IEEE transactions on neural networks and learning systems最新文献

In recent years, deep learning has emerged as one of the prominent technologies in hyperspectral anomaly detection (HAD) due to its powerful feature extraction capability. Deep learning-based methods aim to reconstruct the background image and use the reconstruction error as a criterion for anomaly detection. However, existing methods rarely exploit the spatial and spectral differences between background and anomalies to suppress anomaly reconstruction. To overcome this limitation, we propose a self-adaptive background-attentive network with anomaly suppression (BASNet). Specifically, a pyramid similarity downsampling module (PSDM) is employed to reduce the proportion of anomalies in the input image and improve multiscale background representation. To enhance the network's attention to background information, a multiscale attention module (MAM) is designed, and an inverse attention strategy is adopted to guide the network to focus on background regions. In addition, a depthwise separable blind-block convolution module (DBCM) is constructed for pixel reconstruction. This module applies weighted multiscale blind-block convolutions according to the anomaly score of each pixel, allowing the network to effectively handle anomalies at different scales. Due to the stochastic nature of deep learning, the optimal stopping epoch for network training is not fixed. Therefore, reconstructed image difference (RID) is developed to dynamically determine the stopping epoch. Extensive experiments on eight datasets demonstrate that the proposed BASNet achieves satisfactory detection performance.

Wideband radar is widely used for imaging and recognition in both military and civilian applications, such as remote sensing and guided strikes. With the advancement of electronic support systems, wideband radar is increasingly susceptible to complex electromagnetic environments. Among these jamming types, interrupted sampling repeater jamming (ISRJ) exhibits characteristics such as rapid interrupted sampling, high power, strong correlation, and complex modulation, which can generate numerous false targets and lead to misjudgments in radar systems. To address this issue, we propose an ISRJ suppression method based on joint morphological component analysis (MCA) and deep unfolding networks. We analyze the structural differences between target echoes and ISRJ signals, which can be sparsely represented under two approximately orthogonal dictionaries. The MCA optimization model is then introduced to simultaneously estimate the target echoes and ISRJ. To circumvent the challenges of hyperparameter determination, we derive an unfolding network based on the alternating direction method of multipliers (ADMM), enabling adaptive hyperparameter selection and optimization solutions with limited iterations in a data-driven manner. This method offers the advantages of adaptive hyperparameter tuning and robustness to varying jamming-to-signal-and-noise ratios (JSNRs). In addition, it achieves simultaneous jamming suppression and complete target echo preservation within the MCA framework. Finally, simulation and real-world experiments validate the effectiveness of this approach.

In real-world applications, hierarchical data with complex category structures pose significant challenges due to high dimensionality and potential incompleteness. Existing studies on hierarchical feature selection assume data are complete, without considering incomplete hierarchical data. However, in practice, obtaining a complete and uncorrupted feature space can be challenging due to physical, technical, and cost constraints in data collection. To address this issue, we propose a hierarchical feature selection method for incomplete hierarchical data. First, weight amendment matrices are introduced to shift the primary missing-data compensation from the feature space to the projection-weight space under hierarchical structural guidance. Second, we establish a hierarchical interaction framework that explicitly captures the implicit relationship between the weight amendment matrix and the hierarchical weight matrix via constrained optimization. Third, we design a sparse hierarchical regularization term that adaptively adjusts feature weights based on hierarchy levels. Extensive experiments on benchmark datasets demonstrate the proposed approach's superior performance over state-of-the-art methods, validating its enhanced effectiveness and progress in addressing the target problem domain.

Temporal processing is fundamental for biological and artificial intelligence systems, as it enables the comprehension of dynamic environments and facilitates timely responses. Spiking neural networks (SNNs) are well-suited for efficient temporal processing owing to their rich neuronal dynamics and event-driven computation. Given the recent surge in the development of SNNs, there is an urgent need for a comprehensive evaluation of their temporal processing capabilities. In this article, we first conduct an in-depth assessment of commonly used neuromorphic benchmarks, revealing critical limitations in their ability to evaluate the temporal processing capabilities of SNNs. To bridge this gap, we further introduce a benchmark suite consisting of four tasks characterized by rich temporal dependencies across multiple timescales. Utilizing this benchmark suite, we comprehensively evaluate recently introduced SNN approaches. Our results demonstrate substantial progress in neuron models, normalization methods, and neural architectures, while also uncovering challenges for online learning and long-range temporal dependencies compared to nonspiking sequence models. These findings establish a comprehensive understanding of the current state of SNNs in temporal processing and lay the foundation for the advancement of neuromorphic algorithms toward efficient operation in dynamic environments.