Artificial intelligence (AI), leveraging machine learning and deep learning, is increasingly seen as a key business enabler for today’s software systems including autonomous vehicles, cloud-based services, big data, blockchain, and IoT, among many other industry applications. AI-based systems are heavily reliant on software, with each system growing massively towards having thousands of software components with intricate interdependencies . The AI community has focussed primarily on algorithmic performance and functional aspects of AI-based systems, while neglecting other crucial software quality attributes, including but not limited to observability, maintainability, safety, security, and sustainability. Moreover, as AI-based systems often operate in dynamic, highly complex, and partially observable environments (e.g., cloud-based IoT), an additional challenge in these systems is to ensure that they can continue to deliver high quality of service despite unforeseen changes and uncertainty of operational environments.
The software engineering community has faced variations of these challenges in other domains like mobile embedded systems  and service-based systems . Evidence shows that one of the most critical success factors for the design and development of these systems has been to raise the level of abstraction by focussing on their software architecture [3, 4]. AI-based systems are software systems. Having an architectural perspective will help software engineers in achieving the sustainable development and deployment of efficient, reliable, and maintainable AI-based systems of the future. Industry experiences also demonstrate that AI-based systems, such as those that have machine learning components, get challenged by unmanaged software design and architectural concerns . Indeed, the architecture of a software system describes its structure in terms of software components, their interaction, and key quality attributes (e.g., safety, performance, maintainability, security) . The architecture of a software system has a significant influence on its key quality attributes and this is especially important for AI-based systems, where reliability, safety, resilience, and privacy are the “make or break” factors of many industrial sectors.
This special issue is a step forward towards positioning software architecture at the core of the ability to build and sustain the AI-based systems of the future. Among others, this special issue will contribute with methods, techniques, and tools for supporting software architects and engineers in designing the software architecture of their AI-based systems and ensure that they will meet their quality and functional requirements in addition to case studies where success and failures are shared. Moreover, there is recent research progress that demonstrates application of AI techniques like machine learning, reinforcement learning, deep neural networks to improve typical architecting activities, such as extracting architectural design decisions  and predicting the impact of architectural changes  .
These results are enabled by the large number of publicly available digital artifacts (e.g., GitHub repositories, Stack Overflow discussions, and other social platforms for developers), which allow researchers and practitioners to build rich datasets on which various AI approaches can be tested upon and subsequently validated. We believe that this trend will continue in the future and will attract greater scientific attention towards (i) discovering new architectural knowledge, (ii) building intelligent tools for software architects, and (iii) improving the quality and development process of future software systems.
We invite in this special issue articles with innovative and significant contributions to research at the intersection between software architecture and artificial intelligence. We accept submissions of original and previously unpublished papers and we especially encourage the submission of extended papers from the 14th European Conference on Software Architecture (ECSA 2020).
Surveys, (Systematic) Literature Reviews or Mapping Studies are out of the scope of the special issue and will be desk-rejected.
