Название: Concepts and Techniques of Graph Neural NetworksАвтор: Vinod Kumar, Dharmendra Singh Rajput
Издательство: IGI Global
Год: 2023
Страниц: 267
Язык: английский
Формат: pdf (true), epub
Размер: 33.6 MB
Graph Neural Networks, also known as GNNs, have seen a meteoric rise in popularity over the past few years due to its capacity to analyse data that is shown in the form of graphs. GNNs have been put to use in a broad variety of industries, including social network research, the search for new drugs, recommender systems, and traffic prediction, to mention just a few examples. GNNs are becoming increasingly popular, which has resulted in an increased interest in the issue among scholars and practitioners who are interested in better comprehending the fundamental ideas and procedures that underpin GNNs. The book “Concepts and Techniques of Graph Neural Networks” is a reference to the concepts and procedures that are utilised in Graph Neural Networks (GNNs). GNNs have developed as an effective method for modelling complicated structured data, such as social networks, protein structures, and traffic patterns, among other applications. Applications have been identified for them in a broad variety of domains, such as drug discovery, computer vision, natural language processing, and recommender systems. GNNs are especially helpful for solving situations in which the data is modelled as a graph, in which the nodes of the graph represent entities and the edges reflect the relationships between those entities.
Because GNNs are able to perform reasoning and inference by utilising the graph structure, they are ideally suited for solving problems in which the relationships between the entities are of primary significance. Today, research into GNNs is a field that is expanding at a rapid rate, and new ideas and methods are being developed on a consistent basis. This is an interdisciplinary area that draws on concepts from a variety of different fields, including computer science, mathematics, and statistics, among others. The relevance of GNNs is anticipated to expand as the amount of structured data continues to grow, which makes the ideas and approaches for GNNs an essential topic of research in the world as it exists today. GNNs are able to accomplish tasks such as node classification, link prediction, and graph classification by employing a mix of node and edge attributes. This allows the information to be propagated throughout the graph as the network is traversed. The fundamental concept underlying GNNs is that they should take advantage of the local connection patterns of the graph to extract relational information and then use this knowledge to the task of making predictions. Moreover, GNNs are able to handle large-scale, complicated graphs that contain millions or even billions of nodes and edges, which is one of the primary advantages of using these types of networks.
GNNs also have the potential to overcome some of the difficulties that have arisen as a result of the increasing availability of graph-structured data in the world as it exists today. Large-scale graphs are frequently used to illustrate the intricacies of the interactions that exist between the many entities involved in a variety of areas, such as social networks, online markets, and transportation systems, to name just a few examples. Researchers and practitioners may construct models for analysing these graphs that are more accurate, scalable, and interpretable by utilising the capabilities of GNNs. These models will allow for the extraction of relevant insights. This book on “Concepts and Techniques of Graph Neural Networks” aims to provide a comprehensive resource for anyone interested in understanding GNNs. The book covers both fundamental concepts and recent advances in the field, making it accessible to both beginners and experienced practitioners.
Several prominent researchers and practitioners in the fields of AI, ML, and graph mining contributed to the book's 13 chapters. The book is organized in a logical progression from basic concepts to the corresponding technological solutions. The book's material is structured as follows:
Chapter 1: This talks about basic and very important key terms related to graphs, such as graph and its different types, nodes, edges, degree of a graph, adjacency matrix, modelling of graphs, graph embedding, etc. Overall, this chapter is about the basics of graph theory and graphs for graph neural networks.
Chapter 2: This chapter talks about the Graph Neural Network and its design, the Graph Embedding Process, and the benefits of Graph Neural Networks over Graph Convolutional Networks. It also talks about the problems and new changes in GNN, as well as how it can be used in real life.
Chapter 3: This introduces Graph Neural Networks (GNNs) by outlining graph deep learning, GNN theories, general principles, and GNN types. GNN apps were described at the end, and GNN applications were also looked at.
Chapter 4: This gives a review of GNNs and how they can be used to classify graphs. It also talks about some of the problems and limits of GNNs and how they can be used in different ways to classify graphs. In addition, it gives an overview of standard datasets that can be used for empirical study.
Chapter 5: This chapter gives an overview of recent improvements in adversarial attacks on GNNs, including attack methods, evaluation measures, and how they affect model performance.
Chapter 6: This chapter talks about the basic ideas behind graph attention networks (GAT), including how they are put together. It also goes into more detail about how normal GCN and GAT are different. In addition, it shows how GAT can be used in important ways in the real world.
Chapter 7: This is about Graph Convolutional Neural Networks for Link Prediction in Social Networks. The GCNN design for link detection is covered. This shows how the social network data for GCNN was prepared before it was used.
Chapter 8: The writers want to look at how graph neural networks (GNNs) can be used to solve common computer vision problems, such as visual saliency, salient object recognition, and co-saliency. In this chapter, we look at how graph neural networks have been used to solve a number of problems with visual saliency. Also looked at are the different study methods that used GNN to find salience and co-salience between items.
Chapter 9: This chapter talks about the useful study gaps in the area of forecasting models that use Graph Neural Networks (GNN) and Machine Learning. This includes trial data analysis with different examples. Two case studies are also included to help you understand how GNN applications work.
Chapter 10: This chapter shows how Graph Neural Networks can be used in m-health to track diseases in people. It also says that health information technology (HIT) that is digital is the way of the future.
Chapter 11: This chapter looks at how Graph Neural Network can be used to predict how well kids will do in school. To help students do better, it is suggested that a systematic literature review be done on how to predict student success by using data processing methods.
Chapter 12: This chapter shows how artificial intelligence and inspired algorithms can be used to find fake news by using graph neural networks.
Chapter 13: This chapter is an in-depth look at face recognition using feature extraction and the fusion face method. The main goal of this chapter is to use image fusion to solve the problem of face recognition. It also gives a thorough look at the different image fusion methods for face recognition.
Covering key topics such as graph data, social networks, Deep Learning, and graph clustering, this premier reference source is ideal for industry professionals, researchers, scholars, academicians, practitioners, instructors, and students.
Скачать Concepts and Techniques of Graph Neural Networks