9 results about "Geometric space" patented technology

The invention relates to the field of knowledge graph data processing, and discloses an indoor space automatic layout and rendering method and system based on a design knowledge graph, and the method comprises the steps: maintaining an entity knowledge graph containing a variation path, and discretizing a continuous geometric space into an environment topological graph; executing node mapping and resource topology handshake verification based on subgraph isomorphism; and responding to the mapping deadlock, triggering recursive topology rewriting along the variation path to reconstruct the query sub-graph, and generating a configuration instruction. According to the method, continuous geometric search is converted into discrete atlas isomorphism and attribute propagation, the high-dimensional space layout calculation complexity is reduced, the deadlock problem under the over-constraint condition is solved by utilizing a recursive variation mechanism, and the configuration efficiency is improved. And the logic leakproofness and engineering compliance of the layout scheme are ensured.

The application discloses a geometric embedding knowledge graph question answering method based on a large language model, first, aiming at the structural heterogeneity of knowledge graph chain, level and cycle, entity and relationship embedding is learned in Euclidean space, hyperbolic space (Poincare ball model) and spherical space (unit sphere) respectively, and the triple score of the three geometric spaces is optimized by combining negative sampling and maximum interval loss. Secondly, in the reasoning stage, the large language model analyzes the problem and identifies the key entity, constructs the relevant subgraph, and generates the candidate path by using the breadth-first search or beam search; the path node is projected in three ways and spliced to calculate the attention weight, the fusion node representation and edge level fusion score are obtained, and the path score is accumulated. Finally, the structured prompt is formed by selecting the Top-K evidence path and its weight, contribution value, the large model is constrained to generate the answer and explanation, the traceability and accuracy are improved, the consistency is enhanced, and the illusion is reduced.

This invention relates to the field of thread turning technology, and discloses a thread turning cutting simulation method and system based on the analytical formula of a helical surface. The method includes: acquiring helical surface data during thread turning and constructing a geometric space model of the helical surface that can represent the tool sweep body; acquiring blank data and constructing a blank model using a tri-dexel model; for dexel segments parallel to the rotation axis, calculating the corresponding rotation angle and axial intersection point using the helical surface geometric space model, and calculating the axial cutting result of the tool based on the axial intersection point; for dexel segments perpendicular to the rotation axis, solving multiple solutions for the rotation angle and multiple intersection points using the helical surface geometric space model, selecting the effective intersection points as non-axial intersection points, and calculating the non-axial cutting result of the tool based on the non-axial intersection points. This invention can improve the efficiency and accuracy of thread turning cutting simulation.

The application relates to the field of knowledge graph data processing, and discloses an indoor space automatic layout and rendering method and system based on a design knowledge graph, which comprises the following steps: maintaining an entity knowledge graph containing a variation path, and discretizing a continuous geometric space into an environment topology graph; performing node mapping based on subgraph isomorphism and resource topology handshake verification; in response to a mapping deadlock, triggering recursive topology rewriting along the variation path to reconstruct a query subgraph, and generating a configuration instruction; the application converts continuous geometric search into discrete graph isomorphism and attribute propagation, reduces the high-dimensional space layout calculation complexity, solves the deadlock problem under over-constrained conditions by using a recursive variation mechanism, and ensures the logical rigor and engineering compliance of the layout scheme.

The present disclosure provides a multi-geometry space adaptive visual language modeling system for astronomical simulation. A two-stage training strategy is adopted to receive external images and user queries. First, the external images are input into a multi-geometry space prompt generation module, and the spatial features P E , H , S of three spatial geometry priors of Euclidean, hyperbolic and spherical are output. The spatial features are input into a multi-modal astronomical data input and alignment module together with the user query. After vector splicing, the input is input into a geometry adaptive backbone module, and finally the corresponding numerical value and reply are output through an astronomical task intelligent inference and output module. By integrating the representations of multiple geometry spaces such as Euclidean, spherical and hyperbolic into the visual language model architecture, the defects of existing visual language models that cannot handle complex astronomical physical phenomena due to their limitation to Euclidean space are effectively solved.

A construction method of spatio-temporal neural operator on Riemannian manifolds for complex geometries is disclosed, and relates to the field of machine learning technology, the method comprises the following steps: S1, according to a geometric space of an input spatio-temporal function and an output spatio-temporal function, solving a set of Laplacian eigenfunctions as basis functions, and then constructing a spatial dimension encoding module and a spatial dimension decoding module ; S2, solving a set of Fourier basis functions, and then constructing a temporal dimension encoding module and a temporal dimension decoding module −1; and S3, constructing a Laplace-Fourier nested kernel integration module, and constructing a spatio-temporal neural operator on Riemannian manifolds for complex geometries by serially connecting a plurality of kernel integration modules. The present disclosure adopts the construction method of spatio-temporal neural operator on Riemannian manifolds for complex geometries based on the above steps, and the mapping between two spatio-temporal functions defined on complex geometries is represented by constructing a parameterized model.

The invention discloses a mechanical arm generalization operation method and system based on mixed geometric representation pre-training, and belongs to the field of mechanical arm generalization operation. The method comprises the steps that a 3D-MOV data set containing an object, a fine-grained scene and desktop-level 3D point cloud is constructed, and a mixed geometric representation encoder is trained based on the pre-training; and after mapping to the hyperbolic space optimization learning, returning to the Euclidean space to complete self-supervised reconstruction. The scene observation point cloud is rendered into an RGB image, the RGB image is subjected to double-path distribution after being processed, strategy network fusion information is subjected to fine adjustment, and the action of the mechanical arm is predicted. Hyperbolic geometric representation is introduced into visual self-supervision pre-training based on 3D multi-view angles, 3D multi-view representation pre-training based on mixed geometric space is used for mechanical arm operation, the next operation action is effectively predicted, and the generalization operation capacity of a mechanical arm in different scenes is greatly enhanced.

Systems and methods for persistence of memory on a persistent cognitive machine (PCM) that uses a continuous, differentiable, cognitive manifold in geometric space to allow a computer to engage in human-like thought processes. A PCM with cognitive manifold performs cognition on a cognitive manifold in a continuous, differentiable, cognitive manifold in geometric space as opposed to probabilistic prediction in a discontinuous, anisotropic, and topologically fractured vector space. A mechanism inspired by magnetohydrodynamics is provided for coupling of typed spaces where thoughts on a cognitive manifold are structured as typed entities.