检索结果-内蒙古大学图书馆

Exploring the Boundaries Between LLM code clone detection and code Similarity Assessment on Human and AI-Generated code

引用

BIG DATA AND COGNITIVE COMPUTING 2025年第2期9卷 41-41页

作者： Zhang, Zixian Saber, Takfarinas Univ Galway Sch Comp Sci CRT AI Galway H91TK33 Ireland Univ Galway Sch Comp Sci Lero Galway H91 TK33 Ireland

As Large Language Models (LLMs) continue to advance, their capabilities in code clone detection have garnered significant attention. While much research has assessed LLM performance on human-generated code, the proliferation of LLM-generated code raises critical questions about their ability to detect clones across both human- and LLM-created codebases, as this capability remains largely unexplored. This paper addresses this gap by evaluating two versions of LLaMA3 on these distinct types of datasets. Additionally, we perform a deeper analysis beyond simple prompting, examining the nuanced relationship between code cloning and code similarity that LLMs infer. We further explore how fine-tuning impacts LLM performance in clone detection, offering new insights into the interplay between code clones and similarity in human versus AI-generated code. Our findings reveal that LLaMA models excel in detecting syntactic clones but face challenges with semantic clones. Notably, the models perform better on LLM-generated datasets for semantic clones, suggesting a potential bias. The fine-tuning technique enhances the ability of LLMs to comprehend code semantics, improving their performance in both code clone detection and code similarity assessment. Our results offer valuable insights into the effectiveness and characteristics of LLMs in clone detection and code similarity assessment, providing a foundation for future applications and guiding further research in this area.

关键词： code clone detection code similarity large language model fine-tuning LLM-generated code

来源：评论

学校读者我要写书评

暂无评论

An optimized hybrid deep learning model for code clone detection

引用

International Journal of Information Technology (Singapore) 2025年第3期17卷 1589-1595页

作者： Geetika Kaur, Navdeep Kaur, Amandeep Sri Guru Granth Sahib World University Fatehgarh Sahib India Lovely Professional University Punjab Phagwara India NIT Kurukshetra Punjab Haryana India

code clones in software system are identical or similar pieces of code. The code is repeatedly generated by the copy and paste program. As a result, every duplicate contains a defect that was detected in one unit and the existing techniques are unable to achieve high accuracy for the code clone detection. In this research work, a hybrid deep learning model is proposed which comprises four phases namely pre-processing, feature set generation, feature set optimization and clone detection. We have utilized particle swarm optimization (PSO) and genetic algorithm (GA) for optimization along with convolutional neural network (CNN) and long short-term memory (LSTM) for clone detection. The proposed model is implemented in python and tested on several datasets in terms of accuracy (%), precision (%) and recall (%). In addition to this, the proposed model is compared with existing recent studies in terms of performance and the results show that the proposed hybrid model attains the highest accuracy (94.67%), highest precision (93.12%) and highest recall (93.13%) in case of big clone bench (BCB) dataset. Similarly, our model attains the highest accuracy (93.90%), highest precision (93.50%) and highest recall (93.52%) in case of Google code jam dataset while in case of Java dataset, accuracy, precision and recall are 93.78%, 92.67% and 92.66% respectively. © Bharati Vidyapeeth's Institute of Computer Applications and Management 2024.

关键词： CNN code clone detection Feature set optimization GA LSTM PSO

来源：评论

学校读者我要写书评

暂无评论

An enhanced transformer-based framework for interpretable code clone detection

引用

JOURNAL OF SYSTEMS AND SOFTWARE 2025年 222卷

作者： Nashaat, Mona Amin, Reem Eid, Ahmad Hosny Abdel-Kader, Rabab F. Port Said Univ Fac Engn Elect Engn Dept Port Said 42526 Egypt

