检索结果-内蒙古大学图书馆

Multi-grained contextual code representation learning for commit message generation

INFORMATION AND SOFTWARE TECHNOLOGY 2024年 167卷

作者： Wang, Chuangwei Zhang, Li Zhang, Xiaofang Soochow Univ Sch Comp Sci & Technol Suzhou Peoples R China

Commit messages, precisely describing the code changes for each commit in natural language, makes it possible for developers and succeeding reviewers to understand the code changes without digging into implementation details. However, the semantic and structural gap between code and natural language poses a significant challenge for commit message generation. Several researchers have proposed automated techniques to generate commit messages. Nevertheless, the information about the code is not sufficiently exploited. In this paper, we propose multi-grained contextual code representation learning for commit message generation (COMU). We extract multi-grained information from the changed code at the line and AST levels (i.e., code_Diff and AST_Diff). In code_Diff, we construct global contextual semantic information about the changed code, and mark whether a line of code has changed with three different tokens. In AST_Diff, we extract the code structure from source code changes and combine the extracted structure with four types of editing operations to explicitly focus on the detailed information of the changed part. In addition, we build the experimental datasets, since there is still no publicly sufficient dataset for this task. The release of this dataset would contribute to advancing research in this field. We perform an extensive experiment to evaluate the effectiveness of COMU. The experimental evaluation and human study show that our model outperforms the baseline model.

关键词： code change code representation learning Commit message generation Pre-training

来源：评论

学校读者我要写书评

暂无评论

Contextuality of code representation learning 38

Contextuality of Code Representation Learning

引用

38th IEEE/ACM International Conference on Automated Software Engineering (ASE)

作者： Li, Yi Wang, Shaohua Nguyen, Tien N. New Jersey Inst Technol Dept Informat Newark NJ 07102 USA Univ Texas Dallas Dept Comp Sci Dallas TX USA

ISBN: (纸本)9798350329964

Advanced machine learning models (ML) have been successfully leveraged in several software engineering (SE) applications. The existing SE techniques have used the embedding models ranging from static to contextualized ones to build the vectors for program units. The contextualized vectors address a phenomenon in natural language texts called polysemy, which is the coexistence of different meanings of a word/phrase. However, due to different nature, program units exhibit the nature of mixed polysemy. Some code tokens and statements exhibit polysemy while other tokens (e.g., keywords, separators, and operators) and statements maintain the same meaning in different contexts. A natural question is whether static or contextualized embeddings fit better with the nature of mixed polysemy in source code. The answer to this question is helpful for the SE researchers in selecting the right embedding model. We conducted experiments on 12 popular sequence-/tree-/graph-based embedding models and on the samples of a dataset of 10,222 Java projects with +14M methods. We present several contextuality evaluation metrics adapted from natural-language texts to code structures to evaluate the embeddings from those models. Among several findings, we found that the models with higher contextuality help a bug detection model perform better than the static ones. Neither static nor contextualized embedding models fit well with the mixed polysemy nature of source code. Thus, we develop HYcode, a hybrid embedding model that fits better with the nature of mixed polysemy in source code.

关键词： code representation learning Contextualized Embedding Contextuality of code Embedding

来源：评论

学校读者我要写书评

暂无评论

CoreGen: Contextualized code representation learning for Commit Message Generation

引用

NEUROCOMPUTING 2021年 459卷 97-107页

作者： Nie, Lun Yiu Gao, Cuiyun Zhong, Zhicong Lam, Wai Liu, Yang Xu, Zenglin Harbin Inst Technol Shenzhen Peoples R China Chinese Univ Hong Kong Hong Kong Peoples R China Sun Yat Sen Univ Guangzhou Peoples R China Nanyang Technol Univ Singapore Singapore

Automatic generation of high-quality commit messages for code commits can substantially facilitate software developers' works and coordination. However, the semantic gap between source code and natural language poses a major challenge for the task. Several studies have been proposed to alleviate the challenge but none explicitly involves code contextual information during commit message generation. Specifically, existing research adopts static embedding for code tokens, which maps a token to the same vector regardless of its context. In this paper, we propose a novel Contextualized code representation learning strategy for commit message Generation (CoreGen). CoreGen first learns contextualized code representations which exploit the contextual information behind code commit sequences. The learned representations of code commits built upon Transformer are then fine-tuned for downstream commit message generation. Experiments on the benchmark dataset demonstrate the superior effectiveness of our model over the baseline models with at least 28.18% improvement in terms of BLEU-4 score. Furthermore, we also highlight the future opportunities in training contextualized code representations on larger code corpus as a solution to low-resource tasks and adapting the contextualized code representation framework to other code-to-text generation tasks. (c) 2021 Elsevier B.V. All rights reserved.

