期刊文献+
共找到2,183篇文章
< 1 2 110 >
每页显示 20 50 100
Design of Protograph LDPC-Coded MIMO-VLC Systems with Generalized Spatial Modulation 被引量:4
1
作者 Dai Lin Fang Yi +1 位作者 Guan Yongliang Mohsen Guizani 《China Communications》 SCIE CSCD 2024年第3期118-136,共19页
This paper investigates the bit-interleaved coded generalized spatial modulation(BICGSM) with iterative decoding(BICGSM-ID) for multiple-input multiple-output(MIMO) visible light communications(VLC). In the BICGSM-ID ... This paper investigates the bit-interleaved coded generalized spatial modulation(BICGSM) with iterative decoding(BICGSM-ID) for multiple-input multiple-output(MIMO) visible light communications(VLC). In the BICGSM-ID scheme, the information bits conveyed by the signal-domain(SiD) symbols and the spatial-domain(SpD) light emitting diode(LED)-index patterns are coded by a protograph low-density parity-check(P-LDPC) code. Specifically, we propose a signal-domain symbol expanding and re-allocating(SSER) method for constructing a type of novel generalized spatial modulation(GSM) constellations, referred to as SSERGSM constellations, so as to boost the performance of the BICGSM-ID MIMO-VLC systems.Moreover, by applying a modified PEXIT(MPEXIT) algorithm, we further design a family of rate-compatible P-LDPC codes, referred to as enhanced accumulate-repeat-accumulate(EARA) codes,which possess both excellent decoding thresholds and linear-minimum-distance-growth property. Both analysis and simulation results illustrate that the proposed SSERGSM constellations and P-LDPC codes can remarkably improve the convergence and decoding performance of MIMO-VLC systems. Therefore, the proposed P-LDPC-coded SSERGSM-mapped BICGSMID configuration is envisioned as a promising transmission solution to satisfy the high-throughput requirement of MIMO-VLC applications. 展开更多
关键词 bit-interleaved coded modulation generalized spatial modulation multiple-input multipleoutput protograph LDPC codes visible light communication
在线阅读 下载PDF
PERFORMANCE OF SIMPLE-ENCODING IRREGULAR LDPC CODES BASED ON SPARSE GENERATOR MATRIX
2
作者 唐蕾 仰枫帆 《Transactions of Nanjing University of Aeronautics and Astronautics》 EI 2006年第3期202-207,共6页
A new method for the construction of the high performance systematic irregular low-density paritycheck (LDPC) codes based on the sparse generator matrix (G-LDPC) is introduced. The code can greatly reduce the enco... A new method for the construction of the high performance systematic irregular low-density paritycheck (LDPC) codes based on the sparse generator matrix (G-LDPC) is introduced. The code can greatly reduce the encoding complexity while maintaining the same decoding complexity as traditional regular LDPC (H-LDPC) codes defined by the sparse parity check matrix. Simulation results show that the performance of the proposed irregular LDPC codes can offer significant gains over traditional LDPC codes in low SNRs with a few decoding iterations over an additive white Gaussian noise (AWGN) channel. 展开更多
关键词 belief propagation iterative decoding algorithm sparse parity-check matrix sparse generator matrix H LDPC codes G-LDPC codes
在线阅读 下载PDF
CONTROL STRATEGY FOR ELECTRO-HYDRAULIC POSITION SERVO SYSTEM WITH GENERALIZED PULSE CODE MODULATION 被引量:1
3
作者 LIU Rong PAN Huachen CHEN Ying 《Chinese Journal of Mechanical Engineering》 SCIE EI CAS CSCD 2007年第3期50-53,共4页
A hybrid control strategy has been designed and developed for the electro-hydraulic posi-tion servo control system with generalized Pulse code modulation(GPCM),which is suitable for the area where the work condition i... A hybrid control strategy has been designed and developed for the electro-hydraulic posi-tion servo control system with generalized Pulse code modulation(GPCM),which is suitable for the area where the work condition is poor and a large flow rate is required.It is difficult to control the GPCM system because the system is discrete.With consideration of the stability and speediness of the GPCM position servo control system,a control strategy is developed through the theoretical and ex-perimental analyses.The control strategy integrates the merits of Bang-Bang control,PID control and fuzzy control.With this hybrid control strategy,the electro hydraulic control system has good per-formances,and the servo control is carried out with GPCM through on-off valves. 展开更多
