Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Thesis_XiangLi.pdf (1.72 MB)

ETD Abstract Container

Abstract Header

Development of Deep Learning Models for Attributed Graphs

Li, Xiang
http://rave.ohiolink.edu/etdc/view?acc_num=osu1683467941548213

Abstract Details

2023, Doctor of Philosophy, Ohio State University, Computer Science and Engineering.
Attributed graphs, i.e. graphs with attributes associated with nodes, are popular data representations used to capture interactions between entities. In recent years, interest has been growing in developing data mining techniques on attributed graphs for different learning tasks such as node classification and clustering, link prediction and graph classification. Graph Neural Networks (GNN) are emerging as the state-of-the-art graph mining models. Within GNN, Graph Convolutional Networks (GCN) have been used with great success in various domains such as recommendation systems, social network analysis, AI-powered drug discovery etc. However, there are multiple challenges from different aspects in using GCN based approaches: 1) high time cost resulting from the frequent loading of data to GPUs during training; 2) limited learning ability resulting from over-smoothing issues inherent to GCN; 3) difficulties in scaling to large-scale graphs due to limited GPU memory; 4) restrictions on access to centralized training datasets when data sharing is prohibited due to privacy or commercial restrictions. This dissertation focuses on developing multiple efficient GNN based approaches for attributed graphs learning aimed at solving the aforementioned challenges. First, we present a general framework, in which multiple GCN methods (GraphSage, clusterGCN, GraphSaint) can be accelerated by reducing the frequency of data transfer to GPUs without noticeable degradation on learning ability. Second, to relieve both the over-smoothing and scalability issues of GCN, we describe our scalable deep clustering framework, Random-walk based Scalable Learning (RwSL) focusing on the node clustering task. Previous work like GCN based DGI, SDCN, DMoN or graph filtering based AGC, SSGC, AGE do not scale to large-scale graphs due to their use of non-scalable graph convolution operations. In contrast, RwSL can scale to graphs of arbitrary size by employing a parallelizable random-walk based algorithm as a graph filter and a subsequent DNN based module for clustering-oriented mini-batch training. Third, we present a scalable GNN method, Deep Metric Learning with Multi-class Tuplet Loss (DMT) where a resulting embedding can support multiple downstream learning tasks (node classification/clustering, link prediction) with superior performance and training efficiency. This further extends the application scope of scalable GNN approaches and provides improved learning capability. Finally, we consider scenarios, where graph-structure data are stored in a decentralized manner and transfer of the raw data, is prohibited. We propose a framework, Federated Contrastive Learning of Graph-level representation (FCLG). Our goal is to train a global GNN based model that exploits data from decentralized clients. In order to address the Non-IID (independent and identical distributed) issues inherent to Federated Learning (FL), we employ a two-level contrastive learning mechanism. On each client, we use the contrast between multiple augmented views of input graphs to encode robust characteristics within different graphs into a local GNN model. We then contrast the global model learned on the server with the local models learned on clients to improve the generalization performance of the global model.
Rajiv Ramnath (Advisor)
Gagan Agrawal (Committee Member)
Srinivasan Parthasarathy (Committee Member)
Radu Teodorescu (Committee Member)
Ruoming Jin (Committee Member)
149 p.
Computer Engineering; Information Technology
attributed graphs, deep learning, graph neural networks, scalability, federated learning

Recommended Citations

Citations

  • Li, X. (2023). Development of Deep Learning Models for Attributed Graphs [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1683467941548213

    APA Style (7th edition)

  • Li, Xiang. Development of Deep Learning Models for Attributed Graphs. 2023. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1683467941548213.

    MLA Style (8th edition)

  • Li, Xiang. "Development of Deep Learning Models for Attributed Graphs." Doctoral dissertation, Ohio State University, 2023. http://rave.ohiolink.edu/etdc/view?acc_num=osu1683467941548213

    Chicago Manual of Style (17th edition)

osu1683467941548213
207
© 2023, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.