关键词： Commit message generation code representation learning code-to-text generation Self-supervised learning Contextualized code representation

来源：评论

学校读者我要写书评

暂无评论

Contextuality of code representation learning 23

Contextuality of Code Representation Learning

引用

Proceedings of the 38th IEEE/ACM International Conference on Automated Software Engineering

作者： Yi Li Shaohua Wang Tien N. Nguyen Department of Informatics New Jersey Institute of Technology New Jersey USA Computer Science Department The University of Texas at Dallas Texas USA

ISBN: (纸本)9798350329964

Advanced machine learning models (ML) have been successfully leveraged in several software engineering (SE) applications. The existing SE techniques have used the embedding models ranging from static to contextualized ones to build the vectors for program units. The contextualized vectors address a phenomenon in natural language texts called polysemy, which is the coexistence of different meanings of a word/phrase. However, due to different nature, program units exhibit the nature of mixed polysemy. Some code tokens and statements exhibit polysemy while other tokens (e.g., keywords, separators, and operators) and statements maintain the same meaning in different contexts. A natural question is whether static or contextualized embeddings fit better with the nature of mixed polysemy in source code. The answer to this question is helpful for the SE researchers in selecting the right embedding *** conducted experiments on 12 popular sequence-/tree-/graph-based embedding models and on the samples of a dataset of 10,222 Java projects with +14M methods. We present several contextuality evaluation metrics adapted from natural-language texts to code structures to evaluate the embeddings from those models. Among several findings, we found that the models with higher contextuality help a bug detection model perform better than the static ones. Neither static nor contextualized embedding models fit well with the mixed polysemy nature of source code. Thus, we develop Hycode, a hybrid embedding model that fits better with the nature of mixed polysemy in source code.

关键词： code representation learning

来源：评论

学校读者我要写书评

暂无评论

A hybrid code representation learning approach for predicting method names

引用

JOURNAL OF SYSTEMS AND SOFTWARE 2021年 180卷

作者： Zhang, Fengyi Chen, Bihuan Li, Rongfan Peng, Xin Fudan Univ Sch Comp Sci Shanghai Peoples R China Fudan Univ Shanghai Key Lab Data Sci Shanghai Peoples R China

Program semantic properties such as class names, method names, and variable names and types play an important role in software development and maintenance. Method names are of particular importance because they provide the cornerstone of abstraction for developers to communicate with each other for various purposes (e.g., code review and program comprehension). Existing method name prediction approaches often represent code as lexical tokens or syntactical AST (abstract syntax tree) paths, making them difficult to learn code semantics and hindering their effectiveness in predicting method names. Initial attempts have been made to represent code as execution traces to capture code semantics, but suffer scalability in collecting execution traces. In this paper, we propose a hybrid code representation learning approach, named Meth2Seq, to encode a method as a sequence of distributed vectors. Meth2Seq represents a method as (1) a bag of paths on the program dependence graph, (2) a sequence of typed intermediate representation statements and (3) a sentence of natural language comment, to scalably capture code semantics. The learned sequence of vectors of a method is fed to a decoder model to predict method names. Our evaluation with a dataset of 280.5K methods in 67 Java projects has demonstrated that Meth2Seq outperforms the two state-of-the-art code representation learning approaches in F1-score by 92.6% and 36.6%, while also outperforming two state-of-the-art method name prediction approaches in F1-score by 85.6% and 178.1%. (C) 2021 Elsevier Inc. All rights reserved.

关键词： code representation learning Method name prediction Deep learning

来源：评论

学校读者我要写书评

暂无评论

Improving Bug Detection via Context-Based code representation learning and Attention-Based Neural Networks

引用

PROCEEDINGS OF THE ACM ON PROGRAMMING LANGUAGES-PACMPL 2019年第OOPSLA期3卷 1–30页

作者： Li, Yi Wang, Shaohua Nguyen, Tien N. Son Van Nguyen New Jersey Inst Technol Dept Informat Newark NJ 07102 USA Univ Texas Dallas Comp Sci Dept 800 W Campbell Rd Richardson TX 75080 USA

