检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and signal processing (ICASSP)

作者： Qiangsheng Feng Jiwei Qin Jie Ma School of Computer Science and Technology Xinjiang University Key Laboratory of Signal Detection and Processing

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Bayesian Personalized Ranking (BPR) is a widely used optimization function in GNN-based recommender systems, and negative samples are usually obtained through the Random Negative Sampling (RNS) method during BPR training. However, from the gradient perspective, RNS tends to select low-quality samples with minimal information, resulting in small gradients. These small gradients contribute little to BPR optimization, limiting the model’s ability to effectively distinguish between positive and negative samples. To alleviate this issue, we propose a general negative sample information enhancement method: Enhancing Random Negative Sampling (E-RNS), which constructs hard negative samples by enhancing the information in randomly selected negative samples. Specifically, in the Noise Injection step, it generates initial noise and injects a certain amount of noise in the same direction into the vector dimensions of positive samples to create enriched information. Then, in the Information Fusion step, this enriched information is mixed with the negative samples to synthesize new hard negative samples. Extensive experiments demonstrate that applying E-RNS to GNN-based recommender models significantly improves performance.

关键词： Training Limiting Noise Information filters Vectors Acoustics Bayes methods Speech processing Optimization Recommender systems

来源：评论

学校读者我要写书评

暂无评论

Mainlobe Deceptive Interference Suppression with Planar Frequency Diverse Array MIMO radar

引用

IEEE Transactions on Aerospace and Electronic Systems 2025年

作者： Zhang, Mengdi Lu, Jiahao Xu, Jingwei Li, Shiyin China University of Mining and Technology School of Information and Control Engineering Xuzhou221116 China Xidian University National Laboratory of Radar Signal Processing Xi'an710071 China

Frequency diversity array (FDA) is capable of suppressing mainlobe deceptive interference with delays exceeding one pulse relative to the true target, known as inter-pulse interference. However, current FDA-based interference suppression methods are ineffective when both inter-pulse and intra-pulse interferences coexist. To this end, we propose a two-stage mainlobe deceptive interference suppression method with planar FDA multiple-input-multiple-output (PFDA-MIMO) radar: 1) Horizontal transmit spatial frequency compensation (HTSFC) is applied using the range information of each range bin to differentiate between inter-pulse interferences and target. Subsequently, data-independent beamforming is performed to suppress inter-pulse interferences in horizontal spatial frequency domain;2) To distinguish the target from intra-pulse interferences, vertical transmit spatial frequency compensation (VTSFC) is applied utilizing the initial range of the range bin containing target. Afterwards, an adaptive beamforming (ABF) approach based on transmit-receive spatial smoothing (TRSS) is applied to suppress intra-pulse interference while avoiding target cancellation. Furthermore, the frequency increments in both rows and columns of PFDA are elaborately designed to effectively distinguish between true target and interferences. Extensive simulations are conducted to evaluate the excellent suppression effect of the proposed scheme against mainlobe deceptive interference. © 2025 IEEE.

关键词： Frequency domain analysis

来源：评论

学校读者我要写书评

暂无评论

Hybrid signal fusion for target detection in distributed PA-MIMO radar systems on moving platforms

引用

IEEE Transactions on Aerospace and Electronic Systems 2025年

作者： Zhou, Dingsen Yang, Minglei Lian, Hao Greco, Maria Sabrina Gini, Fulvio Xidian University National Key Laboratory of Radar Signal Processing Shaanxi Xi'an710071 China Science and Technology on Underwater Test and Control Laboratory Dalian116013 China University of Pisa Department of Information Engineering Pisa56122 Italy

Distributed phased multiple-input multiple-output (PA-MIMO) radar systems on moving platforms combine the spatial diversity of MIMO radars with the coherent gain of phased-array radars, making them ideal for dynamic environments. However, challenges such as motion-induced channel variability, Doppler shifts, and complex parameter estimation often hinder target detection. Traditional methods like non-coherent fusion (NCF) and coherent fusion (CF) struggle to adapt to such variability, leading to suboptimal performance. In this paper, we propose a hybrid signal fusion (HSF) scheme that integrates joint space-time-phase compensation (JSTPC) with a compressed fusion network. The JSTPC module mitigates motion-induced phase errors and ensures precise alignment of received signals, while the compressed fusion network combines CF and NCF stages, leveraging their respective strengths to adaptively manage channel correlation variability. We also develop a joint parameter estimation algorithm to accurately extract direction of departure, direction of arrival, and Doppler frequency, thereby ensuring precise signal reconstruction and enhancing the detection process by estimating key parameters in real time. Theoretical analyses of detection thresholds and probabilities, derived using moment-matching techniques, further validate the reliability of the proposed scheme. Monte Carlo simulations and theoretical analyses for diverse radar configurations and channel correlation scenarios show that the proposed HSF scheme achieves 2 dB detection gain improvement over NCF. © 2025 IEEE.

