NGIoT Sustainability Workshop 2023_ Hans-Martin Heyn presentation
•Download as PPTX, PDF•
0 likes•4 views
Methods for requirements engineering in AIoT systems. Hans-Martin Heyn. NGIoT Sustainability Workshop – Use Cases, February 2023
Report
Share
More Related Content
Similar to NGIoT Sustainability Workshop 2023_ Hans-Martin Heyn presentation
Similar to NGIoT Sustainability Workshop 2023_ Hans-Martin Heyn presentation (20)
More from VEDLIoT Project
More from VEDLIoT Project (20)
Recently uploaded
Recently uploaded (20)
NGIoT Sustainability Workshop 2023_ Hans-Martin Heyn presentation
- 1. Methods for requirements engineering in AIoT systems 22. February 2023
- 3. 3 Problem definition • We needed to define an architectural framework, that supports both aspects of co- design. • The framework must support explicitly aspects of distributed systems (IoT) and AI system development. – Learning and data management • The framework also needed to be flexible enough to cover all current use cases, and new future use cases. – A special focus therefore lies on the support of non-functional requirements / quality views – Traceability of design decisions • A single reference architecture would have been too limiting for allowing the variety of (open call) use cases in VEDLIoT and beyond. Co-Design Integrated design Collaborative design Task 2.1 Heyn, H. M., Knauss, E., Muhammad, A. P., Eriksson, O., Linder, J., Subbiah, P., ... & Tungal, S. (2021, May). Requirement engineering challenges for ai-intense systems development. In 2021 IEEE/ACM 1st Workshop on AI Engineering-Software Engineering for AI (WAIN) (pp. 89-96). IEEE
- 4. 4 A compositional architecture framework for VEDLIoT
- 5. 5 How to compositional architecture framework • Step 1: Identify clusters of concern • Step 2: Identify levels of abstraction • Step 3: Add existing architectural decisions. • Step 4: Add missing architectural views. • Step 5: Add missing relations. • Step 6: Iterate if needed. Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ...
- 6. 6 Identify clusters of concern • Step 1: Identify clusters of concern • Step 2: Identify levels of abstraction • Step 3: Add existing architectural decisions. • Step 4: Add missing architectural views. • Step 5: Add missing relations. • Step 6: Iterate if needed. Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ...
- 7. 7 Identify clusters of concern Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ... Step 1: Identify clusters of concern
- 8. 8 Identify levels of abstraction Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ... Step 2: Identify levels of abstraction
- 9. 9 Add existing architectural decisions Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ... Step 3: Add existing architectural decisions
- 10. 10 Add existing architectural decisions Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ... Step 3: Add existing architectural decisions
- 11. 11 Add missing architectural views Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ... Step 4: Add missing architectural views
- 12. 12 Add missing architectural views Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ... Step 4: Add missing architectural views
- 13. 13 Add missing relations Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations . . . ... Step 5: Add missing relations
- 14. 14 Why? Task 2.1 • The architectural framework helps connecting different aspects of a system together. • It allows for “middle-out” development, i.e., existing design decisions are explicitly considered. • It allows to keep an overview over the necessary quality aspects, such as safety, security, ethical, or privacy aspects of the use case. • The traceability of design decisions allows for compliance with upcoming AI regulations. Heyn, H. M., Knauss E., & Pelliccione P. (2023). A Compositional Approach to Architecture Frameworks for distributed AI Systems. In Elsevier Journal of Systems and Software (JSS).
- 15. 15 Thank you for your attention.
- 16. 16 Co-Design of a system • Designing a complex and distributed system is a hierarchical process. – Several, sometimes highly specialized views allow for decomposition of the design task. – Requirements and architecture often co- evolve (Twin Peaks). • Developing complex system is a highly collaborative act between many stakeholders. Task 2.1
- 17. 17 Current approaches to architecture did not help • Providing the right learning setting / training data – No explicit views on the learning perspective of an AI system in common architecture approaches (Bosch et al., 2020, Muccini et al., 2021). • Monitoring solutions must be represented explicitly in the architecture – Some flaws can only be detected after deployment – Therefore, monitoring is needed to ensure functional, and non-functional aspects of an AI system (Bernadri et al., 2019). • New quality aspects arise, such as “explainability”, or “data privacy” – Depending on the use case certain a wide set of quality aspects can be relevant (Habibullah and Horkoff, 2019) – New stakeholders need to be included with their own views on the system (Vogelsang and Borg, 2019) Bosch, J., Olsson, H. H., & Crnkovic, I. (2021). Engineering ai systems: A research agenda. In Artificial Intelligence Paradigms for Smart Cyber-Physical Systems (pp. 1-19). IGI global. Muccini, H., & Vaidhyanathan, K. (2021, May). Software architecture for ml-based systems: what exists and what lies ahead. In 2021 IEEE/ACM 1st Workshop on AI Engineering-Software Engineering for AI (WAIN) (pp. 121-128). IEEE. Bernardi, L., Mavridis, T., & Estevez, P. (2019, July). 150 successful machine learning models: 6 lessons learned at booking. com. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining (pp. 1743-1751). Habibullah, K. M., & Horkoff, J. (2021, September). Non-functional requirements for machine learning: understanding current use and challenges in industry. In 2021 IEEE 29th International Requirements Engineering Conference (RE) (pp. 13-23). IEEE. Vogelsang, A., & Borg, M. (2019, September). Requirements engineering for machine learning: Perspectives from data scientists. In 2019 IEEE 27th International Requirements Engineering Conference Workshops (REW) (pp. 245-251). IEEE. Task 2.1
- 18. 18 A compositional architecture framework Rule 1: Clusters of concern shall contain architectural views with different levels of details of a certain aspect. Rule 2: Architectural views shall be sorted into levels of abstractions, according to their level of details. Rule 3: By using correspondence rules, it shall be possible to arrive at different architectural views of the system without encountering inconsistencies. Rule 4: Architectural views, and relations between them, shall be mapped to the next lower level of abstraction. Task 2.1 Heyn, H. M., Knauss E., & Pelliccione P. (2022, under review). A Compositional Approach to Architecture Frameworks for distributed AI Systems. In Elsevier Journal of Systems and Software (JSS).
- 19. 19 How to compositional architecture framework Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations . . . ... Step 5: Add missing relations
- 20. 20 How to compositional architecture framework • Step 1: Identify clusters of concern • Step 2: Identify levels of abstraction • Step 3: Add existing architectural decisions. • Step 4: Add missing architectural views. • Step 5: Add missing relations. • Step 6: Iterate if needed. Step 1: Identify clusters of concern Step 2: Identify levels of abstraction Step 3: Add existing architectural decisions Step 4: Add missing architectural views Step 5: Add missing relations ... ...