关键词 generalized pulse code modulation(GPCM) Position control Electro hydraulic control
在线阅读 下载PDF
Index modulation aided generalized space-time block coding:A unified MIMO framework 被引量:1
4
作者 Zhiqin Wang Hui Liu +3 位作者 Lixia Xiao Liuke Li Guochao Song Tao Jiang 《Digital Communications and Networks》 CSCD 2024年第5期1459-1470,共12页
In this paper,Index Modulation(IM)aided Generalized Space-Time Block Coding(GSTBC)is proposed,which intrinsically exploits the benefits of IM concept,diversity gain and spatial multiplexing gain.Specifically,the infor... In this paper,Index Modulation(IM)aided Generalized Space-Time Block Coding(GSTBC)is proposed,which intrinsically exploits the benefits of IM concept,diversity gain and spatial multiplexing gain.Specifically,the information bits are partitioned into U groups,with each being modulated by IM symbols(i.e.Spatial Modulation(SM),Quadrature SM(QSM),etc).Next,the structure of GSTBC is invoked for each K IM symbol,and a total ofμ=U/K GSTBC codes are transmitted via T time slots.A Block Expectation Propagation(B-EP)detector is designed for the proposed IM-GSTBC structure.Moreover,the theoretical Average Bit Error Probability(ABEP)is derived for our IM-GSTBC system,which is confirmed by the simulation results and helpful for performance evaluation.Simulation results show that our proposed IM-GSTBC system is capable of striking an efficient trade-off between spatial multiplexing gain,spatial diversity gain as well as implementation cost imposed for both small-scale and large-scale MIMO antenna configurations. 展开更多
关键词 Index modulation(IM) generalized space time block coding(GSTBC) Spatial multiplexing gain Diversity gain Block expectation propagation(B-EP)
在线阅读 下载PDF
PN code generator based on generalized Bent function for SW/SFH System
5
《Wuhan University Journal of Natural Sciences》 CAS 1996年第2期221-224,共4页
Generalized Bent function and generalized Bent function sequences are introduced in this paper.The main performance or these sequences used as SW/SFH(Short Wave/Slow Frequency Hopping) code are studied. And the hardwa... Generalized Bent function and generalized Bent function sequences are introduced in this paper.The main performance or these sequences used as SW/SFH(Short Wave/Slow Frequency Hopping) code are studied. And the hardware circuit and the soflware program flow chart of the SW/SFH PN code generator are also given,which is based on generalized Bent function sequence generator by using a single chip mlcrocomputer. 展开更多
关键词 SW/FH bent function sequence PN code generator single chip microcomputer
在线阅读 下载PDF
A New Generalized Complex Orthogonal Space-Time Block Codes
6
作者 单红梅 胡光锐 《Journal of Shanghai University(English Edition)》 CAS 2005年第3期241-243,共3页
Recent research challenges in the wireless communication include the usage of diversity and efficient coding to improve data transmission quality and spectral efficiency. Space diversity uses multiple transmitting and... Recent research challenges in the wireless communication include the usage of diversity and efficient coding to improve data transmission quality and spectral efficiency. Space diversity uses multiple transmitting and/or receiving antennas to create independent fading channels without penalty in bandwidth efficiency. Space-time block coding is an encoding scheme for communication over Rayleigh fading channels using multiple transmitting antennas. Space-time block codes from complex orthogonal designs exist only for two transmitting antennas. This paper generalizes a new complex orthogonal space-time block code for four transmitting antennas, whose decoding complexity is very low. Simulations show that the generalized complex orthogonal space-time block code has low bit error rate, full rate and possibly large diversity. 展开更多
关键词 space diversity space-time block code multiple antennas generalized complex orthogonal designs.
