Scientific ML promises 'near interactive' design optimization speeds: Page 2 of 2

May 22, 2020 //By Rich Pell
Scientific ML promises 'near interactive' design optimization speeds
Researchers at the Oden Institute for Computational Engineering and Sciences at the University of Texas at Austin say they are developing deep learning methods to dramatically reduce the cost and turnaround of conceptual design computations for complex energy systems.
building a DNN for the inverse problem, generating that training data could quickly become prohibitively expensive, since it entails computing a large number of optimal designs, each requiring many forward model solutions. This, say the researchers, is where they apply their new mathematical ideas, which exploit the low-dimensional structure underlying the inverse design problem to reduce the training data required and guide the training process.

"Our guiding principle is this: the simplest machine learning model that explains the optimal design data is the best," says Ghattas. "Not just to make the technology more widely accessible but because in doing so we avoid overfitting the model to the data, which leads to the inability to predict unseen data."

"It's about making aerodynamic design optimization accessible to a much broader set of industrial users," he says. "We know these technologies can enable faster, cheaper and more energy efficient design and production methods. So they should also be universally available."

Related articles:
GM, Autodesk using AI, 3D printing to make vehicles lighter
Cloud engineering simulation startup gets Intel, Google backing
DARPA selects Cadence for ML electronic design program
Computational fluid dynamics software enhances simulation, boosts productivity
Autonomous AI solution for chip design inspired by AlphaZero

Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.