Post by Anthony Massobrio

There is sometimes the perception of a gap between shapes optimized via Deep Learning on one side, and requirements of industrial CAD on the other side. As a result, frustrating feedbacks can be heard, such as the following: "You simulated an optimal theoretical shape... however, it cannot be constructed/industrialized". I would like therefore to highlight this advanced paper: https://lnkd.in/dxkqVuK. In short, for the rest of us in the CAE community: the approach presented in the paper "MeshSDF: Differentiable Iso-Surface Extraction" will help to bridge the aforementioned gap. This will have a very practical impact in the field of realistic generative design and promote more effective technical communication within organizations! Image courtesy Dr Pierre Baqué. #deeplearning #innovation #technology