In software development, the replication of specific source code segments is known as code cloning. This practice allows reusing source code instead of developing these segments from scratch, enhancing software productivity. However, code cloning can introduce bugs, complicate code refactoring, and increase maintenance costs. Consequently, code clone detection (CCD) is an essential concern for the software industry. While various techniques have been proposed for detecting code clones, many existing tools generate a high ratio of false positives/negatives and a need for more contextual awareness. Therefore, this paper introduces cloneXformer, an innovative framework for code clone detection. The framework adopts a collaborative approach that harnesses multiple large language models for code understanding. The framework employs a preliminary phase to preprocess the input code, which helps the models understand and represent the code efficiently. Then, it captures the semantic level of the code and the syntactic level as it relies on a set of transformer-based models. Afterward, these models are finetuned to detect code clones with interpretable results that explain the detected clone types. Finally, the output of these models is combined to provide a unified final prediction. The empirical evaluation indicates that the framework improves detection performance, achieving an approximately 16.88 % higher F1 score than the state-of-the-art techniques.

关键词： Software development code clone detection Software productivity Large language models Transformer-based models

来源：评论

学校读者我要写书评

暂无评论

code clone detection Based on Contrastive Learning 4

Code Clone Detection Based on Contrastive Learning

引用

4th IEEE International Conference on Software Engineering and Artificial Intelligence (SEAI)

作者： Xie, Chunli Liang, Yao Lv, Quanrun Wan, Zexuan Jiangsu Normal Univ Dept Comp Sci & Technol Xuzhou Jiangsu Peoples R China

ISBN: (纸本)9798350374353;9798350374346

In recent years, an increasing number of deep learning-based code representation learning techniques have achieved great success in the field of program analysis. These methods require large scale labeled samples for training the model. Due to the costly and time-consuming manual labeling, a large amount of data in open-source platforms lacks labels or has a few labels, which leads to the limitation of the effectiveness and practicality of the existing methods. To solve this problem, we propose a method based on contrastive learning to learn source code representation under small sample, that uses gated recurrent unit to extract code features and apply it to the code clone detection. Experimental results show that the method outperforms the state-of-the-art methods by 8% with respect to F1 score on benchmark dataset.

关键词： self-supervised learning contrastive learning code clone detection

来源：评论

学校读者我要写书评

暂无评论

code clone detection based on Event Embedding and Event Dependency 22

Code Clone Detection based on Event Embedding and Event Depe...

引用

13th Asia-Pacific Symposium on Internetware (Internetware) - A Software Paradigm for Internet Computing

作者： Huang, Cheng Zhou, Hui Ye, Chunyang Li, Bingzhuo Hainan Univ Haikou Hainan Peoples R China

ISBN: (纸本)9781450397803

The code clone detection method based on semantic similarity has important value in software engineering tasks (e.g., software evolution, software reuse). Traditional code clone detection technologies pay more attention to the similarity of code at the syntax level, and less attention to the semantic similarity of the code. As a result, candidate codes similar in semantics are ignored. To address this issue, we propose a code clone detection method based on semantic similarity. By treating code as a series of interdependent events that occur continuously, we design a model namely EDAM to encode code semantic information based on event embedding and event dependency. The EDAM model uses the event embedding method to model the execution characteristics of program statements and the data dependence information between all statements. In this way, we can embed the program semantic information into a vector and use the vector to detect codes similar in semantics. Experimental results show that the performance of our EDAM model is superior to state-of-the-art open source models for code clone detection.

关键词： code clone detection event embedding event dependency

来源：评论

学校读者我要写书评

暂无评论

Domain analysis of device drivers using code clone detection method

引用

ETRI JOURNAL 2008年第3期30卷 394-402页

作者： Ma, Yu-Seung Woo, Duk-Kyun ETRI S W & Content Res Lab Taejon South Korea

Domain analysis is the process of analyzing related software systems in a domain to find their common and variable parts. In the case of device drivers, they are highly suitable for domain analysis because device drivers of the same domain are implemented similarly for each device and each system that they support. Considering this characteristic, this paper introduces a new approach to the domain analysis of device drivers. Our method uses a code clone detection technique to extract similarity among device drivers of the same domain. To examine the applicability of our method, we investigated whole device drivers of a Linux source. Results showed that many reusable similar codes can be discerned by the code clone detection method. We also investigated if our method is applicable to other kernel sources. However, the results show that the code clone detection method is not useful for the domain analysis of all kernel sources. That is, the applicability of the code clone detection method to domain analysis is a peculiar feature of device drivers.

