Dr. Yaniv Mordecai
Dr. Yaniv Mordecai
Overview.
Artificial Intelligence (AI) is rapidly transforming various industries and disciplines, and its integration with Systems Engineering (SE) represents a significant advancement in the way we design, develop, deploy, and operate complex systems. Systems Engineering, as an interdisciplinary approach, focuses on the holistic and systematic design, functionality, and performance of complex socio-technical systems. The integration of SE and AI has immense potential in two key pillars:
SE4AI and
AI4SE. Systems Engineering for Artificial Intelligence (
SE4AI) focuses on applying Systems Engineering principles to the development and deployment of AI-centric and AI-driven systems. Systems Engineering offers a structured, holistic approach to managing complex AI system development projects, ensuring that AI and AI-based systems are reliable, scalable, and aligned with user requirements. SE processes ensure that AI is integrated into complex systems and enterprise processes in a customer-driven manner, not just for technological innovation. Artificial Intelligence for Systems Engineering (
AI4SE) emphasizes the application of AI to enhance Systems Engineering processes. AI technologies such as machine learning, predictive analytics, automated reasoning, and generative AI are leveraged to improve the precision, efficiency, and adaptability of system design, management, and optimization. AI-driven tools and methods enable intelligent, robust, and resilient systems engineering processes.
The Israeli Association for Systems Engineering, INCOSE IL, held its international conference on Systems Engineering on June 3-4, 2024, with a focus on AI as an emerging technology and theme, particularly the integration of AI and SE. The second day of the conference featured four deep-dive tutorials by internationally recognized experts on state-of-the-art AI applications in complex systems at various scales (see Figure 1). This paper summarizes the highlights of the tutorials across the various system scales. The tutorials day program is available here:
https://incoseil.org/activity_page/incoseil2024/day2.htmlInspired by the four talks, this summary explores how AI technologies such as machine learning, natural language processing, and autonomous agents cab be incorporated into a variety of systems architectures. It highlights the benefits, challenges, and future prospects of this synergy, illustrating how AI can optimize system performance, improve decision-making, and enable more adaptive and intelligent systems. Our four speakers discussed a broad range of systems-related aspects of AI, including enterprises – such as corporates and government agencies; systems of systems (SoS) – such as mission operation centers and communication networks; socio-technical systems – such as information management and business control systems, and cyber physical systems – such as sensors, radars, and IoT devices. This breadth of scales, typical of the various roles that systems engineers and architects play in various organizations and ecosystems, was particularly curated to provide a holistic and comprehensive overview of AI-based systems challenges and opportunities. Figure 2 captures this spectrum of AI applications in complex system architectures.
Figure 1. Screen Capture with the Introduction of our Keynote Speakers
Figure 2. AI in Complex System Architectures Covered by the Speakers in the Tutorials Day
AI for the Enterprise – Driving Adoption with Stakeholders in MindLarge enterprises and government agencies can reduce costs, improve effectiveness, and achieve better business outcomes With AI powered systems enabled by systems engineering. Systems Engineers are often assigned to lead and accelerate the adoption, integration, and scaling of innovative and disruptive technologies in advanced systems. Thus, enterprise and system architects often own the adoption and utilization of trusted artificial intelligence and machine learning in large-scale socio-technical systems and system development programs.
In the tutorial on
Driving AI Adoption in Large Enterprises using Systems Engineering by
Dr. Jyotirmay Gadewadikar, Aparna Durvasula, and Sean Hwang from the MITRE Corporation, we learned how to use systems perspectives to expand AI above and beyond individual use cases at the enterprise level. Key takeaways included: a) how to identify AI needs and capabilities and generate synergetic themes from these needs and capabilities; and b) how to accelerate AI initiatives at enterprise level using the awareness of system dependence and capability sequencing.
We understood how enterprises, corporates, and agencies perceive the strengths, weaknesses, opportunities, and threats (SWOT) of AI for the business, and how they consider its adoption. Without a systems lifecycle approach the adoption might fail and end up incurring excessive costs, loss of trust, and ineffective adotion. We saw how the systems lifecycle processes and steps address the questions raised by the introduction of AI. We went through an interactive exercise in which we had to map business needs to key technical initiatives and then to map the key technical initiatives to AI-enabled thmes, such as personalized customer/user experience, error reduction and increased accuracy, smart decision making, AI-driven chatbots and communication, and AI-boosted process automation, efficiency, productivity. We also saw how model-based systems engineering (MBSE) can be used for creating value stream maps, multi-attribute project selection based on criteria like impact, feasibility, alignment with goals, return on investment, and stakeholder buy-in, and how the framework can be used for AI governance, policy making, resource allocation, progress monitoring, control, ethics, operational concepts, etc.
