Although unsupervised multiplex graph representation learning (UMGRL) has been a hot research topic, existing UMGRL methods still has limitations to be addressed. For example, previous works either preserve structural information by ignoring the impact of heterophily in the graph structure or only focus on node-level consistency by ignoring class-level consistency. To address these issues, in this paper, we propose a new UMGRL method to explore both homophily and consistency in the multiplex graph. Specifically, we propose to restructure the multi-order relationships of every graph between every node and its multi-order neighbors to improve the homophily and reduce the impact of the heterophily in the graph structure. We also design a contrastive loss based on a self-expression matrix of the node representation to achieve node-level and class-level consistency. Furthermore, we theoretically prove our method to achieve class-level consistency. Extensive experimental results on real datasets verify the effectiveness of the proposed method with respect to node classification tasks, compared to SOTA methods.

We revisit a classic problem of inferring the routing tree for a given source in a packet-switched network from end-to-end measurements, with two critical differences from existing solutions: (i) instead of exclusively relying on active measurements obtained by probing, we strive to maximally utilize passive measurements obtained from data packets; (ii) instead of inferring a logical topology that omits degree-2 nodes, we want to recover the physical topology containing all the nodes. Our main idea is to utilize the detailed queueing dynamics inside the network to estimate a certain parameter (residual capacity) of each queue, and then use the estimated parameters as fingerprints to detect the queues shared across paths and thus infer the topology. To this end, we develop a Laplace-transform-based estimator to estimate the parameters of a tandem of queues from end-to-end delays, and efficient algorithms to infer the topology by identifying the parameters associated with the same queue. To improve the accuracy, we further develop a hybrid algorithm that uses the information from active measurements to identify (generalized) siblings and the information from passive measurements to detect shared queues on the paths from the source to each pair of identified siblings. Our inferred topology is guaranteed to converge to the ground-truth topology as the number of measurements increases, up to a permutation of the queues traversed by the same set of paths. Our evaluations in both queueing-theoretic and packet-level simulations show that the proposed solutions, particularly the hybrid algorithm, significantly improve the accuracy over the state of the art.

The emerging machine learning paradigm of decentralized federated learning (DFL) has the promise of greatly boosting the deployment of artificial intelligence (AI) by directly learning across distributed agents without centralized coordination. Despite significant efforts on improving the communication efficiency of DFL, most existing solutions were based on the simplistic assumption that neighboring agents are physically adjacent in the underlying communication network, which fails to correctly capture the communication cost when learning over a general bandwidth-limited network, as encountered in many edge networks. In this work, we address this gap by leveraging recent advances in network tomography to jointly design the communication demands and the communication schedule for overlay-based DFL in bandwidth-limited networks without requiring explicit cooperation from the underlying network. By carefully analyzing the structure of our problem, we decompose it into a series of optimization problems that can each be solved efficiently, to collectively minimize the total training time. Extensive data-driven simulations show that our solution can significantly accelerate DFL in comparison with state-of-the-art designs.

Overlay network is a non-intrusive mechanism to enhance the existing network infrastructure by building a logical distributed system on top of a physical underlay. A major difficulty in operating overlay networks is the lack of cooperation from the underlay, which is usually under a different network administration. In particular, the lack of knowledge about the underlay topology and link capacities makes the design of efficient overlay routing extremely difficult. In contrast to existing solutions for overlay routing based on simplistic assumptions such as known underlay topology or disjoint routing paths through the underlay, we aim at systematically optimizing overlay routing without causing congestion, by extracting necessary information about the underlay from measurements taken at overlay nodes. To this end, we (i) identify the minimum information for congestion-free overlay routing, and (ii) develop polynomial-complexity algorithms to infer this information with guaranteed accuracy. Our evaluations in NS3 based on real network topologies demonstrate notable performance advantage of the proposed solution over existing solutions.

This paper focuses on learning the representing dictionaries for sparse channel estimation in frequency-division duplexing (FDD) massive multiple-input multiple-output (MIMO) systems. To overcome the energy leakage problem in traditional sparse channel estimation, we propose a geographical dictionary- based spatial channel model to efficiently represent the cell-specific geographical characteristics. Based on that, the properties, especially the block structure, of the expected dictionaries are analyzed, and we design a data-driven joint block-structured dictionary learning algorithm (JBSDL) to obtain the expected representing dictionaries. The simulation environment is generated according to 3GPP standard, and we systematically study the properties of the learned dictionaries, which reveals the physical meaning of the dictionary learning results in massive MIMO systems. The proposed method demonstrates superior downlink channel estimation performance through the simulations.

Inspired by the phenomenon that the received signals naturally form clusters, we propose a novel machine learning framework to design multi-input multi-output (MIMO) communication systems. In the proposed framework, the MIMO detection problem is converted into a clustering problem, and known labels are transmitted to assist the receiver for labeling the clusters. A modulation-constrained Gaussian mixture model (MC-GMM) and the associated optimization algorithm are developed to reduce the number of parameters to be learnt in the clustering algorithm. Furthermore, we propose a method called label reconstruction to minimize the overhead of label transmission, and the design of the optimal labels is studied. Simulation results are presented to verify the effectiveness of the proposed label-assisted clustering (LAC) receiver in approaching the optimal maximum likelihood detection (MLD) with perfectly known channel knowledge for typical MIMO systems.

Blind modulation classification is a fundamental step before signal detection for cognitive radio networks where the users may not have the complete knowledge of the modulation scheme due to the flexibility of operating dynamically in multiple frequency bands. In this paper, a modulation-constrained (MC) clustering classifier is proposed for recognizing the modulation scheme with unknown channel matrix and noise variance for MIMO systems. By recognizing the fact that the received signals within an observation interval form into clusters and exploiting the intrinsic relationships of different digital modulation schemes, the modulation classification problem is transformed into a clustering problem without direct channel estimation for each modulation scheme and the maximum likelihood criterion is applied for the final classification decision. A central component of the proposed classifier is a method called centroid reconstruction, which exploits the structural relationships in constellation diagrams to reconstruct cluster centroids with fewer number of parameters. Furthermore, a method to initialize the cluster centroids is also proposed. The proposed MC classifier together with centroid reconstruction and initialization methods not only reduce the number of parameters to be estimated, but also help to initialize the centroids for enhanced convergence of expectation- maximization (EM) algorithm. Simulation results show that our algorithm can perform excellently even at low SNR and with very short observation interval length.

In cognitive radio networks, the performance of energy detection will be degraded significantly due to the cluster overlapping caused by noise uncertainty. To alleviate the noise uncertainty effect, a novel machine learning algorithm is proposed in this paper for cooperative spectrum sensing. The proposed algorithm incorporates fuzzy support vector machine and nonparallel hyperplane support vector machine. For membership assignment, kernel shadow c-means (KSCM) algorithm is utilized. Furthermore, the test statistics collected by the second users are arranged into a feature vector instead of being combined through weighted sum. Simulations results have shown that the proposed scheme, called NP-FSVM, is more robust to noise uncertainty than the existing methods.