关键词： Frequency estimation

来源：评论

学校读者我要写书评

暂无评论

Complex High-Resolution Range Profile-Based Discrimination Method for Active Deception Targets Within Multistatic radar System

引用

IEEE Sensors Journal 2025年第10期25卷 18001-18014页

作者： Li, Qiang Guo, Yumei Xu, Hong Zhang, Linrang Chen, Zhanye Xidian University Hangzhou Institute of Technology Hangzhou311200 China Hangzhou Dianzi University School of Electronics and Information Engineering Hangzhou310018 China Xidian University School of Electronic Engineering Xi’an710071 China Xidian University National Laboratory of Radar Signal Processing Xi’an710071 China Southeast University State Key Laboratory of Millimeter Waves Nanjing210096 China Southeast University Institute of Electromagnetic Space Nanjing210096 China Chongqing University School of Microelectronics and Communication Engineering Chongqing400044 China

The available electronic counter-countermeasures methods for active deception jamming mainly focus on narrow-band processing in multistatic radar system. Compared with limited information in narrow-band case, the complex high-resolution range profile (C-HRRP), the manifestation of the spatial scattering characteristic in wideband case, contains more information about the structure and size of the target on the radar line of sight (LOS). The sensitivity of C-HRRP to target-aspect makes the C-HRRPs of physical target (PT) decorrelated in two distributed receivers with aspect variations. However, the C-HRRPs of the same active false target (FT) are highly consistent for which the jamming signals are from same source in the distributed receivers. Utilizing the difference, the correlation coefficient is defined first as the maximal sliding Hermitian distance between two C-HRRPs, and an innovative signal fusion approach is presented for discriminating PTs from active false ones through correlation testing for wideband multistatic radar system. A constant misjudgment probability can be achieved for PTs by the proposed discriminator. Due to the unchanged high correlation of jamming signals in different receivers, the presented approach can be applied into countering active deception jamming with arbitrary modulation. Simulations and corresponding result analyses are performed to verify the proposed approach’s feasibility and effectiveness. © 2001-2012 IEEE.

关键词： Multistatic radars

来源：评论

学校读者我要写书评

暂无评论

Leveraging Peripheral Visual Stimuli for Enhanced SSVEP-Based BCIs in Fast Calibration Scenario

引用

IEEE Sensors Journal 2025年第10期25卷 17683-17695页

作者： He, Xinjie Allison, Brendan Z. Liang, Wei Chen, Weijie Wang, Xingyu Cichocki, Andrzej Jin, Jing East China University of Science and Technology Key Laboratory of Smart Manufacturing in Energy Chemical Process Ministry of Education Shanghai200237 China University of California San Diego Department of Cognitive Science La Jolla CA92093 United States Nicolaus Copernicus University Department of Informatics Toruń87-100 Poland Systems Research Institute Polish Academy of Science Warsaw01-447 Poland RIKEN Brain Science Institute Laboratory for Advanced Brain Signal Processing Wako351-010 Japan East China University of Science and Technology Key Laboratory of Smart Manufacturing in Energy Chemical Process Ministry of Education School of Mathematics Shanghai200237 China

The decoding method based on training sample is utilized to effectively improve the recognition ability for steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs). However, these methods are not adapted to fast calibration scenarios because their performance degrades as the number of calibration trials decreases. For fast calibration scenarios, we proposed a multivisual source aliasing matrix estimation task-related component analysis (MVSAME-TRCA) method, which can simultaneously exploit the common information of the source aliasing matrix and the spatial filter across a peripheral visual stimulus (PVS). Experiment results show that the MVSAME-TRCA method outperforms the state-of-the art (SOTA) method such as ensemble TRCA approach (eTRCA), ePRCA, and same-eTRCA in a fast calibration scenario in the two public datasets. Notably, our approach only needs 28 s to calibrate 40 targets in the two public datasets and achieves an averaged information transfer rates (ITRs) of 248.78 ± 86.47 bits/min and 186.86 ± 100.32 bits/min, respectively. This research greatly reduces the need for individualized calibration for SSVEP system and is conducive to the development of practical plug and play SSVEP-BCI. © 2001-2012 IEEE.

