检索结果-内蒙古大学图书馆

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Qiao, Xiaoyu Li, Weisheng Xiao, Bin Huang, Yuping Yang, Lijian Chongqing Key Laboratory of Image Cognition Chongqing University of Posts and Telecommunications Chongqing China Key Laboratory of Cyberspace Big Data Intelligent Security Chongqing University of Posts and Telecommunications Ministry of Education Chongqing China

ISBN: (纸本)9798350368741

Model-based networks have shown convincing performance in MRI reconstruction. However, the unrolled cascades within the networks are constrained to solely obtain information from the preceding counterpart, resulting in potential error accumulation. Moreover, the linear structure fails to address the challenge of recovering fine-grained details. To tackle these problems, we propose to redesign the information flow in model-based networks. Our method features a large U-shaped network, where the nodes are built with unrolled cascades and U-Net-based regularizers. We design an input-level integration module to help the cascades acquire information from adjacent and skip-connected counterparts, building robust mappings to the target. We further design a coarse-to-fine feature-level integration module, aiming at guiding the network to progressively recover fine details. Intermediate reconstructions produced by subnetworks of different scales are integrated, enabling the extraction of complementary information to enhance the final performance. Compared with cutting-edge methods on different datasets, our method exhibits superior performances. © 2025 IEEE.

关键词： compressed sensing deep learning model-based MRI reconstruction U-Net

来源：评论

学校读者我要写书评

暂无评论

Subsampling Decomposition based k-Space Refinement for Accelerated MRI Reconstruction

Subsampling Decomposition based k-Space Refinement for Accel...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

ISBN: (纸本)9798350368741

In accelerated MRI reconstruction problem, directly recovering all the missing k-space data from undersampled measurements is highly ill-posed and often leads to suboptimal performance. To address the problem, we propose a novel deep unfolding network (DUN) with subsampling decomposition (SD) based k-space refinement to mitigate the ill-posedness. Our method employs a parallel network architecture with a primary branch unfolded by gradient descent-inspired optimization process (GD-PB) for reconstruction. Additionally, we introduce an SD-based auxiliary branch (SD-AB) that decompose the inverse problem into moderately corrupted subproblems. We design a novel subsampling mask predictor that captures both global and local spatial correlations in k-space, ensuring the SD-AB effectively preserves the most well-reconstructed subsets as reliable region (RR). The RR in SD-AB is used to periodically refine the intermediate outputs of the GD-PB, achieving improved accuracy. Experimental results reveal that our method significantly outperforms conventional and SD-based DUN techniques, achieving superior PSNR and SSIM results compared with cutting-edge methods. © 2025 IEEE.

关键词： compressed sensing deep unfolded network MRI reconstruction parallel imaging subsampling decomposition

来源：评论

学校读者我要写书评

暂无评论

Learning Preconditioners in Gates-controlled Deep Unfolding Networks based on Quasi-Newton Methods For Accelerated MRI Reconstruction

Learning Preconditioners in Gates-controlled Deep Unfolding ...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

ISBN: (纸本)9798350368741

Deep unfolding networks (DUNs) have made significant progress in MRI reconstruction, successfully tackling the problem of prolonged imaging time. However, the ill-conditioned nature of MRI reconstruction often causes slow convergence in iterative optimization, potentially compromising the performance of DUNs. In this study we propose a preconditioned and gates-controlled DUN (PGDUN) to address these challenges. Our approach starts with optimizing the step size of proximal gradient descent (PGD) through a preconditioner. To improve flexibility and adaptability, we relax the constrains on quasi-newton-based optimization procedure. We design ConvLSTM-based modules, where the gate units automatically preserve necessary long- and short-term information, facilitating the learning of optimized variables and their combinations. Furthermore, we design gate units to modulate the features fed to regularizers across different iterations, boosting their robustness against potential accumulated errors. Evaluations using PSNR and SSIM metrics reveal that our approach outperforms existing state-of-the-art methods, achieving superior reconstruction results across various sequences. © 2025 IEEE.

关键词： Compressed Sensing LSTM MRI reconstruction Preconditioner Quasi-Newton

来源：评论

学校读者我要写书评

暂无评论

Virtual Petrochemical FireFighting Training system Design and Implementation 14th

Virtual Petrochemical FireFighting Training System Design an...