Bug detection has been shown to be an effective way to help developers in detecting bugs early, thus, saving much effort and time in software development process. Recently, deep learning-based bug detection approaches have gained successes over the traditional machine learning-based approaches, the rule-based program analysis approaches, and mining-based approaches. However, they are still limited in detecting bugs that involve multiple methods and suffer high rate of false positives. In this paper, we propose a combination approach with the use of contexts and attention neural network to overcome those limitations. We propose to use as the global context the Program Dependence Graph (PDG) and Data Flow Graph (DFG) to connect the method under investigation with the other relevant methods that might contribute to the buggy code. The global context is complemented by the local context extracted from the path on the AST built from the method's body. The use of PDG and DR;enables our model to reduce the false positive rate, while to complement for the potential reduction in recall, we make use of the attention neural network mechanism to put more weights on the buggy paths in the source code. That is, the paths that are similar to the buggy paths will be ranked higher, thus, improving the recall of our model. We have conducted several experiments to evaluate our approach on a very large dataset with +4.973M methods in 92 different project versions. The results show that our tool can have a relative improvement up to 160% on F-score when comparing with the state-of-the-art bug detection approaches. Our tool can detect 48 true bugs in the list of top 100 reported bugs, which is 24 more true bugs when comparing with the baseline approaches. We also reported that our representation is better suitable for bug detection and relatively improves over the other representations up to 206% in accuracy.

关键词： Bug Detection Deep learning code representation learning Network Embedding Program Graphs Attention Neural Networks

来源：评论

学校读者我要写书评

暂无评论

code Multiview Hypergraph representation learning for Software Defect Prediction

引用

IEEE TRANSACTIONS ON RELIABILITY 2024年第4期73卷 1863-1876页

作者： Qiu, Shaojian Huang, Mengyang Liang, Yun Peng, Chaoda Yuan, Yuan South China Agr Univ Coll Software Engn Guangzhou 510642 Peoples R China Beihang Univ Sch Comp Sci & Engn Beijing 100191 Peoples R China Zhongguancun Lab State Key Lab Software Dev Environm Beijing 100095 Peoples R China

Software defect prediction technology aids the reliability assurance team in identifying defect-prone code and assists the team in reasonably allocating limited testing resources. Recently, researchers assumed that the topological associations among code fragments could be harnessed to construct defect prediction models. Nevertheless, existing graph-based methods only concentrate on features of single-view association, which fail to fully capture the rich information hidden in the code. In addition, software defects may involve multiple code fragments simultaneously, but traditional binary graph structures are insufficient for representing these multivariate associations. To address these two challenges, this article proposes a multiview hypergraph representation learning approach (MVHR-DP) to amplify the potency of code features in defect prediction. MVHR-DP initiates by creating hypergraph structures for each code view, which are then amalgamated into a comprehensive fusion hypergraph. Following this, a hypergraph neural network is established to extract code features from multiple views and intricate associations, thereby enhancing the comprehensiveness of representation in the modeling data. Empirical study shows that the prediction model utilizing features generated by MVHR-DP exhibits superior area under the curve (AUC), F-measure, and matthews correlation coefficient (MCC) results compared to baseline approaches across within-project, cross-version, and cross-project prediction tasks.

关键词： code multiview fusion code representation learning hypergraph construction software defect prediction software reliability

来源：评论

学校读者我要写书评

暂无评论

Improving Source code Pre-Training via Type-Specific Masking

引用

ACM TRANSACTIONS ON SOFTWARE ENGINEERING AND METHODOLOGY 2025年第3期34卷 1-34页

作者： Zou, Wentao Li, Qi Li, Chuanyi Ge, Jidong Chen, Xiang Huang, Liguo Luo, Bin Nanjing Univ Natl Key Lab Novel Software Technol Nanjing Peoples R China Nantong Univ Sch Artificial Intelligence & Comp Sci Nantong Peoples R China Southern Methodist Univ Dept Comp Sci Dallas TX USA

The Masked Language Modeling (MLM) task is widely recognized as one of the most effective pre-training tasks and currently derives many variants in the Software Engineering (SE) field. However, most of these variants mainly focus on code representation without distinguishing between different code token types, while some focus on a specific type, such as code identifiers. Indeed, various code token types exist, and there is no evidence that only identifiers can improve PTMs. Thus, to improve PTMs through different types, we conducted an extensive study to evaluate how different type-specific masking tasks can affect PTMs. First, we extract five code token types, convert them into type-specific masking tasks, and generate their combinations. Second, we pre-train codeBERT and PLBART using combinations and fine-tuned them on four SE downstream tasks. Experimental results show that type-specific masking tasks can enhance codeBERT and PLBART on all downstream tasks. Furthermore, we discuss topics related to low-resource datasets, conflicting PTMs that original pre-training tasks conflict with our methods, the cost and performance of our methods, factors that impact the performance of our methods, and applying our methods on state-of-the-art PTMs. These discussions comprehensively analyze the strengths and weaknesses of different type-specific masking tasks. CCS Concepts: center dot Software and its engineering- Software creation and management;center dot Computing methodologies- Artificial intelligence