在线阅读 下载PDF
HIGH-PERFORMANCE SIMPLE-ENCODING GENERATOR-BASED SYSTEMATIC IRREGULAR LDPC CODES AND RESULTED PRODUCT CODES
7
作者 Yang Fengfan Ye Ming Luo Lin 《Journal of Electronics(China)》 2007年第5期613-621,共9页
Low-Density Parity-Check (LDPC) code is one of the most exciting topics among the coding theory community.It is of great importance in both theory and practical communications over noisy channels.The most advantage of... Low-Density Parity-Check (LDPC) code is one of the most exciting topics among the coding theory community.It is of great importance in both theory and practical communications over noisy channels.The most advantage of LDPC codes is their relatively lower decoding complexity compared with turbo codes,while the disadvantage is its higher encoding complexity.In this paper,a new ap- proach is first proposed to construct high performance irregular systematic LDPC codes based on sparse generator matrix,which can significantly reduce the encoding complexity under the same de- coding complexity as that of regular or irregular LDPC codes defined by traditional sparse parity-check matrix.Then,the proposed generator-based systematic irregular LDPC codes are adopted as con- stituent block codes in rows and columns to design a new kind of product codes family,which also can be interpreted as irregular LDPC codes characterized by graph and thus decoded iteratively.Finally, the performance of the generator-based LDPC codes and the resultant product codes is investigated over an Additive White Gaussian Noise (AWGN) and also compared with the conventional LDPC codes under the same conditions of decoding complexity and channel noise. 展开更多
关键词 LDPC code Iterative decoding Parity-check generator matrix Product code GRAPH Kronecker product
在线阅读 下载PDF
Generalized Minimum Rank Distance of Variable-Rate Linear Network Error Correction Codes
8
作者 ZHOU Hang 《Wuhan University Journal of Natural Sciences》 CAS CSCD 2021年第1期19-23,共5页
By extending the notion of the minimum distance for linear network error correction code(LNEC), this paper introduces the concept of generalized minimum rank distance(GMRD) of variable-rate linear network error correc... By extending the notion of the minimum distance for linear network error correction code(LNEC), this paper introduces the concept of generalized minimum rank distance(GMRD) of variable-rate linear network error correction codes. The basic properties of GMRD are investigated. It is proved that GMRD can characterize the error correction/detection capability of variable-rate linear network error correction codes when the source transmits the messages at several different rates. 展开更多
关键词 network error correction code error pattern generalized minimum distance variable-rate
原文传递
Integrated Joint Source-Channel Symbol-by-Symbol Decoding of Variable-Length Codes Using 3-D MAP Sequence Estimation 被引量:1
9
作者 WU Jing CHEN Shuzhen 《Wuhan University Journal of Natural Sciences》 CAS 2007年第3期471-475,共5页
