Search

Recently, diffusion models have achieved remarkable performance in data generation, e.g., generating high-quality images. Nevertheless, chemistry molecules often have complex non-Euclidean spatial structures, with the behavior changing dynamically and unpredictably. Most existing diffusion models highly rely on computing the probability distribution, i.e., Gaussian distribution, in Euclidean space, which cannot capture internal non-Euclidean structures of molecules, especially the hierarchical structures of the implicit manifold surface represented by molecules. It has been observed that the complex hierarchical structures in hyperbolic embedding space become more prominent and easier to be captured. In order to leverage both the data generation power of diffusion models and the strong capability to extract complex geometric features of hyperbolic embedding, we propose to extend the diffusion model to hyperbolic manifolds for molecule generation, namely, Hyperbolic Graph Diffusion Model (HGDM). The proposed HGDM employs a hyperbolic variational autoencoder to generate the hyperbolic hidden representation of nodes and then a score-based hyperbolic graph neural network is used to learn the distribution in hyperbolic space. Numerical experimental results show that the proposed HGDM achieves higher performance on several molecular datasets, compared with state-of-the-art methods.

With development of new energy, inverter interfaced distributed generator (IIDG) has become one of the main types of distributed generator. Its control strategy has great influence on the dynamic behavior of distribution network. Conventional control modes can not provide inertia and are prone to instability. Based on the technology of virtual synchronous generator (VSG), the active power and frequency (P-o)) control, voltage and reactive power (Q-v) control and the inner loop control of IIDG are optimized respectively. The stability model is established after the transfer function of the system is derived, and the effect of damping factor on the output of the IIDG is analyzed. Based on the mechanism of stability, an adaptive damping control strategy is proposed. The strategy takes the deviation of output power as feedback and it can adaptively regulate the damping factor of the IIDG system. Hence the system has the advantages of fast response speed and small dynamic fluctuation. The strategy is evaluated by simulations based onPSCAD/EMTDC, which indicate that the adaptive damping control strategy has remarkable better dynamic performance compared with the fixed damping factor.

In order to improve the life quality for the cardiac patients and/or hypertensives and as well as reduce the expenses, a home electrocardiograph (ECG) and blood pressure (BP) telemonitoring network has been developed, which consists of home ECG and BP monitoring devices limited to a centralized monitoring server at the hospital. The telemonitoring center of the system is proposed in this paper. It is a surveillance of the home users by receiving the patients' medical data and displaying them in front of the medical staff and as well as an assistant for the medical information management. The software system of the telemonitoring center, which is based on the UNIX operation system, consists of a real-time monitoring system (RTMS) and a medical information management system (MIMS) to meet the requirements of medical service for the home users. A client-server architecture is used in the TRMS in order to exchange different medical data (ECG and blood pressure) with the different home devices via PSTN. The MIMS comprises several tools, including a file manager, an ECG analyzer, a BP analyzer, and a medical data backup tool, to assist diagnosis. The trial applications of the ECG and BP telemonitoring system in the Macao and Tsinghua community demonstrate the feasibility of the telemonitoring center.