31 results about "Function representation" patented technology

The invention discloses an intelligent dynamics rapid simulation method. The invention relates to the technical field of finite element simulation, and solves the problem that real-time prediction of high-dimensional nonlinear structural dynamics and online inversion of impact acceleration cannot be met in the prior art. According to the invention, based on a deep learning algorithm, structure-level reconstruction of an existing operator learning normal form is realized through a mapping thought from a graph structure depth operator function to a function. According to the method, the double-tower type branch-trunk topology on which DeepONet leans is thoroughly got rid of; according to the method, a space correlation mode and a time evolution rule are synchronously coded in a hidden space and are directly projected to a function at a target moment to be expressed without following a frequency domain kernel integral path of the FNO instead of an end-to-end serial progressive single calculation path. According to the design, the function-to-function micro-mapping capability is reserved on the semantics of operator approximation, and induction deviation caused by structure splitting of a frame is overcome in a graph domain time sequence task.

The present invention discloses a method for generating a printable 3D model from a single photo based on a generative adversarial network, comprising the following steps: S1, obtaining a single input image and preprocessing the input image; S2, inputting the preprocessed image into an encoder, extracting multi-scale features, and outputting a deep feature representation; S3, generating a 3D model using a conditional generative adversarial network and optimizing the 3D model quality through comparison; S4, refining the 3D model at multiple scales using an implicit neural representation network and converting it into a continuous function representation; S5, optimizing the 3D model using an improved conditional diffusion inversion module, combining multi-scale discrete wavelet transform and rendering feedback correction; S6, optimizing the feature extraction, 3D model generation, and refinement processes using an end-to-end joint training system; and S7, post-processing the 3D model to output a 3D model that meets printing requirements. By integrating generative adversarial networks and implicit neural representation technology, the present invention achieves the rapid generation of high-precision 3D models with rich details suitable for 3D printing from a single photo.

The present invention relates to a high-speed maneuvering target tracking method based on a continuous space-time trajectory function random process, and relates to the field of target tracking. The method comprises: modeling the continuous space-time trajectory of a high-speed maneuvering target with an unknown motion model as a random process based on a continuous-time trajectory function; modeling the mean function of the target trajectory random process as a continuous space-time trajectory function; and fitting and estimating the continuous space-time trajectory function to output the target state. The method of the present invention can achieve higher-precision real-time target tracking in complex maneuvering scenarios, overcoming the shortcomings of existing random process target tracking methods, which often add a constant mean function and rely solely on covariance kernel function representation for trajectory estimation.

This invention provides a method and related apparatus for establishing a nonlinear mathematical model of a hybrid DC transmission system considering discontinuous characteristics. The method includes: representing the dynamic processes of discontinuous components in the main and control loops of the hybrid DC transmission system using a step function; approximating the step function as a continuous function using a linear combination of hyperbolic tangent functions to obtain a continuous expression; obtaining modular nonlinear differential equations for the rectifier, inverter, and AC / DC networks, and replacing the discontinuous dynamic parts involved in the equations with the continuous expression; and interconnecting the modules according to their input-output relationships to obtain a complete nonlinear mathematical model of the hybrid DC transmission system. This invention establishes a nonlinear mathematical model of a hybrid DC transmission system that accurately reflects discontinuous dynamic characteristics by employing the continuous approximation method using step functions and hyperbolic tangent functions, thereby improving the accuracy and practicality of system modeling.

The application discloses a kind of based on three periodic minimal surface double-fluid porous heat exchanger optimization and design method, comprising: step 1: the function representation of double-fluid porous heat exchanger;Step 2: establish optimization mathematical model;Step 3: optimization process and improvement optimization solver.The application is mainly aimed at the problem of enhancing the heat exchange efficiency of double-fluid heat exchanger, based on the porous structure of three periodic minimal surface has natural heat exchange advantage, for example, the specific surface area ratio is increased to increase the contact area of double-fluid, full connectivity and high smoothness can reduce flow resistance;In optimization model, the wall thickness between double-fluid is directly controlled by geometric parameter variable, to avoid adding complex wall thickness constraint;Secondly, the optimization solver of adding monotone programming operator is more suitable for the optimization of complex structure, and accelerates the convergence speed of optimization;In addition, the function representation of porous heat exchanger can guarantee the accuracy of model, good smoothness, connectivity and self-supporting are also conducive to additive manufacturing.

A path planning method for lattice three-dimensional (3D) printing includes generating a 3D model; slicing the 3D model to generate a 2D contour; determining a region of the 2D contour; and generating a path. The 3D model is represented by an explicit mesh or implicit function. Slicing the 3D model to generate the 2D contour includes generating the 2D contour of the 3D model defined by an explicit or implicit function. For a 3D model defined by an explicit function, a 2D contour is generated by calculating intersections between slice layers and triangular meshes. The determination of the region comprises dividing the lattice structure or the thin-wall structure into two regions including a thin wall and a wall connection. Geometric features of the two regions are defined by the tilt angle and the plate thickness.