关键词： code representation learning pre-trained model pre-training tasks masked language modeling

来源：评论

学校读者我要写书评

暂无评论

TerGEC: A graph enhanced contrastive approach for program termination analysis

引用

SCIENCE OF COMPUTER PROGRAMMING 2024年 237卷

作者： Liu, Shuo Keung, Jacky Wai Yang, Zhen Liao, Yihan Li, Yishu City Univ Hong Kong Dept Comp Sci Hong Kong Peoples R China Shandong Univ Sch Comp Sci & Technol Qingdao Peoples R China

Context: Programs with non -termination behavior induce various bugs, such as denial -of -service vulnerability and memory exhaustion. Hence the ability to detect non -termination programs before software deployment is crucial. Existing detection methods are either execution -based or deep learning -based. Despite great advances, their limitations are evident. The former requires complex sandbox environments for execution, while the latter lacks fine-grained analysis. Objective: To overcome the above limitations, this paper proposes a graph -enhanced contrastive approach, namely TerGEC, which combines both inter -class and intra-class semantics to carry out a more fine-grained analysis and exempt execution during the detection process. Methods: In detail, TerGEC analyzes behaviors of programs from Abstract Syntax Trees (ASTs), thereby capturing intra-class semantics both syntactically and lexically. Besides, it incorporates contrastive learning to learn the discrepancy between program behaviors of termination and nontermination, thereby acquiring inter -class semantics. In addition, graph augmentation is designed to improve the robustness. Weighted contrastive loss and focal loss are also equipped in TerGEC to alleviate the classes -imbalance problem during the non -termination detection. Consequently, the whole detection process can be handled more fine-grained, and the execution can also be exempted due to the nature of deep learning. Results: We evaluate TerGEC on five datasets of both Python and C languages. Extensive experiments demonstrate TerGEC achieves the best performance overall. Among all experimented datasets, TerGEC outperforms state-of-the-art baselines by 8 .20% in terms of mAP and by 17 .07% in terms of AUC on average. Conclusion: TerGEC is capable of detecting non -terminating programs with high precision, showing that the combination of inter -class and intra-class learning, along with our proposed classes -imbalance solutions, is significantly ef

关键词： code representation learning Graph neural networks Contrastive learning Termination analysis

来源：评论

学校读者我要写书评

暂无评论

Ponzi Scheme Detection in Smart Contract via Transaction Semantic representation learning

引用

IEEE TRANSACTIONS ON RELIABILITY 2024年第2期73卷 1117-1131页

作者： Cai, Jie Li, Bin Zhang, Jiale Sun, Xiaobing Yangzhou Univ Sch Informat Engn Yangzhou 225102 Jiangsu Peoples R China

The Ponzi scheme implemented through smart contracts is one of the most common scams on the blockchain platform. Although various learning-based Ponzi smart contract detection approaches have been proposed, they still suffer from several limitations, i.e., 1) extracting insufficient semantics and gathering Ponzi irrelevant components from the smart contract during feature engineering, and 2) underutilizing structured semantic features during model training. As the Ponzi scheme is an economic crime with the typical Rob-Peter-to-Pay-Paul transaction pattern, we propose a transaction semantic learning based approach to mitigate the above limitations. The fundamental idea of our approach is to represent the transaction-related semantics of a smart contract as a graph and utilize a graph convolutional network (GCN) to learn the potential Ponzi-like transaction pattern from it. We define a novel code representation named slice transaction property graph (sTPG) to represent the transaction-related semantics, which can encode multiple transaction-related semantics inside a smart contract function into a graph and eliminate other irrelevant fragments. Then, we propose a relation-sensitive GCN as the learning model to identify potential Ponzi-scheme-like transaction patterns from sTPG by considering both nodes and edges features in sTPG. We evaluate our approach on two datasets: 1) smart contracts collected from Forum and Public datasets, and 2) really deployed smart contracts on the Ethereum blockchain. The experiment results show that our approach outperforms the state-of-the-art learning-based approaches.

关键词： code representation learning graph convolutional network (GCN) Ponzi scheme smart contract

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：