Graph neural networks (GNNs) are specifically designed for graph-structured data and have gained significant attention. However, training GNN on large-scale graphs remains challenging due to the iterative aggregation of high-dimensional features and high computational complexity. Graph sparsity often leads to inefficient memory access and prolonged training times. To address these issues, we propose the Maximum Common Neighbor Graph Reordering (MaCoN-GR) algorithm, which optimizes the memory layout by reducing the physical distance between nodes and their neighbors. We first evaluate MaCoN-GR through its performance on Sparse Matrix-Vector Multiplication (SpMV), a core operation in GNN training. Experimental results demonstrate that MaCoN-GR achieves speedups ranging from 1.10× to 1.35×, indicating that our method effectively enhances memory access efficiency and reduces latency. Building upon this result, we further apply MaCoN-GR to accelerate GNN training. Within the Topology-Oriented Sampling (TOS) model, GraphSAGE, MaCoN-GR achieves up to 1.08× speedup on large-scale datasets, along with notable improvements in accuracy. In the Feature-Oriented Sampling (FOS) framework, we evaluate both MaCoN-GR and METIS partitioning on FOSGNN, achieving speedups between 1.12× and 2.34×, while also improving convergence and model accuracy. These results highlight the effectiveness of MaCoN-GR in optimizing both memory performance and GNN training efficiency.

Graph Neural Networks (GNNs) demonstrate remarkable learning efficacy on graph-structured data across various real-world domains. However, due to large-scale graph datasets, it is impractical to train the complete graph on a single GPU. Subgraph-level parallel training has proven to be an effective approach for GNN training on large-scale graphs. Nonetheless, this approach presents certain issues: 1. Due to the absence of neighboring vertices at the boundaries, gradient estimation during the training process may exhibit systemic bias, potentially compromising training accuracy; 2. Existing training frameworks often neglect to consider computational capability differences among trainers when allocating the workloads. We propose GALA-GNN to alleviate the aforementioned issues: 1. We introduce a heuristic edge partitioning algorithm, Neighbor Expansion Simulated Annealing, i.e., NESA, which minimizes the number of vertices with missing neighborhoods in subgraphs. Additionally, it allows for the division of the original graph into subgraphs with uneven workloads by setting a threshold manually; 2. We present Computation-Aware Subgraph Enlarge, which not only prevents significant loss in training accuracy but also reduces idle time for high-computation trainers during the training process, thereby enhancing overall computational resource utilization. We compare our proposed GALA-GNN with the current state-of-the-art GNN training framework, DGL. On the medium-scale dataset Flickr, GALA-GNN achieves up to 7.63x speedup, and on the large-scale dataset ogbn-products, it achieves up to 4.84x speedup, without causing significant loss in training accuracy.

Rehabilitation therapy involving repetitive exer-cises targeting specific human joints under the supervision of a doctor is crucial for patients with movement disorders. Yet the cost of commuting and the demand for medical resources are inconvenient for patients. Human-computer interaction can provide remote rehabilitation guidance for patients at home through human pose estimation technology, but privacy concerns with optical sensors and the cost and discomfort of wearable sensors have hindered progress in this field. To address the above challenges, we propose mmHPE, an innovative 3D human pose estimation framework that uses millimeter-wave (mmWave) radar sensors. It initially manipulates the raw data captured from radar sensors to generate a spatiotemporal sequence point cloud dataset. Afterward, we create a Convolutional Neural Network (CNN) that is linked to a Bidirectional Long Short-Term Memory (Bi-LSTM) network. Moreover, a multi-head attention mechanism is employed to boost the network's performance and to accurately estimate the locations of human skeletons. Ultimately, the 21 points with the corre-sponding human pose position are successfully reconstructed. We investigate the mmHPE framework's feasibility and cross-domain stability in different home environments in real-world scenarios. This innovation proffers patients a convenient and privacy-conscious solution for their rehabilitation training req-uisites.

