Autonomous Experiments for Structural Design in 3d Printing
Tuesday May 18 2021 15:40 GMT
Please to join the live chat.
Why This Is Interesting
Active learning was used to optimize the mechanical properties of a 3d printed part with fewer experiments than other approaches
Combining physical simulation with active learning reduced the number of experiments required to discover the optimal design by an order of magnitude compared to active learning alone, and two orders of magnitude compared to an experimental grid search approach
The active learning system was used to power an experimental setup that designed, fabricated, and tested components completely autonomously
The results demonstrate promise for drastically improving efficiency and reducing the cost and time required to develop components that are optimized for specific applications