Ming Hui

Sparse General Matrix Multiplication (SpMM) is a fundamental kernel in iterative scientific computing and graph neural network inference. On single-GPU platforms, iterative SpMM suffers from two major performance bottlenecks: excessive memory traffic caused by repeated loading of Krylov basis vectors, and low computational efficiency due to irregular sparsity patterns. To address these challenges, this article proposes BLR-Krylov , an iterative SpMM optimization framework that integrates Communication-Avoiding Krylov (CA-Krylov) methods with a Block Low-Rank Sparse (BLR-Sparse) matrix format. BLR-Krylov first preprocesses sparse matrices into the BLR-Sparse format, where low-rank blocks are compressed through basis vector decomposition to reduce storage and arithmetic overhead. It then employs GPU-optimized matrix power kernels to batch-compute Krylov subspace basis vectors, effectively minimizing global memory accesses across iterations. Experimental results on multiple GPU architectures demonstrate that BLR-Krylov achieves a 4.0–6.0× speedup over existing communication-avoiding approaches in iterative workloads, increases global memory bandwidth utilization to 82%–87%, and improves tensor core utilization to 85%–92%, while maintaining numerical accuracy within 1 × 10 ^-6 . To facilitate reproducibility, partial source code is publicly available at https://github.com/19547035579zz-tech/BLR-Krylov.

Affective events are events that are typically associated with a positive or negative emotional state. For example, get food is a desirable event, while suffer from asthma is an undesirable event. Identifying affective events and their polarities is essential for many artificial intelligence (AI) tasks. Since sentiments toward an affective event is often implicit, affective event acquisition and classification have always been two great challenges. Encounter events are a special type of affective events. The goal of this article is to automatically acquire and classify stereotypically positive and negative encounter events in Chinese. First of all, we collect a comprehensive list of Chinese encounter event indicators based on the literature related to Chinese encounter events. Then, we automatically extract candidate encounter events from a text corpus using these indicators as keywords. In order to acquire high-quality encounter events, we design an event filtering scheme and generate a novel stopword list. Finally, we classify these obtained encounter events exploiting the indicators. Experimental results show that the proposed approaches to acquire and classify Chinese encounter events perform promisingly well. One key advantage of the method is that no labelling and training of data is required. The automatically constructed common affective event knowledge base contains more than 38,000 encounter events annotated with affective polarity labels (negative or positive), which is currently the only resource of common affective events in Chinese language. Related resources including encounter event indicators and encounter events will be released after the publication of this article.

This study proposes a broadband, high-gain, Multi-Resonant antenna design for the 5G Sub-6 GHz band, using a fusion method to effectively cover the key Sub-6 GHz frequency band. The proposed design achieves optimized performance by incorporating symbiotic and parasitic elements to enhance selectivity and etching slotlines to improve impedance matching and expand the bandwidth. Simulation results show that the antenna operates at a center frequency of 3.5 GHz, with a relative bandwidth of 16.04% (3.21–3.77 GHz) and an average gain of 9.5 dBi.

To improve the informatization construction and intelligent decision-making level of river and lake basin management, the water quality of a digital twin basin was considered as the starting point and a water quality evaluation platform for Chuancheng River and Baihe River in Nanyang City, Henan Province was established. Based on digital twin technology, the platform establishes a virtual space city model, uses the long short-term memory algorithm to establish a water quality prediction model, draws the distribution of water pollution factors in two dimensions based on Kriging interpolation, simulates the pollutant diffusion in three dimensions based on numerical simulation, and finally builds a visual platform for evaluation and analysis. The platform combines digital twin with three models: one-dimensional (1D) water quality data processing, two-dimensional pollutant distribution, and three-dimensional (3D) pollutant diffusion simulation to achieve visual and comprehensive management of water quality assessment. Compared with the traditional 1D water quality data management platform, the proposed digital twin 3D urban river water quality evaluation platform system solves the problems of low visualization degree, single management, and incomplete analysis, as well as provides a new technical guarantee for the management of urban river water quality.