引用

14th International Conference on Frontier Computing, FC 2024

作者： Li, Hengdong Ye, Jun Zhang, Dian Trusted and Intelligent Data Analytic Research Group School of Cyberspace Security Hainan University Haikou China Key Laboratory of Internet Information Retrieval of Hainan Province Hainan University Haikou China

ISBN: (纸本)9789819627974

Due to the special and dangerous nature of flammable and explosive petrochemical substances, petrochemical storage and transportation involves great risks and is highly prone to fire and explosion and other accidents, which poses a great challenge to firefighting and rescue. Due to the special characteristics of petrochemical fire accidents, traditional fire training and experience often cannot cope with them, resulting in low firefighting efficiency and heavy economic losses. With the wide application of virtual reality technology in the training field, this study summarises the current research status in the field of virtual reality firefighting, optimises the LOD technology, and develops a petrochemical firefighting training system based on virtual reality, and the research results effectively address the current shortcomings in petrochemical firefighting training. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Fire extinguishers

来源：评论

学校读者我要写书评

暂无评论

Generating Is Believing: Membership Inference Attacks against Retrieval-Augmented Generation

Generating Is Believing: Membership Inference Attacks agains...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Li, Yuying Liu, Gaoyang Wang, Chen Yang, Yang Key Laboratory of Intelligent Sensing System and Security Ministry of Education School of Artificial Intelligence Hubei University Wuhan China Hubei Key Laboratory of Smart Internet Technology Huazhong University of Science and Technology Wuhan China

ISBN: (纸本)9798350368741

Retrieval-Augmented Generation (RAG) is a state-of-the-art technique that mitigates issues such as hallucinations and knowledge staleness in Large Language Models (LLMs) by retrieving relevant knowledge from an external database to assist in content generation. Existing research has demonstrated potential privacy risks associated with the LLMs of RAG. However, the privacy risks posed by the integration of an external database, which often contains sensitive data such as medical records or personal identities, have remained largely unexplored. In this paper, we aim to bridge this gap by focusing on membership privacy of RAG's external database, with the aim of determining whether a given sample is part of the RAG's database. Our basic idea is that if a sample is in the external database, it will exhibit a high degree of semantic similarity to the text generated by the RAG system. We present S2MIA, a Membership Inference Attack that utilizes the Semantic Similarity between a given sample and the content generated by the RAG system. With our proposed S2MIA, we demonstrate the potential to breach the membership privacy of the RAG database. Extensive experimental results demonstrate that S2MIA outperforms five existing MIAs, even when the system is protected by three representative defenses. © 2025 IEEE.

关键词： Large Language Models Membership Inference Attacks Retrieval-Augmented Generation

来源：评论

学校读者我要写书评

暂无评论

Thermal engineering in ALD-grown ZGO thin films for high-performance photodetectors

引用

材料科学技术（英文版） 2025年第6期209卷 19-26页

作者： Si-Tong Ding Yu-Chang Chen Cai-Yu Shi Lei Shen Qiu-Jun Yu Lang-Xi Ou Ze-Yu Gu Na Chen Ting-Yun Wang David Wei Zhang Hong-Liang Lu State Key Laboratory of ASIC and System Shanghai Institute of Intelligent Electronics & SystemsSchool of MicroelectronicsFudan UniversityShanghai 200433China Zhangjiang Fudan International Innovation Center Shanghai 201203China National Integrated Circuit Innovation Center Shanghai 201203China Key Laboratory of Specialty Fiber Optics and Optical Access Networks Shanghai Institute Communication and Data ScienceShanghai UniversityShanghai 200444China State Key Laboratory of ASIC and System Shanghai Institute of Intelligent Electronics & SystemsSchool of MicroelectronicsFudan UniversityShanghai 200433China Zhangjiang Fudan International Innovation Center Shanghai 201203China National Integrated Circuit Innovation Center Shanghai 201203China Jiashan Fudan Institute Jiaxing 314100China

