Chinese Journal of Electronics最新文献_第9页

Priority-Driven Flow Scheduling for Distributed Large Language Model Training 分布式大型语言模型训练的优先级驱动流调度

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-07-03 DOI: 10.23919/cje.2025.00.349

Tianshi Wang;Yiran Zhang;Yiao Zhang;Qiyang Zhang;Ao Zhou;Shangguang Wang;Jihong Yu

引用次数: 0

InterKV: Tail Latency Reduction for LSM-Trees in Hybrid Storage Systems 混合存储系统中lsm树的尾部延迟降低

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-07-03 DOI: 10.23919/cje.2025.00.137

Shiqiang Nie;Menghan Li;Jianqiang Ma;Min Fang;Ping She;Weiguo Wu;Meng Ma

{"title":"InterKV: Tail Latency Reduction for LSM-Trees in Hybrid Storage Systems","authors":"Shiqiang Nie;Menghan Li;Jianqiang Ma;Min Fang;Ping She;Weiguo Wu;Meng Ma","doi":"10.23919/cje.2025.00.137","DOIUrl":"https://doi.org/10.23919/cje.2025.00.137","url":null,"abstract":"The emergence of non-volatile memory (NVM) has introduced new opportunities to enhance the performance of key-value storage systems. However, the high cost of NVM makes its exclusive adoption economically unfeasible. To address this issue, previous research has explored log-structured merge (LSM)-tree split architectures, aiming to achieve cost-effective performance improvements. Nevertheless, these studies often assign the lower levels of the LSM-tree to NVM and the higher levels to low-performance storage devices, which leads to significant compaction overheads at the higher levels and increased tail latency. In this paper, we propose an inter-leaving level placement scheme that maximizes device bandwidth utilization during compaction, thereby enhancing both read and write efficiency. Furthermore, we introduce a speculative read scheme designed specifically for this architecture, which reduces seek waiting overhead and improves random read throughput. Experimental results show that our scheme achieves up to 48.2% performance improvement over state-of-the-art systems.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 3","pages":"1327-1338"},"PeriodicalIF":4.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11595438","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148443331","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Efficient Revocable Attribute-Based Encryption with Intel SGX 基于Intel SGX的高效可撤销属性加密

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-07-03 DOI: 10.23919/cje.2025.00.287

Chenyang Zhao;Ying Gao;Xinrui Yang;Shuqing Zhang;Fengwei Zhang

{"title":"Efficient Revocable Attribute-Based Encryption with Intel SGX","authors":"Chenyang Zhao;Ying Gao;Xinrui Yang;Shuqing Zhang;Fengwei Zhang","doi":"10.23919/cje.2025.00.287","DOIUrl":"https://doi.org/10.23919/cje.2025.00.287","url":null,"abstract":"Cloud storage has transformed data sharing for individuals and organizations, but it also raises significant privacy concerns. Attribute-based encryption (ABE) enables fine-grained access control over encrypted data on untrusted servers, offering advantages over traditional symmetric or public-key encryption. However, with the advent of quantum computing, classical bilinear pairing-based ABE schemes face long-term security risks. Lattice-based ABE emerges as a promising post-quantum alternative, yet existing constructions suffer from large key/ciphertext sizes, inefficient Gaussian preimage sampling, and limited revocation support. To address these challenges, we propose trusted environment revocable ABE (TR-ABE), a lattice-based revocable ciphertext-policy ABE scheme explicitly designed for Intel Software Guard Extensions (SGX). By leveraging the SGX enclave, Gaussian SamplePre operations are securely outsourced while preserving confidentiality against side-channel leakage under the transparent trusted execution environment model. TR-ABE further adopts an attribute-aggregated key structure to enable efficient, resampling-free key updates, and supports efficient revocation with cryptographic binding of user identities to resist collusion. A cloud trusted execution environment framework ensures secure ciphertext updates with enclave integrity guarantees. We formally prove indistinguishability under chosen-plain-text attack security and collusion resistance under the ring learning with errors assumption in the standard model. Both theoretical analysis and experimental evaluation demonstrate significant improvements in revocation efficiency and up to 7.3× performance improvement in SamplePre through secure outsourcing in Intel SGX.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 3","pages":"1224-1245"},"PeriodicalIF":4.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11595436","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148443332","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deconstruct Evaluate and Targeted Intervention Based LLMs Unlearning Algorithm 基于解构评估和目标干预的LLMs学习算法

