Online electroencephalograph (EEG) classification is a core service of recently booming brain e-health, but its performance often becomes unstable because (1) conventional end-to-end models (e.g., deep neural network, DNN) largely remain static, while brain states of diseases are highly dynamic and exhibits significant individuality; and (2) EEG analytics are too complicated and have to be sustained by advanced computing services. This study adopts an automatic machine learning method to construct a dual-CNN (convolutional neural network) of high performance in terms of both accuracy and efficiency. The model can optimize its hyperparameters continuously on its own initiative. Experimental results in the evaluation of depression using real EEG datasets indicate that (1) the proposed method executes 3.5 times faster compared with a conventional counterpart; (2) the dual-CNN gains a significant performance improvement (versus CapsuleNet and Resnet-16) in identifying Major Depression Disorder (MDD) with accuracy, sensitivity, and specificity up to 98.81, 98.36, and 99.31 percent respectively; and those for treatment outcome are 99.52, 99.63, and 99.37 percent respectively, and (3) classification can be completed several hundred times faster than EEG being collected upon a COTS computer.

Dengue is a serious vector-borne disease, and incidence rates have significantly increased during the past few years, particularly in 2014 in Guangzhou. The current situation is more complicated, due to various factors such as climate warming, urbanization, population increase, and human mobility. The purpose of this study is to detect dengue transmission patterns and identify the disease dispersion dynamics in Guangzhou, China. We conducted surveys in 12 districts of Guangzhou, and collected daily data of Breteau index (BI) and reported cases between September and November 2014 from the public health authority reports. Based on the available data and the Ross-Macdonald theory, we propose a dengue transmission model that systematically integrates entomologic, demographic, and environmental information. In this model, we use (1) BI data and geographic variables to evaluate the spatial heterogeneities of Aedes mosquitoes, (2) a radiation model to simulate the daily mobility of humans, and (3) a Markov chain Monte Carlo (MCMC) method to estimate the model parameters. By implementing our proposed model, we can (1) estimate the incidence rates of dengue, and trace the infection time and locations, (2) assess risk factors and evaluate the infection threat in a city, and (3) evaluate the primary diffusion process in different districts. From the results, we can see that dengue infections exhibited a spatial and temporal variation during 2014 in Guangzhou. We find that urbanization, vector activities, and human behavior play significant roles in shaping the dengue outbreak and the patterns of its spread. This study offers useful information on dengue dynamics, which can help policy makers improve control and prevention measures. Dengue transmission is a spatio-temporal process with interactions between hosts, vectors, and viruses. Its transmission also involves multiple complex or even hidden factors, such as climate, social environment, vector ecology, and host mobility. These complexities make the underlying process of dengue transmission difficult to clarify. We address how the patterns of dengue transmission can be inferred by investigating the 2014 dengue outbreak in the city of Guangzhou, China, taking the available surveillance data and applying mathematical models and computational methods. We can then estimate the distribution of dengue infections and identify the transmission mechanisms. In our study, we systematically investigate the critical factors, enabling us to estimate the real patterns and dynamics of dengue transmission beyond the surveillance data.

Evidence shows that a small number of line contingencies in power systems may cause a large-scale blackout due to the effects of cascading failures. With the development of new technologies and the growing number of heterogeneous participants, a modern/smart grid should be able to self-heal its internal disturbances by continually performing self-assessment to deter, detect, respond to and restore from unpredictable contingencies. Along this line, this research focuses on the problem of how to prevent the occurrence of cascading failures through load shedding by considering heterogeneous shedding costs of grid participants. A fair load-shedding algorithm is proposed to solve the problem in a decentralized manner, where a load-shedding participant need only monitor its own operational status and interact with its neighboring participants. Using an embedded feedback mechanism, the fair load-shedding algorithm can determine a marginal compensation price for each load-shedding participant in real time based on the proportional fairness criterion, without knowing the shedding costs of the participants. Such fairly determined compensations can help motivate loaders/generators to actively participate in the load shedding in the face of internal disturbances. Finally, the properties of the load-shedding algorithm are evaluated by carrying out an experimental study on the standard IEEE 30 bus system. The study will offer new insights into emergency planning and design improvement of self-healing smart grids

