3D medical image classification is challenging due to small, subtle lesions and substantial irrelevant context, which often mislead deep models. Inspired by the top-down diagnos-tic process of clinicians—first identifying anatomical context, then locating anomalies—we propose Prompt-Guided 3D Vision Transformer (PG 3D- ViT), a framework that simulates clinical reasoning through prompt-driven attention. To address limited 3D training data, PG3D- ViT leverages 2D masked auto encoder (MAE) pretraining to learn transferable image features. Through the prompt generation module, consistency difference analysis is performed between normal and abnormal samples to extract anatomical structure and global spatial prompt information related to the lesion context. These prompts are injected as query into a cross-attention mechanism, guiding the model to focus on lesion-relevant regions across the 3D volume. Evaluated on 7 public datasets spanning multiple modalities and pathologies, PG3D-ViT achieves a 1.88% average AUC improvement over state-of-the-art methods. The attention map visualizations demonstrate that the model can accurately localize lesion regions, validating the effectiveness of the clinical prompting mechanism in enhancing both the performance and interpretability of 3D medical image classification. The code is available at the provided link11https://github.comJUMED-P/PG3D-ViT

Data-driven transient stability assessment (TSA) methods have proven effectiveness in addressing the security challenges posed by the rapid integration of renewable energy into power systems. However, these methods often face high training costs, physical inconsistency, and limited generalization. To address these limitations, this article proposes a physics-augmented auxiliary learning (PA-AL) framework for TSA, implemented in a multigate mixture of expert model and a physics-informed paradigm for end-to-end stability margin prediction. Unlike conventional physics-informed neural network-based approaches, which rely on rotor angle trajectory predictions prone to cumulative errors, PA-AL embedding physics law by integrating auxiliary electrical velocity predictions. The proposed PA-AL framework employs a dual-phase physics-guided training scheme: first, dynamic-guided state representation learning, which captures system states by minimizing a loss function with auxiliary empirical terms and physics regularization derived from discrete swing equations, and second, mechanism-guided margin estimation calibration, which refines stability margin predictions through a mechanism-enhanced regularization term leveraging the relationship between margin and velocity deviation. Extensive case studies on the WSCC 9-bus, New England 39-bus, IEEE 68-bus, and Iceland 189-bus systems demonstrate the effectiveness of the proposed approach in improving stability assessment accuracy and maintaining physics consistency.

In recent years, numerous studies have explored how to allocate limited defense resource among meters to increase difficulties for launching attacks in power systems. However, existing studies often simplify this problem by assuming the linearity of attack cost functions, which creates an inevitable gap between theoretical analysis and practical scenarios. In this article, general nonlinear attack cost functions are considered in the formulation of the optimal defense resource problem, resulting in a mixed-integer nonlinear programming problem. This poses a challenging task for numerical methods or popular commercial solvers. To address this issue, an advanced slime mould algorithm with specialized initialization and evolutionary schemes is proposed. Specifically, a customized initialization strategy is designed such that the population can be easily initialized within the nonconvex feasible region aligning with the nonlinear constraints. Besides, in the evolutionary process, the fitness function is well-designed to encourage the individuals violating nonlinear constraints to evolve toward feasible directions. Comprehensive case studies verify that, compared to the state-of-the-art solvers, the proposed approach can achieve satisfactory defense resource allocation results in terms of feasibility, optimality, and real-time performance.

Single-cell RNA-sequencing (scRNA-seq) has emerged as a powerful technique for studying gene expression patterns at the single-cell level. Inferring gene regulatory networks (GRNs) from scRNA-seq data provides insight into cellular phenotypes from the genomic level. However, the high sparsity, noise and dropout events inherent in scRNA-seq data present challenges for GRN inference. In recent years, the dramatic increase in data on experimentally validated transcription factors binding to DNA has made it possible to infer GRNs by supervised methods. In this study, we address the problem of GRN inference by framing it as a graph link prediction task. In this paper, we propose a novel framework called GNNLink, which leverages known GRNs to deduce the potential regulatory interdependencies between genes. First, we preprocess the raw scRNA-seq data. Then, we introduce a graph convolutional network-based interaction graph encoder to effectively refine gene features by capturing interdependencies between nodes in the network. Finally, the inference of GRN is obtained by performing matrix completion operation on node features. The features obtained from model training can be applied to downstream tasks such as measuring similarity and inferring causality between gene pairs. To evaluate the performance of GNNLink, we compare it with six existing GRN reconstruction methods using seven scRNA-seq datasets. These datasets encompass diverse ground truth networks, including functional interaction networks, Loss of Function/Gain of Function data, non-specific ChIP-seq data and cell-type-specific ChIP-seq data. Our experimental results demonstrate that GNNLink achieves comparable or superior performance across these datasets, showcasing its robustness and accuracy. Furthermore, we observe consistent performance across datasets of varying scales. For reproducibility, we provide the data and source code of GNNLink on our GitHub repository: https://github.com/sdesignates/GNNLink.