Sparse matrix-vector multiplication (SpMV) is crucial in many scientific and engineering applications, particularly concerning the effectiveness of different sparse matrix storage formats for various architectures, no single format excels across all hardware. Previous research has focused on trying different algorithms to build predictors for the best format, yet it overlooked how to address the issue of the best format changing in the same hardware environment and how to reduce prediction overhead rather than merely considering the overhead in building predictors. This paper proposes a novel classification algorithm for optimizing sparse matrix storage formats, DyLaClass, based on dynamic labeling and flexible feature selection. Particularly, we introduce mixed labels and features with strong correlations, allowing us to achieve ultra-high prediction accuracy with minimal feature inputs, significantly reducing feature extraction overhead. For the first time, we propose the concept of the most suitable storage format rather than the best storage format, which can stably predict changes in the best format for the same matrix across multiple SpMV executions. We further demonstrate the proposed method on the University of Florida’s public sparse matrix collection dataset. Experimental results show that compared to existing work, our method achieves up to 91% classification accuracy. Using two different hardware platforms for verification, the proposed method outperforms existing methods by 1.26 to 1.43 times. Most importantly, the stability of the proposed prediction model is 25.5% higher than previous methods, greatly increasing the feasibility of the model in practical field applications.

Public vehicles and personal navigation assistants have become increasingly equipped with single-frequency global positioning system (GPS) receivers or loggers. These commonly used terminals offer an inexpensive way for acquiring large volumes of GPS traces, which contain information pertaining to road position and traffic rules. Using this new type of spatial data resource, we propose a novel approach for generating high-quality routable road maps. In this approach, a simplified road network graph model uses circular boundaries to separate all GPS traces into road intersections and road segments and builds road networks that maintain their identical geometric topologies through the entry/exit points at the original boundaries. One difficulty inherent to this type of approach is how to best determine the appropriate spatial coverage for road intersections. Conflict points among GPS traces that have large intersection angles usually occur within the physical areas of road intersections, particularly those involving left turns. Therefore, we determined a proper circle boundary for individual road intersections by conducting a spatial analysis of such feature points. This approach was implemented using Python and PostgreSQL/PostGIS and was tested in Huaibei City, China. Based on a comparison with human-interpreted results, the automatically generated routable road map was demonstrated to be of high quality and displayed detailed road networks with turning at various at-grade intersections, interchanges and U-turns.

Coal fires are a common problem in most coal-producing countries in the world, which could cause serious environmental, economic and other problems. Rujigou coalfield in Shizuishan City, Ningxia, NW China, is well known for being the storehouse of anthracite coal. The coalfield is also well known for its coal fires among all coalfields in China. In the study, an attempt was made to study the coal fires dynamics in Rujigou coalfield during the range time from 2000 to 2007 using a TIR approach based on the multi-temporal nighttime Landsat data. A quantitative analysis was made over coal fires changes in the spatial extent. The results showed that the general spreading direction of coal fires was toward the north or northeast. From 2001 to 2007, the coal fires had a huge increasement with annual average area of 0.14 km2.

This study developed a method for Soil Organic Matter (SOM) mapping in a lithoidal mountainous area using a stratified strategy under fuzzy inference framework. Environmental variables derived from terrain analysis and remote sensing incorporated with a small number of field samples were used to predict spatial variation of SOM. A new variable, bare soil ratio, representing the fractions of bare soil area inside a pixel of remote sensing images resulted from sub-pixel unmixing technology is used specifically for rocky upland area. A case study in the Hushiha area of North China has shown an improved spatial detail and accuracy of SOM compared to the available traditional soil map of Heibei Province. The proposed method takes full advantages of data with free access such as SRTM DEM and remote sensing images of HJ-A which exhibits massive potential to map SOM efficiently at low cost for a large area.

This paper presents a weighted clustering algorithm based on the physical attraction model, which improves the physical attraction model by assigning a different weight to the position points on a GPS trace for a fast convergence according to their velocity and directional changes. The physical attraction model pulls together traces that belong on the same road in response to simulated potential energy wells created around each trace. Assuming a vehicle with a high velocity has a little derivation to the road it runs on, we assign a high weight to its trace in the physical attraction model so that the clustering progress converges rapidly and closely to road. Within the clustering process, an angle-threshold based smoothing filter is appended to keep the consistency between the changes of those points on the adjusted trace at each loop. This algorithm was demonstrated to enable to effectively clarify those vehicle GPS traces on the same road. In comparison of previous work, it also shows an improved quality grouping GPS traces near road-crossing area by embedding the smoothing filter in clustering process.

This paper presents a spectral unmixing approach that is implemented using linear unminxing method by a genetic algorithm. The unmixing is constrained not only by the negativity and sum-to-one of the abundances of endmembers at each pixel but also by the spatial autocorrelation of their abundances among eight neighbor pixels. The Moran's I indices are proposed to describe the spatial autocorrelation among a pixel and its neighborhood. Based on the above constraints, the objective of unmixing by genetic algorithm is to minimize the mean square error of mixed spectral values. We tested this approach using Chinese HJ-satellite images and obtained an acceptable result.