Toward graph classification on structure property using adaptive motif based on graph convolutional network

2021 ◽  
Author(s):  
Xingquan Li ◽  
Hongxi Wu
Keyword(s):  
Graph Classification ◽  
Structure Property ◽  
Convolutional Network

Graph Convolutional Networks for Exercise Motion Classification

2021 ◽  
Vol 65 (1) ◽  
pp. 685-689
Author(s):  
Parian Haghighat ◽  
Aden Prince ◽  
Heejin Jeong
Keyword(s):  
Nearest Neighbor ◽  
Current Model ◽  
Model Performance ◽  
Classification Performance ◽  
Great Promise ◽  
Graph Classification ◽  
Convolutional Network ◽  
Motion Data ◽  
Convolutional Networks ◽  
Personal Fitness

The growth in self-fitness mobile applications has encouraged people to turn to personal fitness, which entails integrating self-tracking applications with exercise motion data to reduce fatigue and mitigate the risk of injury. The advancements in computer vision and motion capture technologies hold great promise to improve exercise classification performance. This study investigates a supervised deep learning model performance, Graph Convolutional Network (GCN) to classify three workouts using the Azure Kinect device’s motion data. The model defines the skeleton as a graph and combines GCN layers, a readout layer, and multi-layer perceptrons to build an end-to-end framework for graph classification. The model achieves an accuracy of 95.86% in classifying 19,442 frames. The current model exchanges feature information between each joint and its 1-nearest neighbor, which impact fades in graph-level classification. Therefore, a future study on improved feature utilization can enhance the model performance in classifying inter-user exercise variation.

scholarly journals Motif-Matching Based Subgraph-Level Attentional Convolutional Network for Graph Classification

2020 ◽  
Vol 34 (04) ◽  
pp. 5387-5394
Author(s):  
Hao Peng ◽  
Jianxin Li ◽  
Qiran Gong ◽  
Yuanxin Ning ◽  
Senzhang Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Social Network ◽  
Spatial Information ◽  
Structural Information ◽  
Feature Learning ◽  
Drug Analysis ◽  
Classification Performance ◽  
Learning Approaches ◽  
Graph Classification ◽  
Convolutional Network

Graph classification is critically important to many real-world applications that are associated with graph data such as chemical drug analysis and social network mining. Traditional methods usually require feature engineering to extract the graph features that can help discriminate the graphs of different classes. Although recently deep learning based graph embedding approaches are proposed to automatically learn graph features, they mostly use a few vertex arrangements extracted from the graph for feature learning, which may lose some structural information. In this work, we present a novel motif-based attentional graph convolution neural network for graph classification, which can learn more discriminative and richer graph features. Specifically, a motif-matching guided subgraph normalization method is developed to better preserve the spatial information. A novel subgraph-level self-attention network is also proposed to capture the different impacts or weights of different subgraphs. Experimental results on both bioinformatics and social network datasets show that the proposed models significantly improve graph classification performance over both traditional graph kernel methods and recent deep learning approaches.

Aggregation Enhanced Graph Convolutional Network for Graph Classification

2021 ◽  
Author(s):  
Guixian Zhang ◽  
Boyan Chen ◽  
Lijuan Wu ◽  
Kui Zhang ◽  
Shichao Zhang
Keyword(s):  
Graph Classification ◽  
Convolutional Network

Structure-property relations of liquid crystalline polymers

1987 ◽  
Vol 45 ◽  
pp. 460-463
Author(s):  
Linda C. Sawyer
Keyword(s):  
Solid State ◽  
Liquid Crystalline ◽  
Structural Model ◽  
Crystalline Polymer ◽  
Liquid Crystalline Polymers ◽  
Mechanical Anisotropy ◽  
Structure Property ◽  
Preparation Methods ◽  
Crystalline Polymers ◽  
Extended Chain

Recent liquid crystalline polymer (LCP) research has sought to define structure-property relationships of these complex new materials. The two major types of LCPs, thermotropic and lyotropic LCPs, both exhibit effects of process history on the microstructure frozen into the solid state. The high mechanical anisotropy of the molecules favors formation of complex structures. Microscopy has been used to develop an understanding of these microstructures and to describe them in a fundamental structural model. Preparation methods used include microtomy, etching, fracture and sonication for study by optical and electron microscopy techniques, which have been described for polymers. The model accounts for the macrostructures and microstructures observed in highly oriented fibers and films.Rod-like liquid crystalline polymers produce oriented materials because they have extended chain structures in the solid state. These polymers have found application as high modulus fibers and films with unique properties due to the formation of ordered solutions (lyotropic) or melts (thermotropic) which transform easily into highly oriented, extended chain structures in the solid state.