Doped gallium oxide-based thin films are a class of wide-band semiconductor materials with the ad-vantages of chemically stable,tunable bandgap,and offer the benefit of ultraviolet *** order to obtain photodetectors(PDs)with superior response,higher demands are placed on the quality of growth and processing of doped *** this work,Zn-doped ternary metal oxide ZnGaO(ZGO)thin films were grown using the atomic layer deposition technique and annealed at different temperatures under an oxy-gen *** results showed that the high-quality ZGO films with good uniformity,high visible light transmittance,low roughness,and significant reduction of oxygen vacancies were obtained after ***,metal-semiconductor-metal PDs were prepared based on the studied ZGO *** responsivity(R),detectivity(D*),and external quantum efficiency(EQE)of the optimized device are 61.8 A W-1,1.2 x 1012 Jones,and 255.9％,*** to the unannealed device,the an-nealed ZGO PD achieves a maximum 309-fold increase in *** thermal engineering work may provide a strong reference for the development of low-cost,large-area,high-performance ultraviolet *** it also broadens the application of ternary metal oxides in optoelectronics.

关键词： Atomic layer deposition Oxygen annealing ZGO thin films Photodetector Ultraviolet detection

来源：评论

学校读者我要写书评

暂无评论

POINTERGUESS: targeted password guessing model using pointer mechanism 24

POINTERGUESS: targeted password guessing model using pointer...

引用

Proceedings of the 33rd USENIX Conference on security Symposium

作者： Kedong Xiu Ding Wang College of Cyber Science Nankai University Tianjin China and Key Laboratory of Data and Intelligent System Security (NKU) Ministry of Education Tianjin China and Tianjin Key Laboratory of Network and Data Security Technology Nankai University Tianjin China

ISBN: (纸本)9781939133441

Most existing targeted password guessing models view users' reuse behaviors as sequences of edit operations (e.g., insert and delete) performed on old passwords. These atomic edit operations are limited to modifying one character at a time and cannot fully cover users' complex password modification behaviors (e.g., modifying the password structure). This partially leads to a significant gap between the proportion of users' reused passwords and the success rates that existing targeted password models can achieve. To fill this gap, this paper models users' reuse behaviors by focusing on two key components: (1) What they want to copy/keep; (2) What they want to tweak. More specifically, we introduce the pointer mechanism and propose a new targeted guessing model, namely POINTERGUESS. By hierarchically redefining password reuse from both personal and population-wide perspectives, we can accurately and comprehensively characterize users' password reuse behaviors. Moreover, we propose MS-POINTERGUESS, which can employ the victim's multiple leaked *** employing 13 large-scale real-world password datasets, we demonstrate that POINTERGUESS is effective: (1) When the victim's password at site A (namely pwA) is known, within 100 guesses, the average success rate of POINTERGUESS in guessing her password at site B (namely pwB, pwA ≠ pwB) is 25.21% (for common users) and 12.34% (for security-savvy users), respectively, which is 21.23%~71.54% (38.37% on average) higher than its foremost counterparts; (2) When not excluding identical password pairs (i.e., pwA can equal pwB), within 100 guesses, the average success rate of POINTERGUESS is 48.30% (for common users) and 28.42% (for security-savvy users), respectively, which is 6.31%~15.92% higher than its foremost counterparts; (3) Within 100 guesses, the MS-POINTERGUESS further improves the cracking success rate by 31.21% compared to POINTERGUESS.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DSDIR: A Two-Stage Method for Addressing Noisy Long-Tailed Problems in Malicious Traffic Detection

DSDIR: A Two-Stage Method for Addressing Noisy Long-Tailed P...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Guoliang Li Ruiqi Zhang Zhe Sun Lingkai Xing Yu Zhang College of Cyber Science Nankai University China Tianjin Key Laboratory of Network and Data Security Technology China Key Laboratory of Data and Intelligent System Security Ministry of Education China College of Computer Science Nankai University China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

In recent years, deep learning based malicious traffic detection (MTD) systems have demonstrated remarkable success. However, their effectiveness tend to decrease because most malicious traffic datasets are suffered from noisy-labeled and long-tailed problems. While numerous approaches have been developed to address these two problems individually, they become inefficient when confronting the combined challenge of noisy long-tailed data, as they typically tackle only a single adverse factor at a time. This paper proposes a two-stage method called Distribution-aware sample Selection and Dynamic Instance-based Relabeling (DSDIR), which simultaneously addresses the impacts of noisy-labeled and long-tailed problems. In the first stage, a noise-independent clean sample selection method is designed to obtain a clean dataset, which converts negative effects of the long-tailed problem into positive ones. In the second stage, dynamic instance-based relabeling is designed to train a model and improve the dataset’s quality simultaneously. Eventually, DSDIR not only produces a balanced and noise-tolerant model but also obtains a clean dataset. Experimental results demonstrate that in the high noise condition of 60% and 80%, the accuracy rate of DSDIR is 5% higher than the state-of-the-art methods. Our code is available at https://***/nku-ligl/DSDIR.

