
Democratizing Model Discovery with Neural Networks
Ellen Kuhl, the Walter B. Reinhold Professor in Stanford's School of Engineering and Robert Bosch Chair of Mechanical Engineering, explains how neural networks are changing constitutive modeling of soft materials. She traces the history of model discovery, from hand built equations crafted by a small cadre of trained specialists to automated methods that can infer governing equations directly from data. Soft materials matter for autonomy, sustainability, and human health, and Kuhl argues that the traditional approach has kept accurate modeling out of reach for most researchers. She walks through how neural network based discovery widens access to constitutive modeling, letting a broader and more diverse set of scientists and engineers build reliable models without years of specialized training. The talk closes by connecting this shift to her Stanford graduate course on automated model discovery, situating the webinar as an entry point into the underlying methods.