure

Chi, Zhifeng; Jiang, Zhiyong; Kamruzzaman, M. M.; Hafshejani, Behzad Aalipur; Safarpour, Mehran

doi:10.1007/s00366-021-01335-5

Cited by 9 publications

(2 citation statements)

References 126 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, convolutional neural networks (CNN) are often applied in computational problems [5][6][7]. The combination of techniques from computational mathematics and ANN has resulted in interesting contributions for both fields, since the theoretical results from classical approximation theory can be used to derive results on the approximation properties of ANN [8][9][10][11][12].…”

Section: Introductionmentioning

confidence: 99%

Locally refined quad meshing for linear elasticity problems based on convolutional neural networks

Chan

Scholz

Takacs

2022

Engineering with Computers

View full text Add to dashboard Cite

In this paper we propose a method to generate suitably refined finite element meshes using neural networks. As a model problem we consider a linear elasticity problem on a planar domain (possibly with holes) having a polygonal boundary. We impose boundary conditions by fixing the position of a part of the boundary and applying a force on another part of the boundary. The resulting displacement and distribution of stresses depend on the geometry of the domain and on the boundary conditions. When applying a standard Galerkin discretization using quadrilateral finite elements, one usually has to perform adaptive refinement to properly resolve maxima of the stress distribution. Such an adaptive scheme requires a local error estimator and a corresponding local refinement strategy. The overall costs of such a strategy are high. We propose to reduce the costs of obtaining a suitable discretization by training a neural network whose evaluation replaces this adaptive refinement procedure. We set up a single network for a large class of possible domains and boundary conditions and not on a single domain of interest. The computational domain and boundary conditions are interpreted as images, which are suitable inputs for convolution neural networks. In our approach we use the U-net architecture and we devise training strategies by dividing the possible inputs into different categories based on their overall geometric complexity. Thus, we compare different training strategies based on varying geometric complexity. One of the advantages of the proposed approach is the interpretation of input and output as images, which do not depend on the underlying discretization scheme. Another is the generalizability and geometric flexibility. The network can be applied to previously unseen geometries, even with different topology and level of detail. Thus, training can easily be extended to other classes of geometries.

show abstract

Section: Introductionmentioning

confidence: 99%

Locally refined quad meshing for linear elasticity problems based on convolutional neural networks

Chan

Scholz

Takacs

2022

Engineering with Computers

View full text Add to dashboard Cite

show abstract

“…Only the available advanced technologies can ensure people's access to COVID-19-related information. According to many scholars [2][3][4][5][6][7][8], COVID-19-related information should be made available to people through the use of advanced digital technologies, comprising big data, artificial intelligence (AI), Deep Learning (DL), Internet of Things (IoT), 5G and 6G. These technologies are necessary to solve the main clinical problems as well as removing barriers to people's access to information.…”

Section: Introductionmentioning

confidence: 99%

Technology‐driven 5G enabled e‐healthcare system during COVID‐19 pandemic

et al. 2021

View full text Add to dashboard Cite

Technology-driven control measures could be an important tool to control the COVID-19 pandemic crisis. This study evaluates the potentiality of emerging technologies such as 5G and 6G communication, Deep Learning (DL), big data, Internet of Things (IoT) etc. for controlling the COVID-19 transmission and ensuring health safety. The healthcare sector is able to provide a unified, rapid, and incessant service to people by applying modern wireless connectivity tools like 5G or 6G during the COVID-19 pandemic. This study has identified eight key areas of applications for the COVID-19 management like infection detection; travel history analysis; identification of infection symptoms; early detection; transmission identification; access to information in lockdown; movement of people; and development of medical treatments and vaccines. Data have been collected from the respondents living in Sakaka city, KSA during pandemic. This study reveals that most people receive information from social networking sites, health professionals, and television without facing any challenges. The analysis shows that, during the COVID-19 pandemic, about 42% of respondents felt tense always or most of the time in a day. Only 28.6% of respondents felt tense sometimes, whereas the remainder (about 30%) did not feel tense in relation to the COVID-19 crisis. Satisfaction with COVID-19-related information is also positively correlated with COVID-19-related information literacy (r = 0.53, p < 0.01) that is also positively correlated with depression or emotion, anxiety, and stress (r = -0.15, p < 0.05). The long-term pandemic is creating several psychological symptoms including anxiety, stress, and depression, irrespective of age. INTRODUCTIONThe entire world is now facing a terrible and fatal COVID-19 pandemic caused by SARS-CoV-2. From its discovery in Wuhan, China, in December 2019, COVID-19, as a rapidly transmitted infectious coronavirus, has created panic worldwide. As on May 5, 2021, the total number of identified cases is 154,988,353 and death is 3,241,164 [1]. At the current time, the number of cases is increasing rapidly. Traditional technologies in the medical field have failed to control this pandemic. Schol-This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Performance improvement for metro passenger flow forecast using spatio-temporal deep neural network

Mulerikkal

Thandassery

Rejathalal

et al. 2021

Neural Comput & Applic

View full text Add to dashboard Cite

Adaptive momentum-based optimization to train deep neural network for simulating the static stability of the composite structure

Cited by 9 publications

References 126 publications

Locally refined quad meshing for linear elasticity problems based on convolutional neural networks

Locally refined quad meshing for linear elasticity problems based on convolutional neural networks

Technology‐driven 5G enabled e‐healthcare system during COVID‐19 pandemic

Performance improvement for metro passenger flow forecast using spatio-temporal deep neural network

Contact Info

Product

Resources

About