Most of multimedia schemes employ variable-length codes (VLCs) like Huffman code as core components in obtaining high compression rates. However VLC methods are very sensitive to channel noise. The goal of this pape... Most of multimedia schemes employ variable-length codes (VLCs) like Huffman code as core components in obtaining high compression rates. However VLC methods are very sensitive to channel noise. The goal of this paper is to salvage as many data from the damaged packets as possible for higher audiovisual quality. This paper proposes an integrated joint source-channel decoder (I-JSCD) at a symbol-level using three-dimensional (3-D) trellis representation for first-order Markov sources encoded with VLC source code and convolutional channel code. This method combines source code and channel code state-spaces and bit-lengths to construct a two-dimensional (2-D) state-space, and then develops a 3-D trellis and a maximum a-posterior (MAP) algorithm to estimate the source sequence symbol by symbol. Experiment results demonstrate that our method results in significant improvement in decoding performance, it can salvage at least half of (50%) data in any channel error rate, and can provide additional error resilience to VLC stream like image, audio, video stream over high error rate links. 展开更多
关键词 integrated joint source-channel decoding (I-JSCD) variable-length code (VLC) exp-golomb code convolutional code: maximum a-oosteriori (MAP)
在线阅读 下载PDF
A Note on Characteristic Generators of a Free Cyclic Code and Its Dual
10
作者 YANG Qinqin QIN Zhongping 《Wuhan University Journal of Natural Sciences》 CAS 2009年第1期14-18,共5页
Let C be a free cyclic code over Zp^a and dim pC = k. In the paper, we prove that if the k characteristic generators of C are p-linearly independent then the corresponding nα- k characteristic generators of C^⊥ are ... Let C be a free cyclic code over Zp^a and dim pC = k. In the paper, we prove that if the k characteristic generators of C are p-linearly independent then the corresponding nα- k characteristic generators of C^⊥ are p-linearly independent. We then show that to any trellis that can be constructed from k p-linearly independent characteristic generators of C, there exists a trellis for C^⊥ with the same state-complexity profile, which generalizes the conjecture of Koetter and Vardy to a free cyclic code over Zpo. 展开更多
关键词 Hensel lift cyclic codes characteristic generators trellises state-complexity profile
原文传递
Nondifferentiable Multiobjective Programming under Generalized d_I-G-Type I Invexity
11
作者 闫春雷 《Journal of Donghua University(English Edition)》 EI CAS 2013年第4期293-297,共5页
To relax convexity assumptions imposed on the functions in theorems on sufficient conditions and duality,new concepts of generalized dI-G-type Ⅰ invexity were introduced for nondifferentiable multiobjective programmi... To relax convexity assumptions imposed on the functions in theorems on sufficient conditions and duality,new concepts of generalized dI-G-type Ⅰ invexity were introduced for nondifferentiable multiobjective programming problems.Based upon these generalized invexity,G-Fritz-John (G-F-J) and G-Karnsh-Kuhn-Tucker (G-K-K-T) types sufficient optimality conditions were established for a feasible solution to be an efficient solution.Moreover,weak and strict duality results were derived for a G-Mond-Weir type dual under various types of generalized dI-G-type Ⅰ invexity assumptions. 展开更多
