Vortrag von Aida Farahani

Wir starten das Jahr 2023 mit einem Vortrag, der thematisch an unsere Vortragsreihe "Maschinelle Lernverfahren zur Identifikation von Materialmodellen" aus dem letzten Semester anknüpft. Am

11.1.2023 um 14 Uhr in Raum 2/N106 und online über Zoom

spricht Aida Farahani (TU Chemnitz, Professur Künstliche Intelligenz) zu folgendem Thema:

Titel: Developing deep neural networks to represent 3D shape deformations

Abstract: Geometric deep learning is a promising approach to bring the power of deep neural networks to 3D data. Explicit 3D representations such as meshes are not easily combined with neural networks as there is no unique mesh to represent a single geometry. Point clouds, another explicit form, contain samples of surface points that have a varying and often huge number of dimensions that limit their use as neural network inputs. On the contrary, implicit representations such as signed distance functions (SDF) define a 3D shape as a continuous function that a deep neural network could approximate. The continuous property of implicit representations causes the algorithm to be independent of the size and topology of the shapes. In this talk, I demonstrate how deep neural networks, along with implicit representations, can precisely predict the deformation of a 3D structure after applying a specific load. The model is trained using a set of custom finite element simulations to generalize to unseen forces.

Der Vortrag wird auf Englisch gehalten. Wir sehen uns am 11.1.!