IF 3 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-06-08 DOI: 10.23919/cje.2025.00.105

Yang Wang;Ke Miao;Yuke Hu;Xiaochen Li;Zhan Qin

{"title":"Deconstruct Evaluate and Targeted Intervention Based LLMs Unlearning Algorithm","authors":"Yang Wang;Ke Miao;Yuke Hu;Xiaochen Li;Zhan Qin","doi":"10.23919/cje.2025.00.105","DOIUrl":"https://doi.org/10.23919/cje.2025.00.105","url":null,"abstract":"With the acceleration of digitalization, the widespread application of large language models (LLMs) in many fields has made data privacy and compliance issues increasingly prominent. Balancing the protection of sensitive information, adherence to regulations, and the maintenance of model performance and utility has become a pressing challenge. Machine unlearning technology is a key solution to this problem. How to maintain the model's ability to parse entities and relationships while erasing target knowledge has become an urgent technical bottleneck that needs to be overcome. This paper delves into LLM unlearning, addressing the shortcomings of current evaluation systems and algorithms. We propose a novel framework that integrates deconstruction evaluation with targeted intervention. By parsing discrete knowledge points into (subject <tex>$s$</tex>, relation <tex>$r$</tex>, object <tex>$o$</tex>) triples, we construct a dual dimensional evaluation system, the knowledge deconstruction-based unlearning metric. This framework allows for more precise measurement of model performance during specific knowledge unlearning, supporting technical optimization. Our designed structured targeted unlearning algorithm (STUA) combines targeted token disruption, attention mask replacement, and finetuning. It precisely removes target knowledge while minimizing side effects. Experiments show the proposed evaluation framework accurately assesses specific knowledge unlearning, and STUA outperforms existing algorithms in general knowledge retention and target unlearning degree. Notably, when general knowledge retention is no less than that of retrained models, STUA's target unlearning degree surpasses existing state of the art methods. It improves by 26% on average with 1% unlearning data, 22% with 5% unlearning data, and 41% with 10% unlearning data.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 2","pages":"761-770"},"PeriodicalIF":3.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11554542","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148196132","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An Improved BP Neural Network Model Based on High-Dimensional Feature Selection Strategy and Maxwell's Equations Derived Optimization 基于高维特征选择策略和麦克斯韦方程衍生优化的改进BP神经网络模型

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-06-08 DOI: 10.23919/cje.2025.00.060

Ruizhe Yang;Lilin Li;Xin Zhang;Yunxin Su

{"title":"An Improved BP Neural Network Model Based on High-Dimensional Feature Selection Strategy and Maxwell's Equations Derived Optimization","authors":"Ruizhe Yang;Lilin Li;Xin Zhang;Yunxin Su","doi":"10.23919/cje.2025.00.060","DOIUrl":"https://doi.org/10.23919/cje.2025.00.060","url":null,"abstract":"As systems continue to expand in scale, there is a corresponding increase in the variety and number of electronic devices within these systems, which contributes to a heightened complexity in electromagnetic emissions. This complexity complicates the process of identifying electromagnetic emitters. To tackle this challenge, this paper introduces an improved back propagation (BP) neural network model based on high-dimensional feature selection strategy and Maxwell's equations derived optimization. The proposed model initially utilizes a feature selection approach grounded in random small samples to improve computational efficiency by streamlining the network architecture. Subsequently, an electromagnetic optimization algorithm, referred to as Maxwell's equations derived optimization, is applied to refine the parameters of the neural network, thereby enhancing the model's recognition capabilities. Ultimately, the innovative model is employed to identify a dataset of electromagnetic emission signals, which is constructed based on fundamental emission waveform theory. Experimental findings indicate this approach achieves reduced time complexity and improved recognition accuracy comparing to other existing BP neural network models.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 2","pages":"524-542"},"PeriodicalIF":4.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11554500","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148505384","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Towards Secure In-Vehicle Communication: Surveying CAN and Automotive Ethernet Threats, Standards, and Detection Techniques 迈向安全的车载通信：调查CAN和汽车以太网威胁、标准和检测技术

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-06-08 DOI: 10.23919/cje.2025.00.230

Sifan Li;Gaoda Wei;Boying Fang;Yue Cao;Chakkaphong Suthaputchakun;Faisal Bashir Hussain

