Payment channel networks (PCNs) are generally regarded as one of the most effective and promising scalability solutions for blockchain-based cryptocurrency systems, but suffer the issues of low success ratio and long confirmation latency in processing transactions. In this paper, we demonstrate the feasibility of tremendously increasing the success ratio of transactions and improving their execution efficiency by enhancing network nodes' connectivity and enforcing a balanced network channel capacity. To implement such ideas, multiple designs have been made. First, to extent nodes connectivity, we transform the nearly-linear ordered nodes into a star payment structured typology, and design an incentive financing mechanism to restructure a new landmark routing typology design. Especially, for the marginalized or dissociative nodes, we utilize specific financial loan strategies to encourage them to (re)join the system. Besides, we propose the Power Atomic Multi-Path Payments (Power AMP) traffic distribution method, which hierarchically allocates the bottleneck's currently-available capacity (to replace the random or equal division used in traditional AMP), and thus archives a balanced traffic usage. With such efforts, we improve the transaction success ratio and efficiency of transaction exertion by order of magnitude—compared with traditional PCN using the benchmark of landmark route, our method improves the success ratio by 11.06%.

Machine unlearning is the process through which a deployed machine learning model is enforced to forget about some of its training data items. It normally generates two machine learning models, the original model and the unlearned model, indicating training results before and after data items are deleted. However, recent studies find that machine unlearning is vulnerable to membership inference attacks—as the directivity of training and nontraining data (i.e., data items in the training set have high posterior probabilities), the attackers can utilize this property to infer whether an item has been used for original model training. Nevertheless, such attacks are incapable in label-only settings, in which the attackers are infeasible to get the posteriors. In this paper, we propose a new label-only membership inference attack scheme targeted at machine unlearning to eliminate the dependence on posteriors. Our heuristic is that injected turbulence on candidate samples will present different behaviors for training and nontraining data. Thus, in our scheme, the attacker iteratively query on the original/unlearned models and inject turbulence to change their predicting labels; it determines whether an item is having-been-delated by observing the disturbance amplitude. Extensive experiments (i.e., on MNIST, CIFAR10, CIFAR100, and STL10 data sets) show that our method achieves high inference accuracy (measured by AUC) in label-only settings, for example, AUC = 0.96 for MNIST data set. Besides, we analyze the existing countermeasures in mitigating inference attacks and find that our scheme can bypass most of them.

It is urgent to identify the development of the Corona Virus Disease 2019 (COVID-19) in countries around the world. Therefore, visualization is particularly important for monitoring the COVID-19. In this paper, we visually analyze the real-time data of COVID-19, to monitor the trend of COVID-19 in the form of charts. At present, the COVID-19 is still spreading. However, in the existing works, the visualization of COVID-19 data has not established a certain connection between the forecast of the epidemic data and the forecast of the epidemic. To better predict the development trend of the COVID-19, we establish a logistic growth model to predict the development of the epidemic by using the same data source in the visualization. However, the logistic growth model only has a single feature. To predict the epidemic situation in an all-round way, we also predict the development trend of the COVID-19 based on the Susceptible Exposed Infected Removed epidemic model with multiple features. We fit the data predicted by the model to the real COVID-19 epidemic data. The simulation results show that the predicted epidemic development trend is consistent with the actual epidemic development trend, and our model performs well in predicting the trend of COVID-19.