Identification of drug-target interactions is an indispensable part of drug discovery. While conventional shallow machine learning and recent deep learning methods based on chemogenomic properties of ...drugs and target proteins have pushed this prediction performance improvement to a new level, these methods are still difficult to adapt to novel structures. Alternatively, large-scale biological and pharmacological data provide new ways to accelerate drug-target interaction prediction. Here, we propose DrugMAN, a deep learning model for predicting drug-target interaction by integrating multiplex heterogeneous functional networks with a mutual attention network (MAN). DrugMAN uses a graph attention network-based integration algorithm to learn network-specific low-dimensional features for drugs and target proteins by integrating four drug networks and seven gene/protein networks, respectively. DrugMAN then captures interaction information between drug and target representations by a mutual attention network to improve drug-target prediction. DrugMAN achieves the best prediction performance under four different scenarios, especially in real-world scenarios. DrugMAN spotlights heterogeneous information to mine drug-target interactions and can be a powerful tool for drug discovery and drug repurposing.
In order to improve the accuracy of Indoor Human Activity Recognition based on the spatial location information, we proposed a recognition method using the convolutional neural network(CNN). We ...pre-process the raw spatial location data and transfer them into motion feature, frequency feature and statistic feature. These features are input into the CNN to do local feature analysis. After that, we got the characteristic output items, which have to be processed by the Softmax classifier, which can recognize six activities, including walking, sitting, lying, standing, jogging and jumping. By comparing the experimental results, the best recognition rate of different experimenters is 86.7%, which shows its feasibility.
We consider a new form of reinforcement learning (RL) that is based on opportunities to directly learn the optimal control policy and a general Markov decision process (MDP) framework devised to ...support these opportunities. Derivations of general classes of our control-based RL methods are presented, together with forms of exploration and exploitation in learning and applying the optimal control policy over time. Our general MDP framework extends the classical Bellman operator and optimality criteria by generalizing the definition and scope of a policy for any given state. We establish the convergence and optimality-both in general and within various control paradigms (e.g., piecewise linear control policies)-of our control-based methods through this general MDP framework, including convergence of \(Q\)-learning within the context of our MDP framework. Our empirical results demonstrate and quantify the significant benefits of our approach.
Please cite this paper as: Huai et al. (2010) A primary school outbreak of pandemic 2009 influenza A (H1N1) in China. Influenza and Other Respiratory Viruses DOI: 10.1111/j.1750‐2659.2010.00150.x.
...Background We investigated the first known outbreak of pandemic 2009 influenza A (H1N1) at a primary school in China.
Objectives To describe epidemiologic findings, identify risk factors associated with 2009 H1N1 illness, and inform national policy including school outbreak control and surveillance strategies.
Methods We conducted retrospective case finding by reviewing the school’s absentee log and retrieving medical records. Enhanced surveillance was implemented by requiring physicians to report any influenza‐like illness (ILI) cases to public health authorities. A case–control study was conducted to detect potential risk factors for 2009 H1N1 illness. A questionnaire was administered to 50 confirmed cases and 197 age‐, gender‐, and location‐matched controls randomly selected from student and population registries.
Results The attack rate was 4% (50/1314), and children from all grades were affected. When compared with controls, confirmed cases were more likely to have been exposed to persons with respiratory illness either in the home or classroom within 7 days of symptom onset (OR, 4·5, 95% CI: 1·9–10·7). No cases reported travel or contact with persons who had traveled outside of the country.
Conclusions Findings in this outbreak investigation, including risk of illness associated with contacting persons with respiratory illness, are consistent with those reported by others for seasonal influenza and 2009 H1N1 outbreaks in school. The outbreak confirmed that community‐level transmission of 2009 H1N1 virus was occurring in China and helped lead to changes in the national pandemic policy from containment to mitigation.