关键词 nondifferentiable multiobjective program efficient solution generalized dI-G-type invexity sufficient optimality conditions dualityCLC number:O221.6Document code:AArticle ID:1672-5220(2013)04-0293-05
在线阅读 下载PDF
A Novel Algorithm of Error Check and Code Generation for Structured Flowchart
12
作者 Mingcheng Qu Naigang Cui +1 位作者 Xianghu Wu Yongchao Tao 《Journal of Harbin Institute of Technology(New Series)》 EI CAS 2017年第4期18-30,共13页
Structured flowchart( SFC) and Automatic code generation based on SFC( CG-SFC) have been widely used in software requirements,design and testing phases. Some CG-SFC tools such as Rhapsody have the ability to build flo... Structured flowchart( SFC) and Automatic code generation based on SFC( CG-SFC) have been widely used in software requirements,design and testing phases. Some CG-SFC tools such as Rhapsody have the ability to build flowchart and generate code,but they do not check whether a given flowchart is correct or structural. For unstructured error ‘goto'statements will be generated randomly. We proposed three algorithms and some error recognition criteria to solve those problems. Structure recognition algorithm can recognize Selection,While/for and do-while structures. Error recognition algorithm incorporating criteria can check all the errors. At last,we develop a CG-SFC system,and compared with existing Rhapsody,it shows that the proposed algorithms are correct and effective. 展开更多
关键词 STRUCTURED flowchart CODE generation ERROR CHECK FEATURE RECOGNITION
在线阅读 下载PDF
Research and Application of Code Automatic Generation Algorithm Based on Structured Flowchart
13
作者 Xiang-Hu Wu Ming-Cheng Qu +1 位作者 Zhi-Qiang Liu Jian-Zhong Li 《Journal of Software Engineering and Applications》 2011年第9期534-545,共12页
It is of great significance to automatically generate code from structured flowchart. There are some deficiencies in existing researches, and their key algorithms and technologies are not elaborated, also there are ve... It is of great significance to automatically generate code from structured flowchart. There are some deficiencies in existing researches, and their key algorithms and technologies are not elaborated, also there are very few full-featured integrated development platforms that can generate code automatically based on structured flowchart. By analyzing the characteristics of structured flowchart, a structure identification algorithm for structured flowchart is put forward. The correctness of algorithm is verified by enumeration iteration. Then taking the identified flowchart as input, an automatic code generation algorithm is proposed. Also the correctness is verified by enumeration iteration. Finally an integrated development platform is developed using those algorithms, including flowchart modeling, code automatic generation, CDT\GCC\GDB etc. The correctness and effectiveness of algorithms proposed are verified through practical operations. 展开更多
