New AI “copilot” helps engineers design better structures
Summary
- MIT researchers have developed an AI copilot that introduces human-in-the-loop feedback to topology optimization for structural design.
- The system uses machine learning to identify regions where designers can intervene, improving manufacturability and buckling loads by 39 percent.
- By allowing engineers to infuse domain knowledge as structures evolve, the approach aims to increase trust and industry adoption of algorithmic design.
Why it matters
Integrating human feedback into algorithmic topology optimization bridges the gap between high-performance computer-generated designs and real-world engineering constraints, increasing trust and adoption in architecture and civil engineering.