Topics of interest include, but are not limited to:
● Quality attribute concerns for AI-based systems
● Patterns and tactics for AI-based systems
● Experiences designing and deploying AI-based systems
● Analysis techniques for uncovering architecture issues in AI-based systems
● Verifying and validating AI models as part of system architectures
● Deployment and maintenance of AI-based architectures
● Method and techniques for improving the architecting process of AI-based systems
● Data and related challenges in architecting AI-based systems
● The impact of different algorithms, AI approaches and architecture challenges
● Data for AI research towards improving architecture methods and techniques
● Use of AI to improve architecture, design, conformance and quality
● Monitoring and sustaining AI-based systems architectures
● Iterative and incremental architecting of AI-based systems
● Impact of infrastructure concerns in architecting AI-based systems
● Architecting data/ML pipelines
● Submission Deadline: 9 November 2020
● Initial Author Notifications: 8 March 2021
● Initial Author Revisions Due: 10 May 2021
● Author Notifications for First Revision: 5 July 2021
● Final Author Revisions Due: 6 September 2021
● Final Author Notifications: 11 October 2021
Ivano Malavolta, Vrije Universiteit Amsterdam, The Netherlands (email@example.com)
Henry Muccini, University of L’Aquila, Italy (firstname.lastname@example.org)
Ipek Ozkaya, Carnegie Mellon University, Software Engineering Institute, US (email@example.com)
Paris Avgeriou and David Shepherd
Special Issues Editors
W.K. Chan and R. Mirandola
All manuscripts and any supplementary material should be submitted through the Elsevier Editorial System at http://ees.elsevier.com/jss. Follow the submission instructions given on this site. Please, select the article type as “VSI: SA&AI”, from the “Choose Article Type” pull-down menu during the submission process. All submitted papers should adhere to the general principles of the Journal of Systems and Software articles. Submissions have to be prepared according to the Guide for Authors, available on the journal website. Submitted papers must be original, must not have been previously published or be under consideration for publication elsewhere. The submitted paper must follow the format specified in the JSS Guide for Authors (https://www.elsevier.com/journals/journal-of-systems-and-software/0164-1212/guide-for-authors).
A submission extended from a previous conference version has to contain at least 30% new material. Please note that the papers from ECSA 2020 that are also invited for this special issue, are subject to the same rule. Authors are requested to attach to the submitted paper their relevant, previously published articles and a summary document explaining the enhancements made in the journal version.
For more information about the special issue, contact the guest editors.
Each submission will be reviewed by at least three expert reviewers. The guest editors, together with the Editors-in-Chief and the Special Issues Editors will make the final decisions.
 S. Malek, M. Mikic-Rakic, and N. Medvidovic. A style-aware architectural middleware for resource-constrained, distributed systems. IEEE Transactions on Software Engineering, 31(3), 256-272, 2005.
 E. Di Nitto, M. Di Penta, A. Gambi, G. Ripa, and M. L. Villani. Negotiation of service level agreements: An architecture and a search-based approach. In Proceedings of the International Conference on Service-Oriented Computing, 295-306. Springer, 2007.
 M. Shaw, and D. Garlan. Software architecture. Englewood Cliffs: Prentice Hall, 1996.
 L. Bass, P. Clements, and R. Kazman. Software architecture in practice. Addison-Wesley Professional, 2013.
 D. Sculley, G. Holt, D. Golovin, E. Davydov, T. Phillips, D. Ebner, V. Chaudhary, and M.l Young SE4ML: Software Engineering for Machine Learning (NIPS 2014 Workshop).
 D. Garlan. Software architecture: a travelogue. In Proceedings of the on Future of Software Engineering (FOSE 2014). Association for Computing Machinery, New York, NY, USA, 29–39. DOI:https://doi.org/10.1145/2593882.2593886
 M. Bhat, K. Shumaiev, A. Biesdorf, U. Hohenstein, and F. Matthes. Automatic extraction of design decisions from issue management systems: a machine learning based approach. In Proceedings of the European Conference on Software Architecture, 138-154. Springer, 2017.
 A. Shahbazian, D. Nam, and N. Medvidovic. Toward predicting architectural significance of implementation issues. In Proceedings of the 15th International Conference on Mining Software Repositories, 215-219, 2018.
 A. Horneman, A. Mellinger, and I. Ozkaya. AI Engineering: 11 Foundational Practices (https://resources.sei.cmu.edu/library/asset-view.cfm?assetid=633647), December 2019.
 J, Bosch, I. Crnkovic, and H. H. Olsson. Engineering AI Systems: A Research Agenda. CoRR abs/2001.07522, 2020.
 J. Camara, H.Muccini, and K. Vaidhyanathan. Quantitative Verification-Aided Machine Learning: A Tandem Approach for Architecting Self-Adaptive IoT Systems. In Proceedings International Conference on Software Architecture, ICSA 2020.