Figure 5. Screen Capture from Dr. Jyo Gadewadikar’s Tutorial on Driving AI Adoption in Large Enterprises
AI for System of Systems – Achieving Impact Across System NetworksComplex commercial and military systems with emerging behaviors require decision support well beyond the capacity of human reasoning alone. To fill this gap, Artificial Intelligence and Machine Learning (AI/ML) can assist Systems Engineering (SE) with respect to operationally realizing the full potential (e.g., speed, scale, and accuracy) of the capabilities offered by these systems.
Dr. Paul Hershey from Raytheon spoke about the
Role of AI-ML in Systems-of-Systems Engineering to Support Human Decision Making. This presentation provided a review of the concepts of complex systems and emergent behavior, discusses the fundamentals of AI/ML, and then ties these together to demonstrate the role of AI/ML in systems through specific commercial and military use cases. These use cases include: 1) Object Recognition and Detection Enhancement via Reinforcement Learning Yield, 2) Disaggregated Distributed AI Chat Enabler, and 3) Distributed Disaggregated Communications via Reinforcement Learning and Backpressure.
We covered basic definitions of core concepts in systems engineering, as well as AI, ML, Data Science (DS), and Big Data. We also reviewed the SCOAR approach – a Systems Course-of-Action which consists of plan initiation, mission analysis, COA development, COA analysis, CAO comparison, COA selection, and plan order development – all of which are complemented and enriched through centralized analysis. We saw a variety of examples of the implementation of SCOAR for the integration of AI-based decision automation in defense and homeland security applications. We learned about ORDERLY – Object Recognition and Detection Enhancement via Reinforcement Learning Yield – a proprietary Raytheon AI technology which translates massive amounts of data into actionable insight (see Figure 6). We also learned about the use of reinforcement learning (RL) for chat enrichment and enhancement chat-based mission-critical systems and mission operation centers and for communications management and facilitation across distributed and disaggregated ecosystems.
Figure 6. Screen Capture from Dr. Paul Hershey’s Tutorial on the Role of AI in Decision Support in Systems of Systems
AI-Based Systems – Building Value-Added GenAI-Driven SolutionsGenerative Artificial Intelligence (GenAI) and Large Language Models (LLMs) are changing the world, and show a promising opportunity to revolutionize the engineering of complex systems. A systems approach is critical for conceiving the AI module as a component in a bigger system that intends to support its operators in achieving various effects and goals. Better understanding of LLMs, how to integrate LLMs into comprehensive solutions, and what are the use cases for LLM as part of bigger systems.
In
Integrating Generative AI with Traditional Data Science - A Systems Approach Dr. Yaniv Mordecai (yours truly) showed how GenAI and especially LLMs can be used in complex systems, mainly in conjunction with traditional data analysis methods, for accomplishing various integrated functionalities.
We discussed the basic concepts of LLMs and their capabilities, and how they differ from similar technologies like ChatBots and search engines. We reviewed user and stakeholder perceptions of LLM’s added value and expectations for controllability, observability, explainability, etc. We also reviewed the limitations of LLMs and where they fall short in performance, resource utilization, latency, determinism, repeatability, etc. We did also see a variety of useful applications of LLMs for tasks like fact finding, text generation, translation, summary, classification, similarity analysis, and code generation. We discussed the integration of LLMs and GenAI with traditional data science and machine learning methods, and the incorporation of LLMs into complex solution architectures and how these models can be deployed, and discussed the use of LLM for requirements forecasting and classification .
Figure 4. Screen Capture from Dr. Yaniv Mordecai’s Tutorial on LLM-Based Solution Architectures
AI for the Edge – Critical Components, Sensors, Radars, & Edge DevicesRadar is one of the most complicated complex systems, involving a variety of technologies, and playing a critical role in any system in which it is integrated. Radar system design plays a crucial role in determining the performance and effectiveness of radar technology across various applications. With the emergence of artificial intelligence (AI), there is a growing opportunity to revolutionize radar system design by integrating AI techniques into the design process.
Dr. Niv Regev spoke about
AI-Driven System Design for Advanced Radar Technology and showed how critical components like radars and other types of sensors and actuators can benefit from AI in both design and integrated analytics.
We explored the transformative potential of AI in radar system design, focusing on the benefits, challenges, and practical applications of AI-driven design approaches. We discussed Radar System design, including Antenna and RF design, signal processing, tracking & fusion, and target classification. We saw how AI can be integrated into each and every one of these design and analysis processes through methods like AI(DED) System Design, Machine Learning and Neural Networks for behavioral modeling and optimization, Generative design and evolutionary algorithms for design space exploration, and Transfer Learning and Meta-Learning methods that accelerate design convergence.
