Seminar on From Systems Thinking to Multi-Agent Architectures for Manufacturing
Professor Kim Sang-Gook Massachusetts Institute of Technology (MIT), United States This seminar will be chaired by A/P Moon Seung Ki. | ||
| Seminar Abstract | ||
Despite decades of research and investment in Industry 4.0, manufacturing systems remain largely data-driven rather than knowledge-driven. A critical capability for smart manufacturing is enabling humans and AI agents to reason across machines, processes, and operations by integrating heterogeneous data, engineering knowledge, and operational context. Much of this tacit expertise resides in experienced engineers and operators, embedded in heuristics, pattern recognition, and engineering judgment. Guided by systems thinking, this presentation addresses two key challenges in manufacturing specific agentic AI: contextualizing multimodal industrial data and operationalizing tacit expert knowledge. An AI-based rationale elicitation agent is introduced to reconstruct expert decision rationale using Axiomatic Design, transforming implicit engineering knowledge into structured, machine-interpretable representations. Building on this, a multi-agent reasoning and orchestration architecture enables explainable, auditable, and trustworthy AI agents to collaborate with human experts in complex engineering decisions. The long-term vision is to enable knowledge-driven intelligent manufacturing systems that preserve institutional knowledge and support transparent, scalable AI reasoning and coordination. | ||
| Speaker's Biography | ||
|
