Hierarchical semantic-aware neural code representation

Jiang, Yuan; Su, Xin; Wang, Tiantian

doi:10.1016/j.jss.2022.111355

Cited by 10 publications

(9 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tree-LSTM processes structured input through tree-based units, with Child-Sum Tree-LSTM designed for high branching factor trees, and Binary Tree-LSTM suited for binary trees. Graph-LSTM [31] adapts the Tree-LSTM architecture for processing graph data, incorporating syntax and semantics while preserving the original structures. Nevertheless, LSTM architectures are significantly more complex in the hidden layer, resulting in approximately four times more parameters than a simple RNN architecture [32].…”

Section: Recurrent Neural Network (Rnns)mentioning

confidence: 99%

“…CDLH was compared to Deckard [83], DLC [4], and SourcererCC [82] on the BigCloneBench and OJClone benchmarks, with results indicating that CDLH's superior performance in terms of F1-score on both benchmarks. Ullah et al [31] (S19) proposed a technique for source code representation using a semantic graph to capture both syntax and semantic features. Their study compared this technique to several others, including Deckard [83], DLC [4], SourcererCC [82], CDLH [63], FCCD [90], ASTNN [49], CodeBert [31] (S19) showed strong performance on OJClone, excelling in terms of precision, recall, and F1-score, closely followed by ASTNN [47].…”

Section: Plos Onementioning

confidence: 99%

“…Ullah et al [31] (S19) proposed a technique for source code representation using a semantic graph to capture both syntax and semantic features. Their study compared this technique to several others, including Deckard [83], DLC [4], SourcererCC [82], CDLH [63], FCCD [90], ASTNN [49], CodeBert [31] (S19) showed strong performance on OJClone, excelling in terms of precision, recall, and F1-score, closely followed by ASTNN [47]. Additionally, the DFS technique [31] (S19) outperformed other techniques on both benchmarks in terms of precision, recall, and F1-score, with ASTNN [47] as the next best performer.…”

Section: Plos Onementioning

confidence: 99%

See 2 more Smart Citations

A systematic literature review on the applications of recurrent neural networks in code clone research

Quradaa,

Shahzad,

Almoqbily

2024

PLoS ONE

View full text Add to dashboard Cite

Code clones, referring to code fragments that are either similar or identical and are copied and pasted within software systems, have negative effects on both software quality and maintenance. The objective of this work is to systematically review and analyze recurrent neural network techniques used to detect code clones to shed light on the current techniques and offer valuable knowledge to the research community. Upon applying the review protocol, we have successfully identified 20 primary studies within this field from a total of 2099 studies. A deep investigation of these studies reveals that nine recurrent neural network techniques have been utilized for code clone detection, with a notable preference for LSTM techniques. These techniques have demonstrated their efficacy in detecting both syntactic and semantic clones, often utilizing abstract syntax trees for source code representation. Moreover, we observed that most studies applied evaluation metrics like F-score, precision, and recall. Additionally, these studies frequently utilized datasets extracted from open-source systems coded in Java and C programming languages. Notably, the Graph-LSTM technique exhibited superior performance. PyTorch and TensorFlow emerged as popular tools for implementing RNN models. To advance code clone detection research, further exploration of techniques like parallel LSTM, sentence-level LSTM, and Tree-Structured GRU is imperative. In addition, more research is needed to investigate the capabilities of the recurrent neural network techniques for identifying semantic clones across different programming languages and binary codes. The development of standardized benchmarks for languages like Python, Scratch, and C#, along with cross-language comparisons, is essential. Therefore, the utilization of recurrent neural network techniques for clone identification is a promising area that demands further research.

show abstract

Section: Recurrent Neural Network (Rnns)mentioning

confidence: 99%

Section: Plos Onementioning

confidence: 99%

Section: Plos Onementioning

confidence: 99%

See 1 more Smart Citation

A systematic literature review on the applications of recurrent neural networks in code clone research

Quradaa,

Shahzad,

Almoqbily

2024

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…The study provides valuable insights into the potential of semantic enrichment for improving data quality in sensor networks and offers a promising solution for overcoming data interoperability challenges in this domain. Furthermore, Jiang et al [20] discuss the use of neural networks to represent computer code in an hierarchical, semantically-aware manner. The authors proposed a method for improving computer code representation in order to better capture its meaning and structure.…”

Section: Related Workmentioning

confidence: 99%

Join Operation for Semantic Data Enrichment of Asynchronous Time Series Data

Garcia

Peyman

Piè

et al. 2023

Axioms

View full text Add to dashboard Cite

In this paper, we present a novel framework for enriching time series data in smart cities by supplementing it with information from external sources via semantic data enrichment. Our methodology effectively merges multiple data sources into a uniform time series, while addressing difficulties such as data quality, contextual information, and time lapses. We demonstrate the efficacy of our method through a case study in Barcelona, which permitted the use of advanced analysis methods such as windowed cross-correlation and peak picking. The resulting time series data can be used to determine traffic patterns and has potential uses in other smart city sectors, such as air quality, energy efficiency, and public safety. Interactive dashboards enable stakeholders to visualize and summarize key insights and patterns.

show abstract

“…Additionally, depending on the size and complexity of the solutions being compared, line-level or function-level comparisons might be appropriate. Related research on source code representation [26] has also considered methods such as control flow graphs and program dependence graphs. These approaches could potentially bring even more benefits to the comparison process, but they also present challenges in terms of visual representation and usability for programmers.…”

Section: B Using Artificial Intelligence To Generate Codementioning

confidence: 99%

Navigating Complexity in Software Engineering: A Prototype for Comparing GPT-n Solutions

Treude

2023

Preprint

View full text Add to dashboard Cite

Navigating the diverse solution spaces of non-trivial software engineering tasks requires a combination of technical knowledge, problem-solving skills, and creativity. With multiple possible solutions available, each with its own set of trade-offs, it is essential for programmers to evaluate the various options and select the one that best suits the specific requirements and constraints of a project. Whether it is choosing from a range of libraries, weighing the pros and cons of different architecture and design solutions, or finding unique ways to fulfill user requirements, the ability to think creatively is crucial for making informed decisions that will result in efficient and effective software. However, the interfaces of current chatbot tools for programmers, such as OpenAI's ChatGPT or GitHub Copilot, are optimized for presenting a single solution, even for complex queries. While other solutions can be requested, they are not displayed by default and are not intuitive to access. In this paper, we present our work-in-progress prototype "GPTCOMPARE", which allows programmers to visually compare multiple source code solutions generated by GPT-n models for the same programmingrelated query by highlighting their similarities and differences.

show abstract

Hierarchical semantic-aware neural code representation

Cited by 10 publications

References 23 publications

A systematic literature review on the applications of recurrent neural networks in code clone research

A systematic literature review on the applications of recurrent neural networks in code clone research

Join Operation for Semantic Data Enrichment of Asynchronous Time Series Data

Navigating Complexity in Software Engineering: A Prototype for Comparing GPT-n Solutions

Contact Info

Product

Resources

About