关键词 AUTOMATIC generATION of CODES STRUCTURED Flowchart Identification of Structure INTEGRATED Development PLATFORM
在线阅读 下载PDF
A Solution Based on Modeling and Code Generation for Embedded Control System
14
作者 Guohua WU Dongwu CHENG Zhen ZHANG 《Journal of Software Engineering and Applications》 2009年第3期160-164,共5页
With the development of computer technology, embedded control system plays an important role in modern industry. For the embedded system, traditional development methods are time-consuming and system is not easy to ma... With the development of computer technology, embedded control system plays an important role in modern industry. For the embedded system, traditional development methods are time-consuming and system is not easy to maintain. Domain-specific modeling provides a solution for the problems. In this paper, we proposed development architecture for embedded control systems based on MIC. GME is used to construct meta-model and application model, model in-terpreter interprets model and stores model information in xml format document. The final cross-platform codes are automatically generated by different templates and xml format document. This development method can reduce time and cost in the lifecycle of system development. 展开更多
关键词 Domain-Specific MODELING Model INTERPRETER CODE generATION EMBEDDED Control System
在线阅读 下载PDF
Quad-Level Cell NAND Design and Soft-Bit Generation for Low-Density Parity-Check Decoding in System-Level Application
15
作者 LIU Shijun ZOU Xuecheng WANG Baocun 《Wuhan University Journal of Natural Sciences》 CAS CSCD 2018年第1期70-78,共9页
QLC(Quad-Level Cell) NAND flash will be one of the future technologies for next generation memory chip after three-dimensional(3D) TLC(Triple-Level Cell) stacked NAND flash. In QLC device, data errors will easil... QLC(Quad-Level Cell) NAND flash will be one of the future technologies for next generation memory chip after three-dimensional(3D) TLC(Triple-Level Cell) stacked NAND flash. In QLC device, data errors will easily occur because of 2~4 data levels in the limited voltage range. This paper studies QLC NAND technology which is 4 bits per cell. QLC programming methods based on 16 voltage levels and reading method based on "half-change" Gray coding are researched. Because of the probable error impact of QLC NAND cell's voltage change, the solution of generating the soft information after XOR(exclusive OR) the soft bits by internal read mechanism is presented for Low-Density Parity-Check(LDPC) Belief Propagation(BP) decoding in QLC design for its system level application. 展开更多
关键词 QLC (Quad-Level Cell)NAND error-correcting code(ECC) Low-Density Parity-Check (LDPC) Soft-Bit generation
原文传递
VCodePPA:面向集成电路物理约束优化的Verilog代码数据集
16
作者 陈溪源 姜宇轩 +2 位作者 夏莹杰 胡冀 周奕兆 《电子与信息学报》 北大核心 2025年第11期4606-4619,共14页