Structure-property relationships of PE/PP sandwiched films

1987 ◽  
Vol 45 ◽  
pp. 454-455
Author(s):  
J. Petermann ◽  
G. Broza ◽  
U. Rieck ◽  
A. Jaballah ◽  
A. Kawaguchi
Keyword(s):  
Mechanical Tests ◽  
Polymer Materials ◽  
Ionic Crystals ◽  
Structure Property ◽  
Polymer Systems ◽  
Structure Property Relationships ◽  
Oriented Films ◽  
Materials Used ◽  
Polymer Laminates ◽  
Heated Glass

Oriented overgrowth of polymer materials onto ionic crystals is well known and recently it was demonstrated that this epitaxial crystallisation can also occur in polymer/polymer systems, under certain conditions. The morphologies and the resulting physical properties of such systems will be presented, especially the influence of epitaxial interfaces on the adhesion of polymer laminates and the mechanical properties of epitaxially crystallized sandwiched layers.Materials used were polyethylene, PE, Lupolen 6021 DX (HDPE) and 1810 D (LDPE) from BASF AG; polypropylene, PP, (PPN) provided by Höchst AG and polybutene-1, PB-1, Vestolen BT from Chemische Werke Hüls. Thin oriented films were prepared according to the method of Petermann and Gohil, by winding up two different polymer films from two separately heated glass-plates simultaneously with the help of a motor driven cylinder. One double layer was used for TEM investigations, while about 1000 sandwiched layers were taken for mechanical tests.

Probing vacancies and structural distortions at individual defects in ceramics using EELS

1996 ◽  
Vol 54 ◽  
pp. 678-679
Author(s):  
D. J. Wallis ◽  
N. D. Browning
Keyword(s):  
Structural Information ◽  
Core Loss ◽  
Nearest Neighbors ◽  
Ceramic Materials ◽  
Scanning Transmission Electron Microscope ◽  
Real Space ◽  
Atomic Scale ◽  
Structure Property ◽  
Scanning Transmission ◽  
Key Issues

In electron energy loss spectroscopy (EELS), the near-edge region of a core-loss edge contains information on high-order atomic correlations. These correlations give details of the 3-D atomic structure which can be elucidated using multiple-scattering (MS) theory. MS calculations use real space clusters making them ideal for use in low-symmetry systems such as defects and interfaces. When coupled with the atomic spatial resolution capabilities of the scanning transmission electron microscope (STEM), there therefore exists the ability to obtain 3-D structural information from individual atomic scale structures. For ceramic materials where the structure-property relationships are dominated by defects and interfaces, this methodology can provide unique information on key issues such as like-ion repulsion and the presence of vacancies, impurities and structural distortion.An example of the use of MS-theory is shown in fig 1, where an experimental oxygen K-edge from SrTiO3 is compared to full MS-calculations for successive shells (a shell consists of neighboring atoms, so that 1 shell includes only nearest neighbors, 2 shells includes first and second-nearest neighbors, and so on).

Approaches for TEM imaging of cavitation in toughened nylon

1989 ◽  
Vol 47 ◽  
pp. 352-353
Author(s):  
Barbara A. Wood
Keyword(s):  
Morphological Characteristics ◽  
Structure Property ◽  
Polymer Systems ◽  
Selective Staining ◽  
Structure Property Relationships ◽  
Rubber Toughened ◽  
Shear Yield ◽  
Controversial Topic ◽  
Selection Of ◽  
Diamond Knife

A controversial topic in the study of structure-property relationships of toughened polymer systems is the internal cavitation of toughener particles resulting from damage on impact or tensile deformation.Detailed observations of the influence of morphological characteristics such as particle size distribution on deformation mechanisms such as shear yield and cavitation could provide valuable guidance for selection of processing conditions, but TEM observation of damaged zones presents some experimental difficulties.Previously published TEM images of impact fractured toughened nylon show holes but contrast between matrix and toughener is lacking; other systems investigated have clearly shown cavitated impact modifier particles. In rubber toughened nylon, the physical characteristics of cavitated material differ from undamaged material to the extent that sectioning of heavily damaged regions by cryoultramicrotomy with a diamond knife results in sections of greater than optimum thickness (Figure 1). The detailed morphology is obscured despite selective staining of the rubber phase using the ruthenium trichloride route to ruthenium tetroxide.

Structure–property studies of self-assembling [n.n]paracyclophanes

2020 ◽  
Author(s):  
Will R Henderson ◽  
Danielle E. Fagnani ◽  
Yu Zhu ◽  
Guancen Liu ◽  
Ronald K. Castellano
Keyword(s):  
Structure Property ◽  
Self Assembling
Sign in / Sign up

Export Citation Format

Share Document