Abstract: Apparatuses, systems, and techniques to use different resolutions of data as part of inferencing. In at least one embodiment, embeddings corresponding to different resolutions of data to be encoded by the embeddings may be obtained in order to be used as part of one or more neural networks that include the embeddings.

Abstract: A generative machine learning model, such as a convolutional neural network (CNN), can be trained with solutions from a topology optimization solver for a solution for a topology of a set of structures so that the generative machine learning model can generate a plurality of alternative designs for a structure that are alternative topology optimizations (for the structure) for a set of initial setup parameters. The generative model when being trained includes a generative network and a discriminator network. The generative model can be trained using outputs from a CNN autoencoder for densities and a CNN autoencoder for strain energies.

Abstract: Partial differential equations used to simulate physical systems can be solved, in one embodiment, by a solver that has been trained with a set of generative neural networks that operated at different resolutions in a solution space of a domain that defines the physical space of the physical system. The solver can operate in a latent vector space which encodes solutions to the PDE in latent vectors in the latent vector space. The variables of the PDE can be partially decoupled in the latent vector space while the solver operates. The domain can be divided into subdomains that are classified based on their positions in the domain.

Abstract: Partial differential equations used to simulate physical systems can be solved, in one embodiment, by a solver that has been trained with a set of generative neural networks that operated at different resolutions in a solution space of a domain that defines the physical space of the physical system. The solver can operate in a latent vector space which encodes solutions to the PDE in latent vectors in the latent vector space. The variables of the PDE can be partially decoupled in the latent vector space while the solver operates. The domain can be divided into subdomains that are classified based on their positions in the domain.

Abstract: A method and apparatus for determining spatial characteristics of three-dimensional objects is described. In an exemplary embodiment, the device receives a point cloud representation of a three-dimensional surface structure of a plurality of objects. The device may further generate a set of bins to represent the three-dimensional surface structure based on the point cloud representation, each bin corresponding to a spatial occupancy related to the point cloud representation, each bin including a respective type indicating a spatial relationship of the surface structures and a corresponding spatial occupancy of the bin. In addition, the device may encode the set of bins using a convolutional neural network. The device may further determine a classification for the spatial characteristic of the surface structures based on the convolutional neural network with the encoded set of bins.

Abstract: Partial differential equations used to simulate physical systems can be solved, in one embodiment, by a solver that has been trained with a set of generative neural networks that operated at different resolutions in a solution space of a domain that defines the physical space of the physical system. The solver can operate in a latent vector space which encodes solutions to the PDE in latent vectors in the latent vector space. The variables of the PDE can be partially decoupled in the latent vector space while the solver operates. The domain can be divided into subdomains that are classified based on their positions in the domain.