关键词： Deep learning Computer vision Heavily-tailed distribution Codes Accuracy Noise Speech recognition Acoustics Noise measurement Speech processing

来源：评论

学校读者我要写书评

暂无评论

Efficient Cross-modal Prompt Learning with Semantic Enhancement for Domain-robust Fake News Detection 31

Efficient Cross-modal Prompt Learning with Semantic Enhancem...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Wu, Fei Jin, Hao Hu, Changhui Ji, Yimu Jing, Xiao-Yuan Jiang, Guo-Ping College of Automation College of Artificial Intelligence Nanjing University of Posts and Telecommunications Nanjing China The State Key Laboratory of Blockchain and Data Security Zhejiang University Hangzhou China Guangdong Provincial Key Laboratory of Petrochemical Equipment Intelligent Security Guangdong University of Petrochemical Technology Maoming China School of Computer Wuhan University Wuhan China

ISBN: (纸本)9798891761964

With the development of multimedia technology, online social media has become a major medium for people to access news, but meanwhile, it has also exacerbated the dissemination of multi-modal fake news. An automatic and efficient multi-modal fake news detection (MFND) method is urgently needed. Existing MFND methods usually conduct cross-modal information interaction at later stage, resulting in insufficient exploration of complementary information between modalities. Another challenge lies in the differences among news data from different domains, leading to the weak generalization ability in detecting news from various domains. In this work, we propose an efficient Cross-modal Prompt Learning with Semantic enhancement method for Domain-robust fake news detection (CPLSD). Specifically, we design an efficient cross-modal prompt interaction module, which utilizes prompt as medium to realize lightweight cross-modal information interaction in the early stage of feature extraction, enabling to exploit rich modality complementary information. We design a domain-general prompt generation module that can adaptively blend domain-specific news features to generate domain-general prompts, for improving the domain generalization ability of the model. Furthermore, an image semantic enhancement module is designed to achieve image-to-text translation, fully exploring the semantic discriminative information of the image modality. Extensive experiments conducted on three MFND benchmarks demonstrate the superiority of our proposed approach over existing state-of-the-art MFND methods. © 2025 Association for Computational Linguistics.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

IMPRESS: an importance-informed multi-tier prefix KV storage system for large language model inference 25

IMPRESS: an importance-informed multi-tier prefix KV storage...

引用

Proceedings of the 23rd USENIX Conference on File and Storage Technologies

作者： Weijian Chen Shuibing He Haoyang Qu Ruidong Zhang Siling Yang Ping Chen Yi Zheng Baoxing Huai Gang Chen The State Key Laboratory of Blockchain and Data Security Zhejiang University and Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security and Zhejiang Key Laboratory of Big Data Intelligent Computing The State Key Laboratory of Blockchain and Data Security Zhejiang University Huawei Cloud

ISBN: (纸本)9781939133458

Modern advanced large language model (LLM) applications often prepend long contexts before user queries to improve model output quality. These contexts frequently repeat, either partially or fully, across multiple queries. Existing systems typically store and reuse the keys and values of these contexts (referred to as prefix KVs) to reduce redundant computation and time to first token (TTFT). When prefix KVs need to be stored on disks due to insufficient CPU memory, reusing them does not always reduce TTFT, as disk I/O latency is high. In this paper, we propose IMPRESS, an importance-informed multi-tier prefix KV storage system to reduce I/O delay for LLM inference by only loading important prefix KVs. IMPRESS first leverages the insight that there is significant similarity in important token index sets across attention heads and introduces an I/O-efficient important KV identification algorithm. It then optimizes prefix KV storage and caching through importance-informed KV management, reducing TTFT during model inference. Our experimental results show that IMPRESS can reduce TTFT by up to 2.8× compared to state-of-the-art systems, while maintaining comparable inference accuracy.

关键词：

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：