关键词： Brain computer interface

来源：评论

学校读者我要写书评

暂无评论

Stable Extended U-Net for Noise-Robust Speaker Verification

Stable Extended U-Net for Noise-Robust Speaker Verification

引用

2025 IEEE International Conference on Acoustics, Speech, and signal processing, ICASSP 2025

作者： Wang, Zonghui Fang, Zhihua He, Liang School of Computer Science and Technology Xinjiang University Urumqi830017 China School of Intelligence Science and Technology Xinjiang University Urumqi830017 China Xinjiang Key Laboratory of Signal Detection and Processing Urumqi830017 China Department of Electronic Engineering Tsinghua University Beijing100084 China Beijing National Research Center for Information Science and Technology Tsinghua University Beijing100084 China

ISBN: (纸本)9798350368741

With advancements in deep learning, speaker verification systems have significantly improved their performance in noisy environments. Researchers typically demonstrate the effectiveness of their improved models by comparing performance on specific datasets, such as the VoxCeleb benchmark. However, in diverse real-world noise conditions, the out-of-domain generalization ability is also a crucial factor in evaluating a model's performance improvement. Research on stable learning indicates that eliminating the spurious correlation between training and testing data can enhance the generalization of the model. Building on this idea, we propose an improved speaker verification system with high generalization based on the extended U-Net (ExU-Net). It uses the sample reweighting method from stable learning to eliminate sample correlations and retains more effective speaker information through sub-pixel convolutions and coordinate attention mechanisms. We validate the effectiveness of this approach through extensive evaluations on VoxCeleb1, VOiCES, and other out-of-domain noise test sets, highlighting its generalization capability and model robustness. © 2025 IEEE.

关键词： noisy environments out-of-domain generalization Speaker verification stable learning

来源：评论

学校读者我要写书评

暂无评论

Improved SOCP Relaxation for SRI-Unknown Emitter Localization Using a Moving Receiver

引用

IEEE signal processing Letters 2025年 32卷 2199-2203页

作者： Liu, Sining Yue, Yaxing Zhao, Jiayi Shi, Zhiguo Liao, Guisheng Xidian University National Key Laboratory of Radar Signal Processing Xi'an710071 China Zhejiang University College of Information Science and Electronic Engineering Hangzhou310027 China

Employing a single moving receiver for emitter localization offers numerous advantages, including low cost, ease of implementation, and elimination of clock synchronization. A novel localization method is proposed for a stationary emitter with an unknown signal Repetition Interval (SRI). We first construct a localization model based on time-of-arrival, taking into account potential missed detections, and then derive the Cramér-Rao Lower Bound. Given that the established model is non-convex, we reformulate the problem into a convex form and convert it to Second-order Cone Programming (SOCP) format. However, the SOCP formulation encounters convex hull issues, which may hinder far-field localization. To address this challenge, we introduce a penalty term to alleviate the convex hull problem for the first time. Simulation results demonstrate the effectiveness of the proposed approach in comparison to existing methods, and the validity of the penalty term is also confirmed. © 1994-2012 IEEE.

关键词： Emitter localization second-order cone programming (SOCP) single moving receiver time-of-arrival (TOA)

来源：评论

学校读者我要写书评

暂无评论

Short-term wind power prediction with a new PCC-GWO-VMD and BiGRU hybrid model enhanced by attention mechanism

引用

Journal of Renewable and Sustainable Energy 2025年第3期17卷

作者： Zhang, Xiaoyu Qin, Xizhong Qin, Jiwei Wang, Ben Yuan, Yubo School of Computer Science and Technology Xinjiang University Urumqi830046 China Xinjiang Key Laboratory of Signal Detection and Processing Urumqi830046 China