Verilog作为主流硬件描述语言之一,其代码质量直接影响电路的功耗、性能和面积(PPA)。当前,基于大语言模型(LLM)生成硬件描述语言的应用面临一个关键挑战:如何引入基于PPA指标的设计反馈机制,以有效指导模型优化,而不仅停留在Verilog语... Verilog作为主流硬件描述语言之一,其代码质量直接影响电路的功耗、性能和面积(PPA)。当前,基于大语言模型(LLM)生成硬件描述语言的应用面临一个关键挑战:如何引入基于PPA指标的设计反馈机制,以有效指导模型优化,而不仅停留在Verilog语法和功能正确性层面。针对这一问题,该文公开了名为VCodePPA的数据集,该数据集将Verilog代码结构与功耗、性能、面积指标进行了精准关联,共包含17342条高质量样本数据。该文所做工作为:构建了包含基本算术运算模块、存储架构模块等20种功能分类的3500条规模的种子数据集;设计了基于蒙特卡罗搜索(MCTS)的多维代码数据增强方案,通过架构层、逻辑层和时序层3个维度的9种变换器对种子数据集进行代码变化,生成大规模功能等同但PPA指标差异显著的Verilog代码集。每条数据均包含如板上资源占用量、关键路径延迟、最大工作频率等多种硬件设计指标,用于训练模型在硬件设计PPA指标的冲突-平衡规律知识。实验表明经此数据集训练后,新模型相比基线模型在各种Verilog设计任务上平均减少了10%~15%的板上资源占用,降低了8%~12%的功耗,并缩短了5%~8%的关键路径延迟。 展开更多
关键词 PPA优化 硬件语言生成 多维代码数据增强方案 双任务架构训练
在线阅读 下载PDF
系统电磁脉冲的定标关系及在腔体系统电磁脉冲中的应用
17
作者 孙会芳 易涛 +1 位作者 张玲玉 杨冬 《强激光与粒子束》 北大核心 2026年第1期175-181,共7页
研究系统电磁脉冲(SGEMP)的定标理论,掌握SGEMP的定标法则,并将其应用于腔体SGEMP,提出保持腔壁材料厚度不变的规范变换方法,应用全电磁粒子模拟方法方法进行数值模拟研究,通过对比实验室激光等离子体源条件下的原模型与10倍扩比模型的... 研究系统电磁脉冲(SGEMP)的定标理论,掌握SGEMP的定标法则,并将其应用于腔体SGEMP,提出保持腔壁材料厚度不变的规范变换方法,应用全电磁粒子模拟方法方法进行数值模拟研究,通过对比实验室激光等离子体源条件下的原模型与10倍扩比模型的模拟结果,发现两者在电场、磁场及电荷密度的空间分布规律上呈现严格一致性,幅值大小完全遵循定标关系:发射端面表面电场峰值原模型为2.0 MV/m,扩比模型为200 kV/m;磁场峰值原模型为0.8×10^(-3) T,扩比模型为0.8×10^(-4) T;电荷密度峰值原模型为6.0×10^(-3) m^(-3),扩比模型为6.0×10^(-5) m^(-3)。模拟结果验证了定标法则应用于腔体SGEMP的有效性。该研究为腔体SGEMP的物理机制解析及实验室模拟实验设计提供了理论依据。 展开更多
关键词 系统电磁脉冲 定标关系 全电磁粒子模拟方法 扩比模型
在线阅读 下载PDF
基于大语言模型的SQL注入漏洞检测载荷生成方法
18
作者 顾兆军 李丽 隋翯 《信息网络安全》 北大核心 2026年第2期274-290,共17页
针对现有SQL注入漏洞检测方法存在鲁棒性不足以及测试用例缺乏针对性等问题,文章提出一种基于大语言模型的SQL注入漏洞检测载荷生成方法。该方法通过生成针对性的检测载荷实现SQL注入漏洞检测,借助提示工程与DeepSeek-V3模型自动提取和... 针对现有SQL注入漏洞检测方法存在鲁棒性不足以及测试用例缺乏针对性等问题,文章提出一种基于大语言模型的SQL注入漏洞检测载荷生成方法。该方法通过生成针对性的检测载荷实现SQL注入漏洞检测,借助提示工程与DeepSeek-V3模型自动提取和统一构建漏洞特征;利用贡献度对漏洞特征进行分析和选择,构建模型的核心输入;通过将关键特征组织成思维链的形式促进多维度漏洞表征融合,并采用低秩适配技术对Qwen模型进行领域自适应监督微调。实验在多个公开漏洞靶场中验证Qwen模型与SqliGPT、GPT-2-web和SQLMap等模型的性能差异和生成质量,并深入分析DeepSeek-V3模型在复杂SQL注入漏洞数据中的特征提取能力。实验结果表明,Qwen模型的平均检测准确率达到75%以上,比SqliGPT、GPT-2-web和SQLMap模型分别提升49.18%、59.64%和15.19%,且载荷生成质量显著优于现有模型,证明了基于大语言模型生成检测载荷,实现SQL注入漏洞检测方法的有效性与优越性。 展开更多
关键词 大语言模型 SQL注入漏洞 代码生成 检测载荷
在线阅读 下载PDF
基于多码深度特征融合生成对抗网络的文本生成图像方法
19
作者 顾广华 孙文星 伊柏宇 《电子与信息学报》 北大核心 2026年第1期287-296,共10页
文本生成图像是一项极具挑战的跨模态任务,其核心在于生成与文本描述高度一致、细节丰富的高质量图像。当前基于生成对抗网络的方法多依赖单一噪声输入,导致生成图像细粒度不足;同时,单词级特征利用不充分,也制约了文本与图像之间的语... 文本生成图像是一项极具挑战的跨模态任务,其核心在于生成与文本描述高度一致、细节丰富的高质量图像。当前基于生成对抗网络的方法多依赖单一噪声输入,导致生成图像细粒度不足;同时,单词级特征利用不充分,也制约了文本与图像之间的语义对齐精度。为此,该文提出一种多码深度特征融合生成对抗网络(mDFAGAN)。该方法通过设计多噪声输入生成器与多码先验融合模块,提升生成图像的细节表现力;在生成器中引入多头注意力机制,从多角度对齐单词与图像子区域,增强语义一致性;此外,提出多码先验融合损失以稳定训练过程。在CUB和COCO数据集上的实验结果表明,所提方法在IS与FID评价指标上均优于当前主流生成对抗网络方法,能够生成更逼真、细节更丰富、语义一致性更强的图像。 展开更多
关键词 文本生成图像 生成对抗网络 跨模态 多码先验融合
在线阅读 下载PDF
基于语义重排序的代码注释生成方法 被引量:1
20
作者 李重 施超煊 +3 位作者 潘敏学 张天 王林章 李宣东 《软件学报》 北大核心 2026年第2期601-620,共20页
代码注释是对源代码功能的自然语言描述,其可以帮助开发人员快速地理解代码语义及功能,从而提高软件开发和维护的效率.然而,书写与维护代码注释费时费力,导致代码注释经常出现缺失、不匹配以及过时等问题.因此,如何自动化地为源代码生... 代码注释是对源代码功能的自然语言描述,其可以帮助开发人员快速地理解代码语义及功能,从而提高软件开发和维护的效率.然而,书写与维护代码注释费时费力,导致代码注释经常出现缺失、不匹配以及过时等问题.因此,如何自动化地为源代码生成注释引起了大量研究人员的关注.现有方法通常利用信息检索技术或深度学习技术来进行代码注释自动生成,但这二者均存在自身的一些局限.目前已有一些对信息检索技术和深度学习技术进行集成的研究工作,但它们无法有效利用这两种技术优势.针对这些问题,提出一种基于语义重排序的代码注释生成方法SRBCS,该方法通过语义重排序模型对不同方法所生成代码注释进行排序选择来实现代码注释生成,从而在实现对不同方法集成的同时最大化地利用不同方法在代码注释生成上的优势.在两个数据集上将SRBCS与14种代码注释生成方法进行比较.实验评估结果表明SRBCS可以有效地对不同代码注释生成方法进行集成,实现了优于现有14种代码注释生成方法的性能. 展开更多
关键词 代码注释生成 语义重排序模型 对比学习
在线阅读 下载PDF
上一页 1 2 110 下一页 到第
使用帮助 返回顶部