No abstract

No abstract

Path Diversification is a new mechanism that can be used to select multiple paths between a given ingress and egress node pair using a quantified diversity measure to achieve maximum flow reliability. The path diversification mechanism is targeted at the end-to-end layer, but can be applied at any level for which a path discovery service is available. Path diversification also takes into account service requirements for low-latency or maximal reliability in selecting appropriate paths. Using this mechanism will allow future internetworking architectures to exploit naturally rich physical topologies to a far greater extent than is possible with shortest-path routing or equal-cost load balancing. We describe the path diversity metric and its application at various aggregation levels, and apply the path diversification process to 13 real-world network graphs as well as 4 synthetic topologies to asses the gain in flow reliability. Based on the analysis of flow reliability across a range of networks, we then extend our path diversity metric to create a compos- ite compensated total graph diversity metric that is representative of a particular topology's survivability with respect to distributed simultaneous link and node failures. We tune the accuracy of this metric having simulated the performance of each topology under a range of failure severities, and present the results. The topologies used are from nationalscale backbone networks with a variety of characteristics, which we characterize using standard graph-theoretic metrics. The end result is a compensated total graph diversity metric that accurately predicts the survivability of a given network topology.

The Generative Models have gained considerable attention in unsupervised learning via a new and practical framework called Generative Adversarial Networks (GAN) due to their outstanding data generation capability. Many GAN models have been proposed, and several practical applications have emerged in various domains of computer vision and machine learning. Despite GANs excellent success, there are still obstacles to stable training. The problems are Nash equilibrium, internal covariate shift, mode collapse, vanishing gradient, and lack of proper evaluation metrics. Therefore, stable training is a crucial issue in different applications for the success of GANs. Herein, we survey several training solutions proposed by different researchers to stabilize GAN training. We discuss (I) the original GAN model and its modified versions, (II) a detailed analysis of various GAN applications in different domains, and (III) a detailed study about the various GAN training obstacles as well as training solutions. Finally, we reveal several issues as well as research outlines to the topic.