{"title":"Towards Secure In-Vehicle Communication: Surveying CAN and Automotive Ethernet Threats, Standards, and Detection Techniques","authors":"Sifan Li;Gaoda Wei;Boying Fang;Yue Cao;Chakkaphong Suthaputchakun;Faisal Bashir Hussain","doi":"10.23919/cje.2025.00.230","DOIUrl":"https://doi.org/10.23919/cje.2025.00.230","url":null,"abstract":"As modern vehicles evolve into increasingly connected and autonomous systems, the security of invehicle networks has become a critical concern. The controller area network (CAN) bus and automotive Ethernet are the two dominant communication protocols used within vehicles, each with distinct data structures and associated threat models. This survey provides a comprehensive overview of these two networks, detailing their data characteristics, known vulnerabilities, and publicly available datasets. Furthermore, we summarize the global landscape of automotive cybersecurity standards and regulations, highlighting efforts to standardize security practices across the industry. Building upon this foundation, we systematically review existing intrusion detection system (IDS) for CAN, including voltage-based physical-layer methods, rule-based and AI-driven detection techniques, and emerging cloud-vehicle collaborative frameworks. Meanwhile, we also examine state-of-the-art intrusion detection strategies tailored for automotive Ethernet. A key innovation of this review lies in its integrated perspective that bridges physical-layer analysis, intelligent detection algorithms, and cloud-based architectures across both CAN and Ethernet domains. By aligning technological insights with regulatory trends, this work aims to inform future research and development of robust, real-world-deployable in-vehicle IDSs.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 2","pages":"625-650"},"PeriodicalIF":4.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11554557","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148506317","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

PowerMG: Early-Stage Dynamic Power Estimation in HLS FPGA via Multi-Grained GNN Information 基于多粒度GNN信息的HLS FPGA早期动态功率估计

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-07-03 DOI: 10.23919/cje.2025.00.021

Zebin Wu;Zhe Lin;Yonghong Tian

{"title":"PowerMG: Early-Stage Dynamic Power Estimation in HLS FPGA via Multi-Grained GNN Information","authors":"Zebin Wu;Zhe Lin;Yonghong Tian","doi":"10.23919/cje.2025.00.021","DOIUrl":"https://doi.org/10.23919/cje.2025.00.021","url":null,"abstract":"Accurate field-programmable gate array (FPGA) power estimation is inherently time-consuming, often requiring iterative evaluations within power-oriented optimization loops, significantly prolonging the design turnaround time. Existing graph neural network (GNN)-based power estimation approaches predominantly rely on node-centric GNN architectures and necessitate register transfer level (RTL) implementation flows, which are computationally expensive and time-intensive. To address these limitations, we propose PowerMG, a novel dynamic power estimation framework that leverages early-stage information from electronic design automation tools, specifically from the high-level synthesis (HLS) front-end and back-end, bypassing the need for RTL-level data. We just need to pass through HLS flow to get the graph data and then feed it into the power estimation model, skipping the time-consuming RTL implementation stage in power inference. A key innovation of PowerMG lies in that it makes use of multi-grained GNN information. The power estimation value of each layer of our GNN is built on different GNN receptive field. Additionally, we introduce a complementary aggregation mechanism that integrates both edge-centric and node-centric layers, mitigating neighbourhood information loss and enhancing the representational capacity of the model. This dual-aggregation strategy is empirically validated to be effective for power estimation tasks. Experimental results demonstrate that PowerMG achieves superior estimation accuracy and strong generalization. It outperforms state-of-the-art methods on 8 out of 9 subsets of the PolyBench benchmark suite and surpasses PowerGear by 4.74% relatively at single model mode, highlighting its effectiveness and transferability.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 3","pages":"961-969"},"PeriodicalIF":4.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11595422","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148506400","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The MC-CC Method Applied for Electromagnetic-Thermal Multi-Physics Statistical Analysis MC-CC方法在电磁-热多物理场统计分析中的应用

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-06-08 DOI: 10.23919/cje.2024.00.206

Xiaojin Tang;Zijing Xiao;Wenshuai Ding;Yingsong Li;Zhixiang Huang;Xianliang Wu;Kaikun Niu

引用次数: 0

Satellite Edge Intelligent Computing Network: A Resource Management Method Based on Multi-Agent Deep Reinforcement Learning 卫星边缘智能计算网络：基于多智能体深度强化学习的资源管理方法

