Fuqi Mao

Accurate traffic state prediction plays an important role in traffic guidance, travel planning, etc. Due to the existence of complex spatio-temporal relationships, there are some challenges in forecasting. Firstly, in terms of spatial correlation, some models only consider the road network structure information, and ignore the relative location relationships between nodes. Secondly, some models ignore the different impacts of nodes in the global road network on traffic. To solve these problems, we propose a new traffic state-forecasting model, namely, spatio-temporal attention-gated recurrent neural network (ST-AGRNN). In the proposed model, structure-based and location-based localized spatial features are obtained simultaneously by Graph Convolutional Networks (GCNs) and DeepWalk. The localized temporal features are obtained by gated recurrent unit (GRU). The attention-based approach is used to obtain global spatio-temporal features. Experimental validation is performed with two real-world public datasets, and the results show that the ST-AGRNN model outperforms the state-of-the-art methods.

As a typical spatiotemporal problem, there are three main challenges in traffic forecasting. First, the road network is a nonregular topology, and it is difficult to extract complex spatial dependence accurately. Second, there are short- and long-term dependencies between traffic dates. Third, there are many other factors besides the influence of spatiotemporal dependence, such as semantic characteristics. To address these issues, we propose a spatiotemporal DeepWalk gated recurrent unit model (ST-DWGRU), a deep learning framework that fuses spatial, temporal, and semantic features for traffic speed forecasting. In the framework, the spatial dependency between nodes of an entire road network is extracted by graph convolutional network (GCN), whereas the temporal dependency between speeds is captured by a gated recurrent unit network (GRU). DeepWalk is used to extract semantic information from road networks. Three publicly available datasets with different time granularities of 15, 30, and 60 min are used to validate the short- and long-time prediction effect of this model. The results show that the ST-DWGRU model significantly outperforms the state-of-the-art baselines.

As an important tool to study the microstructure and properties of materials, High Resolution Transmission Electron Microscope (HRTEM) images can obtain the lattice fringe image (reflecting the crystal plane spacing information), structure image and individual atom image (which reflects the configuration of atoms or atomic groups in crystal structure). Despite the rapid development of HTTEM devices, HRTEM images still have limited achievable resolution for human visual system. With the rapid development of deep learning technology in recent years, researchers are actively exploring the Super-resolution (SR) model based on deep learning, and the model has reached the current best level in various SR benchmarks. Using SR to reconstruct high-resolution HRTEM image is helpful to the material science research. However, there is one core issue that has not been resolved: most of these super-resolution methods require the training data to exist in pairs. In actual scenarios, especially for HRTEM images, there are no corresponding HR images. To reconstruct high quality HRTEM image, a novel Super-Resolution architecture for HRTEM images is proposed in this paper. Borrowing the idea from Dual Regression Networks (DRN), we introduce an additional dual regression structure to ESRGAN, by training the model with unpaired HRTEM images and paired nature images. Results of extensive benchmark experiments demonstrate that the proposed method achieves better performance than the most resent SISR methods with both quantitative and visual results.