Sea surface temperature (SST) prediction has received increasing attention in recent years due to its paramount importance in the various fields of oceanography. Existing studies have shown that neural networks are particularly effective in making accurate SST predictions by efficiently capturing spatiotemporal dependencies in SST data. Among various models, the ConvLSTM framework is notably prominent. This model skillfully combines convolutional neural networks (CNNs) with recurrent neural networks (RNNs), enabling it to simultaneously capture spatiotemporal dependencies within a single computational framework. To overcome the limitation that CNNs primarily capture local spatial information, in this paper we propose a novel model named DatLSTM that integrates a deformable attention transformer (DAT) module into the ConvLSTM framework, thereby enhancing its ability to process more complex spatial relationships effectively. Specifically, the DAT module adaptively focuses on salient features in space, while ConvLSTM further captures the temporal dependencies of spatial correlations in the SST data. In this way, DatLSTM can adaptively capture complex spatiotemporal dependencies between the preceding and current states within ConvLSTM. To evaluate the performance of the DatLSTM model, we conducted short-term SST forecasts in the Bohai Sea region with forecast lead times ranging from 1 to 10 days and compared its efficacy against several benchmark models, including ConvLSTM, PredRNN, TCTN, and SwinLSTM. Our experimental results show that the proposed model outperforms all of these models in terms of multiple evaluation metrics short-term SST prediction. The proposed model offers a new predictive learning method for improving the accuracy of spatiotemporal predictions in various domains, including meteorology, oceanography, and climate science.

The transmission networks of Plasmodium vivax characterize how the parasite transmits from one location to another, which are informative and insightful for public health policy makers to accurately predict the patterns of its geographical spread. However, such networks are not apparent from surveillance data because P. vivax transmission can be affected by many factors, such as the biological characteristics of mosquitoes and the mobility of human beings. Here, we pay special attention to the problem of how to infer the underlying transmission networks of P. vivax based on available tempo-spatial patterns of reported cases. We first define a spatial transmission model, which involves representing both the heterogeneous transmission potential of P. vivax at individual locations and the mobility of infected populations among different locations. Based on the proposed transmission model, we further introduce a recurrent neural network model to infer the transmission networks from surveillance data. Specifically, in this model, we take into account multiple real-world factors, including the length of P. vivax incubation period, the impact of malaria control at different locations, and the total number of imported cases. We implement our proposed models by focusing on the P. vivax transmission among 62 towns in Yunnan province, People's Republic China, which have been experiencing high malaria transmission in the past years. By conducting scenario analysis with respect to different numbers of imported cases, we can (i) infer the underlying P. vivax transmission networks, (ii) estimate the number of imported cases for each individual town, and (iii) quantify the roles of individual towns in the geographical spread of P. vivax. The demonstrated models have presented a general means for inferring the underlying transmission networks from surveillance data. The inferred networks will offer new insights into how to improve the predictability of P. vivax transmission. The transmission of Plasmodium vivax has induced enormous public health problems at the global level. Natural transmission of P. vivax depends on interactions between anopheles mosquitoes and human beings. There are two important factors that influence its geographical spread. First, different locations may have different risks of infection due to their heterogeneous environmental and demographical profiles. Second, human mobility may bring pathogens from high-transmission locations to low-transmission locations. In view of this, to effectively and efficiently control the geographical spread of P. vivax, it would be desirable for us to characterize how it transmits from one location to another. To achieve this, we first build a spatial transmission model to characterize both the heterogeneous infection risks at individual locations and the underlying mobility of infected populations. By doing so, we can further infer the underlying P. vivax transmission networks from tempo-spatial surveillance data by using a machine learning method (i.e., based on a recurrent neural network model). Our study offers new insights into malaria surveillance and control from the viewpoint of both system modeling and machine learning.