Browsing by Author "Li, Xue"

Now showing 1 - 4 of 4

Exploiting Centrality Information with Graph Convolutions for Network Representation Learning
(2019-01-01)
Chen, Hongxu
;
Yin, Hongzhi
;
Chen, Tong
;
Quoc Viet Hung Nguyen
;
Peng, Wen-Chih
Network embedding has been proven effective to learn low-dimensional vector representations for network vertices, and recently received a tremendous amount of research attention. However, most of existing methods for network embedding merely focus on preserving the first and second order proximities between nodes, and the important properties of node centrality are neglected. Various centrality measures such as Degree, Closeness, Betweenness, Eigenvector and PageRank centralities have been designed to measure the importance of individual nodes. In this paper, we focus on a novel yet unsolved problem that aims to learn low-dimensional continuous nodes representations that not only preserve the network structure, but also keep the centrality information. We propose a generalizable model, namely GraphCSC, that utilizes both linkage information and centrality information to learn low-dimensional vector representations for network vertices. The learned embeddings by GraphCSC are able to preserve different centrality information of nodes. In addition, we further propose GraphCSC-M, a more comprehensive model that can preserve different centrality information simultaneously through learning multiple centrality specific embeddings, and a novel attentive multi-view learning approach is developed to compress multiple embeddings of one node into a compact vector representation. Extensive experiments have been conducted to demonstrate that our model is able to preserve different centrality information of nodes, and achieves better performance on several benchmark tasks compared with recent state-of-the-art network embedding methods.
Joint Dynamic Spectrum Access and Multi-Relay Selection: A Matching-Theory-based Approach
(2017-01-01)
Xu, Wenjun
;
Li, Xue
;
Wang, Yixin
;
Lee, Chia-Han
;
Feng, Zhiyong
In this paper, the problem of joint dynamic spectrum access and multi-relay selection is investigated in relay-enabled cooperative communication systems to maximize the system sum-capacity. Since the considered problem is a mixed integer nonlinear program, which is generally intractable to find the optimal solution, two matching theory-based suboptimal algorithms are proposed to reduce the computational complexity for two different cases. For the case that each source node can only be assisted by one relay, a cyclic three-sided matching algorithm is firstly proposed to attain the stable matching results for the selection of the source node and the relay with the spectrum band used. Then, for the case that each source node can be assisted by more than one relay, a two-step matching algorithm is proposed to perform joint dynamic spectrum access and multi-relay selection. Simulation results show that the proposed algorithms, with much lower complexity compared to the optimal exhaustive search, can achieve the near-optimal performance with a gap to the optimum being less than 5%.
Joint Sensing Duration Adaptation, User Matching, and Power Allocation for Cognitive OFDM-NOMA Systems
(2018-02-01)
Xu, Wenjun
;
Li, Xue
;
Lee, Chia-Han
;
Pan, Miao
;
Feng, Zhiyong
In this paper, the non-orthogonal multiple access (NOMA) technology is integrated into cognitive orthogonal frequency-division multiplexing (OFDM) systems, called cognitive OFDM-NOMA, to boost the system capacity. First, a capacity maximization problem is considered in half-duplex cognitive OFDM-NOMA systems with two accessible users on each subcarrier. Due to the intractability of the considered problem, we decompose it into three subproblems, i.e., the optimization of, respectively, sensing duration, user scheduling, and power allocation. By investigating and exploiting the characteristics of each subproblem, the optimal sensing duration adaptation, a matching-theory-based user scheduling, and the optimal power allocation are proposed correspondingly. An alternate iteration framework is further proposed to jointly optimize these three subproblems, with its convergence proved. Moreover, based on the non-cooperative game theory, a generalized power allocation algorithm is proposed and then used in the framework to accommodate half-duplex cognitive OFDM-NOMA systems with multiple users on each subcarrier. Finally, the proposed framework is extended to solve the capacity maximization problem in full-duplex cognitive OFDM-NOMA systems. Simulation results validate the superior performance of the proposed algorithms. For example, for the case of two accessible users, the proposed framework approaches the optimal solution with less than 1% capacity loss and 120 times lower complexity compared with exhaustive search.
Mining Health Examination Records-A Graph-Based Approach
(2016-09-01)
Chen, Ling
;
Li, Xue
;
Sheng, Quan Z.
;
Peng, Wen-Chih
;
Bennett, John
General health examination is an integral part of healthcare in many countries. Identifying the participants at risk is important for early warning and preventive intervention. The fundamental challenge of learning a classification model for risk prediction lies in the unlabeled data that constitutes the majority of the collected dataset. Particularly, the unlabeled data describes the participants in health examinations whose health conditions can vary greatly from healthy to very-ill. There is no ground truth for differentiating their states of health. In this paper, we propose a graph-based, semi-supervised learning algorithm called SHG-Health (Semi-supervised Heterogeneous Graph on Health) for risk predictions to classify a progressively developing situation with the majority of the data unlabeled. An efficient iterative algorithm is designed and the proof of convergence is given. Extensive experiments based on both real health examination datasets and synthetic datasets are performed to show the effectiveness and efficiency of our method.