Automatic human action recognition plays an important role in many real applications, such as video surveillance, virtual reality and human-computer intelligent interaction, etc. The spatial ...complexity and time variability are the main challenges to be addressed. Most of the traditional methods are based on handcrafting video features, which leads to limited expressive power and difficulties of generalization. In recent years, with the rise of deep network, the deep learning method is applied in automatic human action recognition and achieves better performanceIn this paper we present a novel Convolutional Neural Network (CNN) based automatic human action recognition method, which automatically learns the spatial and temporal characteristics of the data to improve the recognition performance. Specifically, we preprocess the dataset to extract keyframes by using interframe difference method to reduce data redundancy and preserve the spatiotemporal characteristics of the data simultaneously; then we utilize the real-time key point recognition system Openpose to get the skeleton information. It consists of human joint points which are the input features of our recognition model. For model training, we use the large data set of UCF-101 which is the common benchmark in this filed. For model evaluation, we compare our method with the state-of-the-art methods. The experimental results show that our method achieves significant performance improvement on the dataset of UCF-101. Finally, based on the model we implement a system by using a Kinect V2 to record human action in real environment. Our system can automatically mark the range of human action and output the corresponding action labels in real time.
The mathematical analysis of epidemic-like behavior has a rich history, going all the way back to the seminal work of Bernoulli in 1766 5. More recently, mathematical models of epidemic-like behavior ...have received considerable attention in the research literature based on additional motivation from areas such as communication and social networks, cybersecurity systems, and financial markets; see, e.g., 6. The types of viral behaviors exhibited in many of these applications tend to be characterized by epidemic-like stochastic processes with time-varying parameters 12, 13. In this paper we consider variants of the classical mathematical model of epidemic-like behavior analyzed by Kurtz 8,7, Chapter 11, extending the analysis and results to first incorporate time-varying behavior for the infection and cure rates of the model and to then investigate structural properties of the interactions between local (micro) and global (macro) behaviors within the process. Specifically, we start by formally presenting an epidemic-like continuous-time, discretestate stochastic process in which each individual comprising the population can be either in a non-infected state or in an infected state, and where the rate at which the noninfected population is infected and the rate at which the infected population is cured are both functions of time. We established that, under general assumptions on the timevarying processes and under a mean-field scaling with respect to population size n, the stochastic processes converge to a continuous-time, continuous-state time-varying dynamical system. Then we study the stationary behavior of both the original stochastic process and the mean-field limiting dynamical system, and verify that they, in fact, have similar asymptotic behavior with respect to time. In other words, we establish that the following diagram is commutative.
In the paper, the low-frequency part of the random-drift error on micro inertial measurement unit (MIMU) is modeled. In order to solve the problem of high-frequency interference and ensure the ...original correlation of the samples, the wavelet soft-threshold method is used to remove high-frequency noise and the autoregressive integrated moving average model (ARIMA), which is more complex than the first-order Markov, is used to describe the slow-drift of the random-drift error. Finally, the model is applied to Kalman filter to smooth, filter and predict the MIMU measurement data. The experimental results show that the model has good dynamic tracking performance, the model better reflects the drift process of MIMU random error, and improves the measurement accuracy of micro-inertial devices.
Research on forward scattering visibility sensor Wu, Yushang; Zhang, Keke; Yang, Yingdong ...
2021 IEEE 4th International Conference on Automation, Electronics and Electrical Engineering (AUTEEE),
2021-Nov.-19
Conference Proceeding
A visibility sensor is designed based on the principle of forward scattering, and the sensor transmitting unit is equipped with one photodetector for light intensity feedback and the other ...photodetector for monitoring the light transmittance of the lens, which can produce stable emitted light and monitor the light transmittance. The principle of phase-locked amplification is used to measure the visibility in the receiving unit. The self-developed visibility sensor is compared with Vaisala PWD20 visibility sensor. In the range of 10 \sim 10000 meters, the measurement deviation of one minute visibility data is basically within ± 10%, and the measurement deviation of ten minute data is within ± 5%; In the range of 10000 ∼ 20000 meters, the measurement deviation of one minute data is basically within ± 20%, and the measurement deviation of ten minute data is within ± 10%. The visibility trend measured by the Vaisala PWD20 visibility sensor and the self-developed visibility sensor is consistent, which proves that the self-developed visibility sensor can meet the needs of visibility monitoring.