China
Africa
Explore chapters and articles related to this topic
T
Published in Phillip A. Laplante, Dictionary of Computer Science, Engineering, and Technology, 2017
type checking a form of semantic analysis in which rules not specified as part of the syntax of the language are used to determine if the data types of the variables of the program conform to those rules. For some languages, type checking can be performed at compile time, and is a form of static semantics. In other languages, type checking cannot be performed until the program executes, and is a form of execution semantics.
A Brief Overview of Natural Language Processing and Artificial Intelligence
Published in Brojo Kishore Mishra, Raghvendra Kumar, Natural Language Processing in Artificial Intelligence, 2020
Sushree Bibhuprada B. Priyadarshini, Amiya Bhusan Bagjadab, Brojo Kishore Mishra
Semantic analysis means the task of ensuring that some statements and declaration of programs are semantically right. That means the meaning is clear and consistent with the way in which the data types, as well as the control structures, are supposed to be employed [6].
Semantic Role Labeling Based on Valence Structure and Deep Neural Network
Published in IETE Journal of Research, 2023
In the post-processing stage of semantic role labeling, in addition to post-processing according to some inherent constraints of semantic role labeling itself, a reasonable semantic role combination is selected according to the joint learning result of syntactic parsing and semantic role labeling Traditional natural language processing tasks (part-of-speech tagging, syntactic parsing, semantic analysis, information extraction, etc.) are usually performed in order, that is, the latter task is performed on the basis of the previous task. For example, semantic role analysis is usually based on the results of syntactic parsing. Performing tasks in order is not the only option. If the joint learning of two or more continuous tasks can be realized, the tasks can use each other's information to benefit each other. A potential application of joint learning is syntactic parsing, semantic information annotation and analysis fusing valence structure.
Identifying rename refactoring opportunities based on feature requests
Published in International Journal of Computers and Applications, 2022
Ally S. Nyamawe, Khadidja Bakhti, Sulis Sandiwarno
To this end, this paper proposes a novel approach to automatically identify rename refactoring opportunities based on feature requests. Our approach differs from the existing approaches in the sense that it identifies rename refactoring opportunities based on the analysis of textual similarity between a to-be-implemented feature and the implementing source code. Whereas, existing approaches require complex semantic analysis of source code to uncover identifiers naming inconsistency, or analyzing lexicon, and grammatical structure of identifiers [16–19]. The proposed approach suggests a simple and effective means to identify whether an existing code entity (i.e. method) would demand rename refactoring or not to reflect its new responsibility (i.e. a requested feature). Therefore, the need for rename refactoring is promptly suggested to a developer as she modifies the code to implement the requested features. As a result, that would help in reducing effects of deferring refactoring such as the need for substantial efforts of cleaning the code later [20]. The evaluation results show that, the proposed approach can identify renaming opportunities on up to precision and recall. Such results are promising and further confirm the potentiality of exploiting high-level artifacts in identifying refactoring opportunities.
Hybrid Attention-based Approach for Arabic Paraphrase Detection
Published in Applied Artificial Intelligence, 2021
Cosma (2011) introduced a semantic-based approach for detecting and investigating source-code plagiarism using Latent Semantic Analysis (LSA). This technique was integrated with the PlaGate plagiarism detection tool to extract semantics between source code fragments. For estimating similarities, a parse-tree kernel method was applied to give the structure of the source code functionality. To detect plagiarism in students’ programming assignments based on semantics, multimedia e-learning-based smart assessment methodology was propounded by Ullah et al. (2020). It was processed as follows: Source codes were converted to tokens. Next, the Document Term Frequency Matrix (DTFM) was prepared and weighted according to the terms used in the source code. Then, they extracted the semantic features of each token using the LSA technique. It efficiently measured the semantic similarity without a parser requirement for any programming language.