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-07-03 DOI: 10.23919/cje.2025.00.119

Hangyu Zhang;Yuhong Huang;Wei Deng;Pingke Deng;Jianyin Zhang

{"title":"Satellite Edge Intelligent Computing Network: A Resource Management Method Based on Multi-Agent Deep Reinforcement Learning","authors":"Hangyu Zhang;Yuhong Huang;Wei Deng;Pingke Deng;Jianyin Zhang","doi":"10.23919/cje.2025.00.119","DOIUrl":"https://doi.org/10.23919/cje.2025.00.119","url":null,"abstract":"The integration of artificial intelligence with satellite edge computing (SEC) networks presents a transformative solution for time-sensitive emergency remote sensing applications. However, it encounters critical challenges in resource-constrained low Earth orbit (LEO) environments. This paper proposes a novel multi-agent deep reinforcement learning framework to address the distributed resource management problem in dynamic SEC networks. Facing urgent remote sensing processing tasks, we establish a fully cooperative satellite edge intelligent transmission and computing architecture, in which each LEO satellite acts independently as an autonomous agent based on local observations. In order to jointly optimize the task offloading ratio, i.e. the corresponding allocated communication and computing resources of distributed multi-satellites, we propose a multi-agent proximal policy optimization algorithm with an actor-critic and centralized training-distributed execution framework. This approach implements the adaptive policy update by the centralized critic and the on-board real-time decision-making by the distributed actors in stages, significantly improving the efficiency of strategy learning and execution. The simulation results demonstrate the proposed scheme effectively reduces task execution time and satellite energy consumption during multi-satellite collaborative transmission and computation, and outperforms the benchmark algorithms. This verifies that the solution is able to balance mission-critical quality of service requirements and sustainable resource utilization in emergency situations, and is expected to achieve space-edge intelligence for the next-generation satellite-ground integrated network.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 3","pages":"1029-1041"},"PeriodicalIF":4.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11595427","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148507052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Gridless Methods for Two-Dimensional Direction-of-Arrival Estimation in Single-Snapshot Scenario 单快照场景下二维到达方向估计的无网格方法

IF 4.5 4区计算机科学

Chinese Journal of Electronics Pub Date : 2026-03-01 Epub Date: 2026-07-03 DOI: 10.23919/cje.2025.00.080

Ninghui Li;Xiaokuan Zhang;Fan Lyu

{"title":"Gridless Methods for Two-Dimensional Direction-of-Arrival Estimation in Single-Snapshot Scenario","authors":"Ninghui Li;Xiaokuan Zhang;Fan Lyu","doi":"10.23919/cje.2025.00.080","DOIUrl":"https://doi.org/10.23919/cje.2025.00.080","url":null,"abstract":"Two-dimensional (2D) direction-of-arrival (DOA) estimation in single-snapshot super-resolution scenarios constitutes a pivotal challenge in radar detection for unmanned aerial vehicles. Existing algorithms can generally be categorized into two main types: Those based on atomic norm minimization (ANM) and those relying on enhanced matrix techniques. However, the ANM-based methods are constrained by the need for prior knowledge of noise power, while the enhanced matrix-based approaches suffer from a reduction in array aperture due to the reconstruction process of the enhanced matrix. To address these issues, we introduce a gridless version of the sparse iterative covariance-based estimation (SPICE) method, termed GLS, which estimates noise power in a vectorized atom form for 2D DOA estimation, leading to the development of a novel approach called V-GLS. Like ANM, V-GLS leverages semi-definite programming optimization, ensuring that the array aperture is not compromised. Additionally, we ingeniously reconfigure the 2D single-snapshot received signal vector into an approximated one-dimensional multiple-snapshot matrix, thereby decoupling the 2D DOAs into two separate one-dimensional DOAs. Moreover, we propose a decoupled version of GLS for 2D DOA estimation, referred to as D-GLS. Compared to V-GLS, D-GLS exhibits superior computational efficiency due to its low-dimensional semi-definite programming optimization. Both V-GLS and D-GLS exhibit robust performance in noisy environments and show promise for estimating multiple sources. Extensive simulations are provided to illustrate the superiority of the proposed method over the state-of-the-art methods.","PeriodicalId":50701,"journal":{"name":"Chinese Journal of Electronics","volume":"35 3","pages":"1130-1138"},"PeriodicalIF":4.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11595450","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148507888","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0