Machine Learning Model Development from a Software Engineering Perspective: A Systematic Literature Review

Lorenzoni, Giuliano; Alencar, Paulo; Nascimento, Nathália; Cowan, Don A.

doi:10.48550/arxiv.2102.07574

Cited by 5 publications

(6 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first group deals with data acquisition and data cleaning, while the second group is dedicated, among others, to training, evaluation, deployment, and monitoring. According to findings in Lorenzoni et al ( 2021 ), a very recent work giving a systematic literature review on the development of machine learning solutions, this process is the most comprehensive and accepted one in the literature.…”

Section: New Approaches and Methodsmentioning

confidence: 99%

A review on AI Safety in highly automated driving

Waschle¹,

Thaler²,

Berres³

et al. 2022

Front. Artif. Intell.

View full text Add to dashboard Cite

Remarkable progress in the fields of machine learning (ML) and artificial intelligence (AI) has led to an increased number of applications of (data-driven) AI systems for the partial or complete control of safety-critical systems. Recently, ML solutions have been particularly popular. Such approaches are often met with concerns regarding their correct and safe execution, which is often caused by missing knowledge or intransparency of their exact functionality. The investigation and derivation of methods for the safety assessment of AI systems are thus of great importance. Among others, these issues are addressed in the field of AI Safety. The aim of this work is to provide an overview of this field by means of a systematic literature review with special focus on the area of highly automated driving, as well as to present a selection of approaches and methods for the safety assessment of AI systems. Particularly, validation, verification, and testing are considered in light of this context. In the review process, two distinguished classes of approaches have been identified: On the one hand established methods, either referring to already published standards or well-established concepts from multiple research areas outside ML and AI. On the other hand newly developed approaches, including methods tailored to the scope of ML and AI which gained importance only in recent years.

show abstract

Section: New Approaches and Methodsmentioning

confidence: 99%

A review on AI Safety in highly automated driving

Waschle¹,

Thaler²,

Berres³

et al. 2022

Front. Artif. Intell.

View full text Add to dashboard Cite

show abstract

“…Also, there has been a lot of research in this direction. A comprehensive study of the available tools and applications for engineering projects is given by Mentor (Lorenzoni et al, 2021;Meyer, 2001). The study has offered a theoretical framework for understanding the nature and fieldwork of engineering projects.…”

Section: Introductionmentioning

confidence: 99%

Comparative Analysis of Data Driven Software Tools in Asian Countries: Advancing the Development of Engineering Projects

Shar,

Memon,

Maher

et al. 2023

ABBDM

View full text Add to dashboard Cite

A number of software tools are developed to improve the quality of the products developed in various fields of life, including engineering. The main purpose of the data-driven software tools the engineers design is to develop products on time with the requirements and specifications given by the customers. The most famous data-driven software tools specialized for engineering projects are AutoCAD, CATIA, Primavera, and ZW3D, which are widely considered good and appropriate tools in Asian countries for working on civil, electrical, and mechanical projects. Each software tool has unique characteristics; some of them are useful for civil engineering projects, while others are good for mechanical and electrical engineering projects. The main objective of conducting this research is to determine the best data-driven engineering tool. In this research, quantitative research methodology through survey techniques is employed for data collection. Data was collected from the already-existing users of the above-mentioned engineering tools. The resultant data is statistically analyzed. The results of the current study show that CATIA is better and preferable in terms of all identified features. However, with respect to robustness, Primavera is found to be slightly higher than CATIA. The results of this study aim to assist engineers, especially in Asian countries, in selecting appropriate tools for their data-driven engineering projects and thus be more effective and efficient in the engineering process.

show abstract

“…The software engineering field has widely applied various machine learning (ML) techniques to enhance organizations' efficiency and effectiveness. A recent study (Lorenzoni et al, 2021) revealed that ML models could be employed in different development stages of the software lifecycle, especially in the quality and analytics process. Software quality assurance saves time and money, as it helps developers discover errors and mistakes early in the development process.…”

Section: Introductionmentioning

confidence: 99%

“…Nevertheless, requirements problems, such as insufficient or ambiguous specifications, can cause misunderstandings during the requirement analysis stage. Research reveals that ML could help develop models that can enhance the quality of requirements classification tasks (Lorenzoni et al, 2021). However, selecting and implementing ML techniques to identify ambiguous requirements is not a trivial task.…”

Section: Introductionmentioning

confidence: 99%

Ambiguity Detection in Requirements Classification Task using Fine-Tuned Transformation Technique

Alharbi¹

2022

Machine Learning Techniques and Data Science Trends

View full text Add to dashboard Cite

The software requirement specification (SRS) document is essential in software development. This document influences all subsequent steps in software development. Nevertheless, requirements problems, such as insufficient or ambiguous specifications, can cause misunderstandings during the requirement analysis stage. This influences testing activities and increases the project’s duration and cost overrun risk. This paper represents an intuitive approach to detecting ambiguity in software requirements. The classifiershould learn ambiguous features and characteristics extracted from the text on a training set and try to detect similar characteristics from a testing set. To achieve this, this study experimented with two main approaches. The first approach is feature extraction, which uses the hidden states as features and trains asupport vector machine (SVM) classifier to assess software requirement ambiguity without modifying the pre-trained model. Unfortunately, this approach only identified 68% of the requirement ambiguity. The second approach is training an end-to-end model that updates the parameters of the pre-trained model. This approach enhanced the baseline results by 13%.

show abstract

Machine Learning Model Development from a Software Engineering Perspective: A Systematic Literature Review

Cited by 5 publications

References 30 publications

A review on AI Safety in highly automated driving

A review on AI Safety in highly automated driving

Comparative Analysis of Data Driven Software Tools in Asian Countries: Advancing the Development of Engineering Projects

Ambiguity Detection in Requirements Classification Task using Fine-Tuned Transformation Technique

Contact Info

Product

Resources

About