Star Topology Convolution for Graph Representation Learning

<p>We present a novel graph convolutional method called star topology convolution (STC). This method makes graph convolution more similar to conventional convolutional neural networks (CNNs) in Euclidean feature spaces. STC learns subgraphs which have a star topology rather than learning a fixed graph like most spectral methods. Due to the properties of a star topology, STC is graph-scale free (without a fixed graph size constraint). It has fewer parameters in its convolutional filter and is inductive, so it is more flexible and can be applied to large and evolving graphs. The convolutional filter is learnable and localized, similar to CNNs in Euclidean feature spaces, and maintains a good weight sharing property. To test the method, STC was compared with state-of-the-art graph convolutional methods in a supervised learning setting on six node properties prediction benchmark datasets: Cora, Citeseer, Pubmed, PPI, Ogbn-Arxiv, and Ogbn-MAG. The experimental results showed that STC achieved state-of-the-art performance on all these datasets and maintained good robustness. In an essential protein identification task, STC outperformed state-of-the-art essential protein identification methods.</p>

Network efficient topology for low power and lossy networks in smart corridor design using RPL

Purpose This paper aims to determine the network efficient topology for low power and lossy networks (LLNs) using routing protocol for LLN (RPL) with respect to the increase in network size and propose a novel approach to overcome the shortcomings of the existing models. Design/methodology/approach The authors have used Contiki OS/Cooja simulator to conduct experiments on primarily four topologies (star, bus/linear, ring/eclipse and random). They have implemented RPL protocol using Sky motes for each topology from 10, 20, 30 and up to 70 nodes. Consequently, after 24 h of experimentation, the readings have been noted and, alongside, a comprehensive comparative analysis has been performed based on the network density and metric parameters: packet delivery ratio (PDR), expected transmission (ETX) and power consumption. Further, a hybrid model is proposed where the additional factors of mobility, multiple sink and a combination of static and mobile nodes are introduced. The proposed model is then compared with the star model (all static nodes and star topology) and the dynamic model (all mobile nodes) to analyze the efficiency and network performance for different network sizes (28, 36, 38 and 44 nodes). The mobility is introduced using BonnMotion tool in Contiki OS. Findings Simulation results have shown that the star topology is most network efficient when compared with bus/linear, ring/eclipse and random topologies for low density and high scalable network. But when the same setup is compared with the proposed hybrid model, the proposed model shows a significant improvement and gives the best and efficient network performance with highest PDR (average improvement approximately 44.5%) and lowest ETX (average improvement approximately 49.5%) comparatively. Practical implications Also, these findings will benefit the deployment of smart devices used in advanced metering infrastructure, road side units and in various industrial applications such as traffic monitoring system, electronic toll collection and traffic analysis in the smart grid infrastructure. Originality/value The impact of topology is significant and detailed analysis is required to understand the impact of different topologies of the nodes in the network for the present and the future scenarios. As very few research studies have discussed this gap, this research paper is quintessential and shall open novel future potential direction. Also, the proposed approach of hybrid model with mobility has not been considered in the literature yet.

WDM Local-Area Networks

WDM-based broadcast-and-select transmission over optical passive-star topology can significantly enhance the speed of optical LANs/MANs. This type of optical LAN/MAN (or simply WDM LAN) can function using a variety of network architectures. In particular, WDM LANs can transmit packets between two nodes using direct or single-hop transmission or through intermediate nodes using multihop transmission, leading to broadly two types of network architectures: single-hop and multihop. The nodes in WDM LANs can employ two types of transmitters and receivers: tunable transmitter (TT) or fixed transmitter (FT) and tunable receiver (TR) or fixed receiver (FR), giving four types of WDM transceiver configurations: TT-TR, TT-FR, FT-TR, and FT-FR. Of the four configurations, the first three can realize single-hop communication, while the fourth generally leads to multihop networks. In this chapter, we describe various types of passive-star-based WDM LANs and examine their salient performance features. (141 words)

On the Capacity Region of Bipartite and Tripartite Entanglement Switching

We study a quantum switch serving a set of users in a star topology. The function of the switch is to create bipartite or tripartite entangled state among users at the highest possible rates at a fixed ratio. We model a set of randomized switching policies. Discovering that some are better than others, we present analytical results for the case where the switch stores one qubit per user, and find that the best policies outperform a time division multiplexing (TDM) policy for sharing the switch between bipartite and tripartite state generation. This performance improvement decreases as the number of users grows. The model is easily augmented to study the capacity region in the presence of qubit decoherence, obtaining similar results. Moreover, decoherence appears to have little effect on capacity. We also study a smaller class of policies when the switch stores two qubits per user.