The question of how AI will impact engineers in their professional lives as well as engineering as a profession remains a hot topic, with reactions to the technology ranging from awe to exhaustion. Whatever the case, there’s been no shortage of AI tools appearing with the promise of making engineers’ lives easier.

The latest example comes from MIT, where researchers have developed a machine learning model trained on tabular data that automatically identifies the most significant variables for improving system performance. According to the researchers, the model found the best solutions to various engineering problems, such as power system optimization, 10-100 times faster than conventional approaches.

“Modern AI and machine-learning models can fundamentally change the way engineers and scientists create complex systems,” said lead author of the published research, Rosen Yu, in a MIT press release. “We came up with one algorithm that can not only solve high-dimensional problems, but is also reusable so it can be applied to many problems without the need to start everything from scratch.”

The basic idea they came up with is to use a type of generative AI, a tabular foundation model, as a surrogate model inside a Bayesian optimization algorithm. As the name suggests, the foundation model is pre-trained on tabular data which, according to the researchers, makes it well-equipped to handle a variety of engineering problems.

“A car might have 300 design criteria, but not all of them are the main driver of the best design if you are trying to increase some safety parameters,” Yu explained in the same release. “Our algorithm can smartly select the most critical features to focus on.”

Pour lire la suite : https://www.engineering.com/using-ai-for-engineering-optimization/?spMailingID=190504&puid=3218464&E=3218464&utm_source=newsletter&utm_medium=email&utm_campaign=190504