Figure 3. Screen Capture from Dr. Nir Regev's Tutorial on AI-Driven System Design for Advanced Radar Technology
ConclusionThis comprehensive summary captures the essence of the Tutorials Day at the INCOSE IL 2024 International Conference on Systems Engineering, showcasing the transformative potential of integrating AI and SE across different system scales—enterprise, systems of systems (SoS), solution architectures, and critical components—in various complex systems and applications. We look forward to seeing many applications and developments in these contexts and to showcasing them in future INCOSE IL events. We hope that this day provided a holistic perspective, inspiration, and useful resources for all attendees. Tutorial recordings would be made available to INCOSE IL members on the
www.incoseil.org website.
I would like to thank the INCOSE IL Secretariat, Mr. Moshe Salem, Kfir and Maya, and the rest of the team for facilitating the conferences in general and the tutorials day, and to the INCOSE IL leadership and board for entrusting me with organizing and chairing this tutorials day. I hope that I was able to curate and showcase high-quality, engaging and inspiring content that would benefit INCOSE IL members in their various endeavors.
Speaker Biographies
Dr. Jyotirmay (Jyo) Gadewadikar is the Chief Scientist for AI Integration and Systems Engineering at the MITRE Corporation, a not-for-profit corporation committed to the public interest, operating federally funded R&D centers on behalf of U.S. government sponsors. Dr. Gadewadikar is overseeing the Systems Engineering Innovation Center’s Artificial Intelligence capabilities, amplifying the integration of AI-enabled technologies into Systems of Systems, and integrating Machine Learning algorithms into systems to improve performance and achieve significant gains in system utility. He previously led conversational AI and data science initiatives at Deloitte and Accenture. He holds a Ph.D. in Autonomous Vehicles from the University of Texas in Austin (2007) and a M.S. in System Design and Management from Massachusetts Institute of Technology (2014). He is a certified Project Management Professional (PMP).
Dr. Paul C. Hershey is presently in his 20th year with RTX, Dulles, Virginia, where he is a Principal Engineering Fellow focusing on data analytics, AI/ML, and modeling and simulation. He has published 41 patents (granted), along with 9 patents pending with the US Patent Office, and over 75 peer-reviewed technical publications. Previously, he was an adjunct professor at George Washington University where he also served on the Curriculum Advisory Board. He is an IEEE Fellow and serves on technical program committees for the IEEE International Systems Conference (also on the conference steering committee) and the IEEE International System of Systems Engineering Conference (also an industrial liaison). He is a Distinguished Lecturer on data analytics for the IEEE Systems Council. Dr. Hershey received the A.B. degree in mathematics from the College of William and Mary, Williamsburg, VA, USA, and the Ph.D. and M.S. degrees in electrical engineering from the University of Maryland, College Park, MD, USA. His Ph.D. research, sponsored by IBM, created a novel information collection, analysis, and decision system that resulted in direct customer sales.
Dr. Nir Regev is an expert in AI and radar signal processing, specializing in FMCW and Pulse Doppler radars. His work focuses on the development of algorithms for target detection and tracking in MATLAB, Python, and C++. Dr. Regev has a deep background in statistical signal processing and applies his expertise to both radar and computer vision technologies and their intersection with AI. He is proficient in various AI domains including generative AI and deep learning for semantic segmentation. Dr. Regev is the founder of AlephZero.ai and serves as an Adjunct Professor in Electrical and Computer Engineering at Cal Poly Pomona. With a strong record in both academia and industry, Dr. Regev leads projects that bridge theoretical research and practical application, enhancing real-time technology solutions. Dr. Regev holds a Ph.D., MS and BS in Electrical Engineering from Ben-Gurion University of the Negev, Beer Sheva, Israel.
Dr. Yaniv Mordecai is a senior teaching fellow at Tel-Aviv University, Tel-Aviv, Israel, a Senior Research Scientist at Amazon, Bellevue, Washington, USA, and a former Technion-MIT Post-Doctoral Fellow at MIT’s Engineering Systems Laboratory (2019-21). He holds a Teaching Certificate from MIT (2020), Ph.D. in information systems engineering from Technion – Israel Institute of Technology, Israel (2016), and M.Sc. (2010, cum laude) and B.Sc. (2002) in industrial engineering & management from Tel-Aviv University, Israel. His research interests include model-based systems engineering, model analytics, cybernetics, interoperable systems, decision automation, operations research, and applications of artificial intelligence for geospatial optimization. Dr. Mordecai is a senior member of IEEE, member of INCOSE and INFORMS, and Board Member of the Israeli Society for Systems Engineering – INCOSE_IL. He won multiple international research awards and is a recognized international expert in modeling and simulation of complex systems.