This study proposed a millimetre‐wave integrated substrate gap waveguide (ISGW) filtering antenna featuring four adjustable radiation nulls, with two located within each of its upper and lower stop bands. These nulls can be tuned by altering the dimensions of stepped‐impedance resonators, complementary U‐slots, and passive coplanar parasitic patches. Antennas offer the benefits of individually adjustable radiation nulls and the flexible modification of selectivity and gain curve roll‐off. Simulation results indicated that the antenna operated at a central frequency of 25.4 GHz, with a relative bandwidth of 14.3% (23.76–27.04 GHz), and attained an average gain of 7.6 dBi.

This paper introduces a millimeter-wave Integrated Substrate Gap Waveguide (ISGW) filtering antenna featuring four adjustable radiation nulls, with two located in each of its upper and lower stop bands. These nulls are tunable by altering the dimensions of stepped-impedance resonators, complementary U-slots, and passive coplanar parasitic patches. The antenna offers the benefits of individually adjustable radiation nulls and the flexible modification of selectivity and gain curve roll-off. Simulation results indicate the antenna operates at a central frequency of 25.4 GHz, with a relative bandwidth of 14.3% (23.76–27.04 GHz), and attains an average gain of 7.6 dBi.

To fulfill the connectivity demands in massive machine-type communications (mMTC), this paper investigates a generalized multi-user sparse superposition transmission (GMUSST) technology based on position index modulation. Due to the high computation complexity of maximum likelihood (ML) multi-user detection, a low complexity multi-path successive interference cancellation (MSIC) multi-user detector is introduced to achieve near-ML detector's block error ratio (BLER) performance. Furthermore, considering that each user is only concerned with their own transmitted signal in the downlink GMUSST system, we propose a minimum mean square error-based SIC (MMSE-SIC) detector, which can directly extract the user's transmission signal from the received superimposed signal of multiple users and is verified compared with MSIC detector. Simulation results show that the GMUSST can achieve better transmission reliability than the existing polar coded sparse code multiple access (PC-SCMA) in the short packet communication scenarios. Especially with the hybrid automatic repeat request mechanism, GMUSST requires fewer retransmissions to achieve the same BLER performance compared to PC-SCMA.

Unsupervised salient object detection is an important task in many real-world scenarios where pixel-wise label information is of scarce availability. Despite its significance, this problem remains rarely explored, with a few works that consider unsupervised salient object detection methods based on the fused graph from the sum fusion of multiple deep feature similarity matrices. However, these methods ignore the interrelation of the low-level feature similarity matrices and the high-level semantic similarity matrice, which degrades the quality of the fused graph. In this article, we propose a semantic-consistency-guided multi-graph fusion learning algorithm for unsupervised saliency detection, where the consistency and inconsistency between multiple low-level feature similarity matrices and the high-level semantic similarity matrice are explored to promote the robustness and quality of the fused graph. In the first stage, a semantic-consistency-guided multi-graph fusion learning method is proposed to exploit consistency and inconsistency of multiple low-level deep features and the high-level semantic feature. The semantic-consistency-guided similarity matrices are computed for preliminary saliency ranking. In the following saliency refinement stage, the semantic-enhanced similarity matrices are built by the cross diffusion to fuse the multiple low-level deep features and the high semantic deep feature. Based on the semantic-enhanced similarity matrices, the refinement saliency maps are calculated in a semantic-enhanced cellular automata manner. Furthermore, the final ensemble stage of the large margin semi-supervised classification views the preliminary ranking results and refinement results as features, adopts the large margin graphs for saliency ensemble. Extensive evaluations over four benchmark datasets show that the proposed unsupervised method performs favorably against the state-of-the-art approaches and is competitive with some supervised deep learning-based methods.

This paper presents a millimeter-wave Integrated Substrate Gap Waveguide (ISGW) filtering antenna with four controllable radiation nulls, two on each of its upper and lower stop bands. These radiation nulls can be adjusted by manipulating the dimensions of the stepped-impedance resonators (SIRs), complementary U-slots, and passive coplanar parasitic patches. This filtering antenna has the advantages of separately controllable radiation nulls and flexible adjustment of selectivity and gain curve roll-off. The simulation results demonstrate that the antenna operates at a center frequency of 25.4 GHz, with a relative bandwidth of 14.3% (23.76-27.04 GHz), and achieves a realizable average gain of 7.6 dBi.