关键词： device drivers code clone detection

来源：评论

学校读者我要写书评

暂无评论

Federated Learning for Software Engineering: A Case Study of code clone detection and Defect Prediction

引用

IEEE TRANSACTIONS ON SOFTWARE ENGINEERING 2024年第2期50卷 296-321页

作者： Yang, Yanming Hu, Xing Gao, Zhipeng Chen, Jinfu Ni, Chao Xia, Xin Lo, David Zhejiang Univ Dept Comp Sci & Technol Ningbo 310058 Peoples R China Zhejiang Univ Sch Software Technol Ningbo 315048 Peoples R China Zhejiang Univ Shanghai Inst Adv Study Hangzhou 310085 Peoples R China Wuhan Univ Sch Comp Sci Wuhan Peoples R China Software Engn Applicat Technol Lab Hangzhou 310007 Peoples R China Singapore Management Univ Sch Comp & Informat Syst Singapore Singapore

In various research domains, artificial intelligence (AI) has gained significant prominence, leading to the development of numerous learning-based models in research laboratories, which are evaluated using benchmark datasets. While the models proposed in previous studies may demonstrate satisfactory performance on benchmark datasets, translating academic findings into practical applications for industry practitioners presents challenges. This can entail either the direct adoption of trained academic models into industrial applications, leading to a performance decrease, or retraining models with industrial data, a task often hindered by insufficient data instances or skewed data distributions. Real-world industrial data is typically significantly more intricate than benchmark datasets, frequently exhibiting data-skewing issues, such as label distribution skews and quantity skews. Furthermore, accessing industrial data, particularly source code, can prove challenging for Software Engineering (SE) researchers due to privacy policies. This limitation hinders SE researchers' ability to gain insights into industry developers' concerns and subsequently enhance their proposed models. To bridge the divide between academic models and industrial applications, we introduce a federated learning (FL)-based framework called Almity. Our aim is to simplify the process of implementing research findings into practical use for both SE researchers and industry developers. Almity enhances model performance on sensitive skewed data distributions while ensuring data privacy and security. It introduces an innovative aggregation strategy that takes into account three key attributes: data scale, data balance, and minority class learnability. This strategy is employed to refine model parameters, thereby enhancing model performance on sensitive skewed datasets. In our evaluation, we employ two well-established SE tasks, i.e., code clone detection and defect prediction, as evaluation tasks. We

关键词： Data models Training codes Cloning Task analysis Benchmark testing Industries Federated learning parameter aggregation strategy skewed data distribution code clone detection defect prediction

来源：评论

学校读者我要写书评

暂无评论

Semantic code clone detection Based on Community detection

引用

INTERNATIONAL JOURNAL OF SOFTWARE ENGINEERING AND KNOWLEDGE ENGINEERING 2024年第10期34卷 1661-1692页

作者： Wan, Zexuan Xie, Chunli Lv, Quanrun Fan, Yasheng Technol Jiangsu Normal Univ Sch Comp Sci Xuzhou 221116 Peoples R China