The integration of wind power into the grid significantly depends on the accuracy and reliability of wind power prediction. However, the task of wind power prediction faces significant challenges due to the stochastic nature of wind speed. This study proposes a novel deep hybrid short-term wind power prediction model: PCC-GWO-VMD-BiGRU-Attention. The model integrates Pearson correlation coefficient (PCC), variational mode decomposition (VMD), gray wolf optimization (GWO), and attention mechanism optimized bidirectional gated recurrent unit (BiGRU-Attention) to enhance prediction accuracy and robustness, while quantifying prediction uncertainty through deep ensemble methods. First, PCC is used to identify the key factors affecting wind power, thereby improving the model's computational performance. Second, GWO can dynamically adjust the key parameters [K, α ] in VMD for optimal decomposition of the input time series, thereby improving prediction accuracy. Finally, the BiGRU-Attention model is utilized to extract global temporal features in historical sequences, while the attention mechanism focuses on key information in the sequences to further enhance prediction performance. Experimental results show that, compared to other deep learning models, this model achieves the highest accuracy in short-term wind power prediction with an root mean square error of 0.2677 Megawatt (MW), mean absolute error of 0.1509 Megawatt (MW), mean square error of 0.0717 Megawatt (MW), and coefficient of determination (R2) of 0.9605. This method has significant practical value and contributes to ensuring safe operation of wind farms. © 2025 Author(s).

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A Two-Stage Iteratively Reweighted Least Squares Estimator for Robust Elliptic Localization in the Presence of Impulsive Noise

引用

IEEE Transactions on Aerospace and Electronic Systems 2025年

作者： Zhao, Xiaoyu Lu, Zukun Li, Jun Huang, Long Yu, Yi Ou, Gang The College of Electronic Science and Technology National University of Defense Technology Changsha410073 China The Key Laboratory of Satellite Navigation Technology Changsha410073 China The National Key Laboratory of Radar Signal Processing Xidian University Xi’an710071 China The Hunan Xinghe Electronics Co.LTD Changsha410073 China

As a range-based localization approach, elliptic localization based on bistatic ranges (BRs) has wide range of applications in multistatic systems. This paper addresses the problem of robust elliptic localization using the minimum `pnorm criterion in scenarios affected by impulsive α-stable noise. To solve the `p-norm formulation, we develop an asymptotically efficient two-stage iteratively reweighted least squares (TSIRLS) estimator via incorporating pseudolinear estimation and multistage weighted least squares estimation into the framework of iteratively reweighted least squares (IRLS) methods. The estimator consists of two cascaded `p-norm minimization estimators. To obtain a closed-form solution at each iteration, we transform nonlinear BR measurement equations into pseudolinear forms through introducing auxiliary variables, thus resulting in an iteratively reweighted pseudolinear least squares (IRPLS) estimator in the first stage. The dependencies of unknown parameters and the estimate error derived from the IRPLS are utilized in the second stage to enhance the performance of target location estimate. The analytical derivation demonstrates that, under the assumptions of a sufficiently large number of measurements and small noise, the final position estimate obtained from the TSIRLS achieves the theoretical covariance of the general `pnorm minimization estimation. Extensive numerical simulations highlight the advantages of TSIRLS over existing least squares and robust estimators in terms of positioning performance and runtime. The TSIRLS is also observed to generate approximately unbiased estimates with mean square errors that closely approach the Cramér-Rao lower bound. © 2025 IEEE.

关键词： Impulse noise

来源：评论

学校读者我要写书评

暂无评论

Duplex: Dual Prototype Learning for Compositional Zero-Shot Learning

arXiv

引用

arXiv 2025年

作者： Peng, Zhong Xu, Yishi Wang, Gerong Chen, Wenchao Chen, Bo Zhang, Jing National Key Laboratory of Radar Signal Processing Xidian University Xi’an710071 China Research Institute of Systems Engineering Academy Military Science Beijing100089 China

Compositional Zero-Shot Learning (CZSL) enables models to recognize novel compositions of visual states and objects that not emerge during training. Existing methods struggle with generalization due to the complex entanglement of states and objects in visual representations, limiting their ability to handle unseen compositions. To address this challenge, we propose Duplex (Dual Prototype Learning), a novel approach that employs a well-designed dual-branch architecture to integrate both semantic and visual prototypes, thereby facilitating efficient composition learning. Duplex utilizes a graph neural network (GNN) to adaptively update visual prototypes, capturing intricate interactions between states and objects. It then leverages the strong visual-semantic alignment of pre-trained vision-language models (VLMs) and incorporates a multi-path architecture with prompt engineering to align image and text representations, ensuring robust generalization. Extensive experiments on three benchmark datasets—MIT-States, UT-Zappos, and C-GQA—demonstrate that Duplex outperforms the state of the art across all datasets in both closed-world and open-world settings. Copyright © 2025, The Authors. All rights reserved.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：