Machine learning-based intelligent systems enhanced with Internet of Things (IoT) technologies have been widely developed and exploited to enable the real-time stability assessment of a large-scale electricity grid. However, it has been extensively recognized that the IoT-enabled communication network of power systems is vulnerable to cyberattacks. In particular, system operating states, critical attributes that act as input to the data-driven stability assessment, can be manipulated by malicious actors to mislead the system operator into making disastrous decisions and thus cause major blackouts and cascading events. In this article, we explore the vulnerability of the data-driven power system stability assessment, with a special emphasis on decision tree-based stability assessment (DTSA) approaches, and investigate the feasibility of constructing a physics-constrained adversarial attack (PCAA) to undermine the DTSA. The PCAA is formulated as a nonlinear programming problem considering the misclassification constraint, power limits, and bad data detection, computing potential adversarial perturbations that reverse the “stable/unstable” prediction of the real-time input while remaining invisible/stealthy. Extensive experiments based on the IEEE 68-bus system are conducted to evaluate the impact of PCAAs on predictions of DTSA and their transferability.

BACKGROUND: Building biological networks with a certain function is a challenge in systems biology. For the functionality of small (less than ten nodes) biological networks, most methods are implemented by exhausting all possible network topological spaces. This exhaustive approach is difficult to scale to large-scale biological networks. And regulatory relationships are complex and often nonlinear or non-monotonic, which makes inference using linear models challenging. RESULTS: In this paper, we propose a multi-layer perceptron-based differential equation method, which operates by training a fully connected neural network (NN) to simulate the transcription rate of genes in traditional differential equations. We verify whether the regulatory network constructed by the NN method can continue to achieve the expected biological function by verifying the degree of overlap between the regulatory network discovered by NN and the regulatory network constructed by the Hill function. And we validate our approach by adapting to noise signals, regulator knockout, and constructing large-scale gene regulatory networks using link-knockout techniques. We apply a real dataset (the mesoderm inducer Xenopus Brachyury expression) to construct the core topology of the gene regulatory network and find that Xbra is only strongly expressed at moderate levels of activin signaling. CONCLUSION: We have demonstrated from the results that this method has the ability to identify the underlying network topology and functional mechanisms, and can also be applied to larger and more complex gene network topologies.

This paper presents the design and implementation of a highly efficient Double-precision General Matrix Multiplication (DGEMM) based on Open BLAS for 64-bit ARMv8 eight-core processors. We adopt a theory-guided approach by first developing a performance model for this architecture and then using it to guide our exploration. The key enabler for a highly efficient DGEMM is a highly-optimized inner kernel GEBP developed in assembly language. We have obtained GEBP by (1) maximizing its compute-to-memory access ratios across all levels of the memory hierarchy in the ARMv8 architecture with its performance-critical block sizes being determined analytically, and (2) optimizing its computations through exploiting loop unrolling, instruction scheduling and software-implemented register rotation and taking advantage of A64 instructions to support efficient FMA operations, data transfers and prefetching. We have compared our DGEMM implemented in Open BLAS with another implemented in ATLAS (also in terms of a highly-optimized GEBP in assembly). Our implementation outperforms the one in ALTAS by improving the peak performance (efficiency) of DGEMM from 3.88 Gflops (80.9%) to 4.19 Gflops (87.2%) on one core and from 30.4 Gflops (79.2%) to 32.7 Gflops (85.3%) on eight cores. These results translate into substantial performance (efficiency) improvements by 7.79% on one core and 7.70% on eight cores. In addition, the efficiency of our implementation on one core is very close to the theoretical upper bound 91.5% obtained from micro-benchmarking. Our parallel implementation achieves good performance and scalability under varying thread counts across a range of matrix sizes evaluated.

HPCG has become a new metric for the design and ranking of HPC. By incorporating a local symmetric Gauss-Seidel preconditioned, HPCG implements the Conjugate Gradient method to solve a sparse linear system. HPCG performs poorly with irregular memory access and may consume a great deal of MPI resources when it is executed on supercomputers. This paper focuses on optimizing SpMV and the Gauss-Seidel preconditioned, the two most important kernels in HPCG. By evaluating the performance impacts of several representative sparse matrix formats, ELLPACK is selected due to its suitability for SIMD, resulting in a speedup of 2.3x for the SpMV kernel. Multi-coloring is performed for Gauss-Seidel, resulting in a speedup of 7.3x over the reference implementation. The CG convergence rate may also be improved after multi-coloring. Our experimental results show that our optimization process works well on supercomputers, achieving 6.5 Gflops on a CPU-only node. This has boosted the total HPCG Gflops by about 7x, giving rise to 80,151 Gflops on 8192 CPU-only Tianhe-2 nodes.

Emerging hardware and communication technologies enable new data stream applications that deal efficiently with very high rates of data updates. In this paper, we propose a novel index structure, termed the QDM-tree (quantized R*-tree with double memos), to support efficient similarity search over data streams. We integrate quantized minimum bounding spheres (QMBSs) and quantized minimum bounding rectangles (QMBRs) which can improve the cache behavior of QDM-tree due to effectively pack more entries in a node and reduce the tree height. Two compact main memory memos (Insert Memo and Delete Memo) can accelerate the speed of insert operations and convert the cost of update to the cost of insert. Therefore, the RamDisk technique reduces the cost of disk accesses to the cost of memory accesses. Theoretical analysis and experimental evaluation demonstrate that the QDM-tree significantly outperforms other state of the art R-tree variants with frequent updates, and is more suitable for massive data streams.