Semantic code clone detection is to find code snippets that are structurally or syntactically different, but semantically identical. It plays an important role in software reuse, code compression. Many existing studies have achieved good performance in non-semantic clone, but semantic clone is still a challenging task. Recently, several works have used tree or graph, such as Abstract Syntax Tree (AST), Control Flow Graph (CFG) or Program Dependency Graph (PDG) to extract semantic information from source codes. In order to reduce the complexity of tree and graph, some studies transform them into node sequences. However, this transformation will lose some semantic information. To address this issue, we propose a novel high-performance method that utilizes community detection to extract features of AST while preserving its semantic information. First, based on the AST of source code, we exploit community detection to split AST into different subtrees to extract the underlying semantics information of different code blocks, and use centrality analysis to quantify the semantic information as the weight of AST nodes. Then, the AST is converted into a sequence of tokens with weights, and a Siamese neural network model is used to detect the similarity of token sequences for semantic code clone detection. Finally, to evaluate our approach, we conduct experiments on two standard benchmark datasets, Google code Jam (GCJ) and BigcloneBench (BCB). Experimental results show that our model outperforms the eight publicly available state-of-the-art methods in detecting code clones. It is five times faster than the tree-based method (ASTNN) in terms of time complexity.

关键词： code clone detection semantic clone community detection centrality analysis siamese network

来源：评论

学校读者我要写书评

暂无评论

Scalable code clone detection and search based on adaptive prefix filtering

引用

JOURNAL OF SYSTEMS AND SOFTWARE 2018年 137卷 130-142页

作者： Nishi, Manziba Akanda Damevski, Kostadin Virginia Commonwealth Univ Dept Comp Sci Richmond VA 23284 USA

code clone detection is a well-known software engineering problem that aims to detect all the groups of code blocks or code fragments that are functionally equivalent in a code base. It has numerous and wide ranging important uses in areas such as software metrics, plagiarism detection, aspect mining, copyright infringement investigation, code compaction, virus detection, and detecting bugs. A scalable code clone detection technique, able to process large source code repositories, is crucial in the context of multi project or Internet-scale code clone detection scenarios. In this paper, we focus on improving the scalability of code clone detection, relative to current state of the art techniques. Our adaptive prefix filtering technique improves the performance of code clone detection for many common execution parameters, when tested on common benchmarks. The experimental results exhibit improvements for commonly used similarity thresholds of between 40% and 80%, in the best case decreasing the execution time up to 11% and increasing the number of filtered candidates up to 63%. (C) 2017 Elsevier Inc. All rights reserved.

关键词： code clone detection Prefix filtering Software maintenance

来源：评论

学校读者我要写书评

暂无评论

Evaluating few-shot and contrastive learning methods for code clone detection

引用

EMPIRICAL SOFTWARE ENGINEERING 2024年第6期29卷 163页

作者： Khajezade, Mohamad Fard, Fatemeh H. Shehata, Mohamed S. Univ British Columbia Okanagan Kelowna BC Canada

Contextcode clone detection (CCD) is a software engineering task that is used for plagiarism detection, code search, and code comprehension. Recently, deep learning-based models have achieved an F1-Score (a metric used to assess classifiers) of similar to\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sim $$\end{document}95% on the codeXGLUE benchmark. These models require many training data, mainly fine-tuned on Java or C++ datasets. However, no previous study evaluates the generalizability of these models where a limited amount of annotated data is *** main objective of this research is to assess the ability of the CCD models as well as few-shot learning algorithms for unseen programming problems and new languages (i.e., the model is not trained on these problems/languages).MethodWe assess the generalizability of the state-of-the-art models for CCD in few-shot settings (i.e., only a few samples are available for fine-tuning) by setting three scenarios: i) unseen problems, ii) unseen languages, iii) combination of new languages and new problems. We choose codeNet and conduct our experiments on Java, C++, and Ruby languages. Then, we employ Model Agnostic Meta-learning (MAML), where the model learns a meta-learner capable of extracting transferable knowledge from the train set;so that the model can be fine-tuned using a few samples. Finally, we combine contrastive learning with MAML to further study whether it can improve the results of *** results show that the performance of the models drops similar to 50%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sim 50\%$$\end{documen

关键词： Contrastive learning Few-shot learning code clone detection

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：