Machine learning (ML) is used increasingly in realworld applications. In this paper, we describe our ongoing endeavor to define characteristics and challenges unique to Requirements Engineering (RE) for ML-based systems. As a first step, we interviewed four data scientists to understand how ML experts approach elicitation, specification, and assurance of requirements and expectations. The results show that changes in the development paradigm, i.e., from coding to training, also demands changes in RE. We conclude that development of ML systems demands requirements engineers to: (1) understand ML performance measures to state good functional requirements, (2) be aware of new quality requirements such as explainability, freedom from discrimination, or specific legal requirements, and (3) integrate ML specifics in the RE process. Our study provides a first contribution towards an RE methodology for ML systems.
The quality of requirements engineering artifacts is widely considered a success factor for software projects. Currently, the definition of high-quality or good RE artifacts is often provided through normative references, such as quality standards, textbooks, or generic guidelines. We see various problems of such normative references: (1) It is hard to ensure that the contained rules are complete, (2) the contained rules are not context-dependent, and (3) the standards lack precise reasoning why certain criteria are considered bad quality. To change this understanding, we postulate that creating an RE artifact is rarely an end in itself, but just a means to understand and reach the project's goals. Following this line of thought, the purpose of an RE artifact is to support the stakeholders in whatever activities they are performing in the project. This purpose must define high-quality RE artifacts. To express this view, we contribute an activity-based RE quality meta model and show applications of this paradigm. Lastly, we describe the impacts of this view onto research and practice.
Context] Model-based Systems Engineering (MBSE) comprises a set of models and techniques that is often suggested as solution to cope with the challenges of engineering complex systems. Although many practitioners agree with the arguments on the potential benefits of the techniques, companies struggle with the adoption of MBSE. [Goal] In this paper, we investigate the forces that prevent or impede the adoption of MBSE in companies that develop embedded software systems. We contrast the hindering forces with issues and challenges that drive these companies towards introducing MBSE. [Method] Our results are based on 20 interviews with experts from 10 companies. Through exploratory research, we analyze the results by means of thematic coding.[Results] Forces that prevent MBSE adoption mainly relate to immature tooling, uncertainty about the return-on-investment, and fears on migrating existing data and processes. On the other hand, MBSE adoption also has strong drivers and participants have high expectations mainly with respect to managing complexity, adhering to new regulations, and reducing costs. [Conclusions] We conclude that bad experiences and frustration about MBSE adoption originate from false or too high expectations. Nevertheless, companies should not underestimate the necessary efforts for convincing employees and addressing their anxiety.
Model-based Systems Engineering (MBSE) advocates the integrated use of models throughout all development phases of a system development life-cycle. It is also often suggested as a solution to cope with the challenges of engineering complex systems. However, MBSE adoption is no trivial task and companies, especially large ones, struggle to achieve it in a timely and effective way. [Goal] We aim to discover what are the best practices and strategies to implement MBSE in companies that develop embedded software systems. [Method] Using an inductive-deductive research approach, we conducted 14 semi-structured interviews with experts from 10 companies. Further, we analyzed the data and drew some conclusions which were validated by an on-line questionnaire in a triangulation fashion. [Results] Our findings are summarized in an empirically validated list of 18 best practices for MBSE adoption and through a prioritized list of the 5 most important best practices. [Conclusions] Raising engineers' awareness regarding MBSE advantages and acquiring experience through small projects are considered the most important practices to increase the success of MBSE adoption.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.