The invention discloses a neural operator modeling method with a structure adjustment capability and a multi-stage optimization mechanism. The neural operator modeling method is suitable for efficiently solving a complex partial differential equation under a small sample condition. According to the method, a primary function regulation and control network is provided, a traditional fixed primary function is replaced, and the function representation capacity is dynamically adjusted through a structure guiding module; in the aspect of feature fusion, a collaborative modulation mechanism is introduced, and deep semantic integration of input conditions and space coordinates is realized. The training process adopts a staged design, and comprises structure adaptive preheating, self-supervised reconstruction and physical consistency collaborative optimization, so that the precision, stability and interpretability of a solution are effectively improved. Experimental verification shows that the method has good generalization and physical rationality in the problems of Darcy flow, Poisson equations and the like, and is suitable for scientific calculation tasks of complex boundaries, multi-scale modeling and the like.

The invention discloses a table structuring method, related equipment and a program product, and relates to the technical field of computer vision and table recognition. The method comprises the following steps: inputting a to-be-structured table image into a configured control parameter prediction model to obtain a group of control parameters respectively corresponding to each row and column line in the table image; generating a curvilinear function corresponding to each row and column line based on the group of control parameters corresponding to each row and column line and a preset curvilinear function containing a group of unknown control parameters; and generating a structured table of the table image according to the curvilinear function corresponding to each row and column line. The model directly predicts the function representation of each row and column line, and eliminates the additional error caused by the fitting process, thereby improving the accuracy of the structured table finally generated by the curvilinear function of the row and column lines.

A method of initial residual hypothesis derivation for joint sign prediction of transform coefficients of a residue block. According to this method, an initial residual hypothesis is derived by applying inverse transformation to de-quantized transform coefficients with the set of sign-predicted transform coefficients set to zero or by applying the inverse transformation to the de-quantized transform coefficients corresponding to non-sign-predicted transform coefficients only to form a first part of the initial residual hypothesis, and adding a second part by summing up products of normalized template functions and absolute values of respective sign-predicted transform coefficients for the set of sign-predicted transform coefficients. The normalized template functions represent approximated entry values of inverse transformation matrixes along top and left block boundaries for corresponding transform coefficient indexes associated with the respective sign-predicted transform coefficients. The initial residual hypothesis is used to derive sign prediction for the sign-predicted transform coefficients.

The present invention discloses a nonlinear constraint coupling-driven implicit modeling optimization method, system, and memory. The method first obtains nonlinear constraint data for a target geological interface. Then, based on the initial control points of the target geological interface, an initial implicit function representation is generated using a Hermite radial basis function. Next, based on the data type of the nonlinear constraint, a nonlinear coupling-driven optimization objective function for the implicit function representation of the target geological interface is obtained. This objective function is optimized using an implicit gradient flow algorithm to obtain a geological interface reconstruction model and a globally optimized target geological interface model. The present invention also provides a computer-readable memory containing a computer program that implements the aforementioned nonlinear constraint coupling-driven implicit modeling optimization method. The memory is readable and executable, and together with a programmable logic controller and a processor, forms an optimization system. This system achieves high-precision modeling of three-dimensional geological interfaces by fully coupling a large amount of nonlinear constraint data for the geological interface and constraining the implicit modeling framework.

The present invention discloses a method for simulating incompressible viscous fluids based on signed distance function representation, and relates to the technical fields of numerical calculation and numerical simulation. The method comprises spatially dividing a calculation area into Cartesian grids; using signed distance functions to represent virtual points in the grid at the solid wall boundary of the calculation area, calculating the perpendicular point from the virtual point to the boundary of the geometric area using a perpendicular calculation algorithm based on signed distance function representation, solving the differential information of the equation at the perpendicular point using an explicit Inverse Lax-Wendroff technique at the perpendicular point, obtaining the grid function value at the virtual point using the differential information, and simultaneously solving the algebraic relationship between the virtual point value and the internal point value, and auxiliary matrix assembly; and using the Crank-Nicolson format to advance the time step forward to realize the numerical simulation of incompressible viscous fluids based on SDF. The present invention realizes complex boundary processing while ensuring the accuracy of the algorithm.

A path planning method for three-dimensional (3D) printing of a lattice includes generating a 3D model; slicing to generate 2D contours of the 3D model; determining regions of the 2D contours; and generating a path. The 3D model is represented by explicit mesh or implicit functions. The slicing to generate 2D contours of the 3D model includes generating 2D contours of the 3D models defined by either explicit or implicit functions. For the 3D model defined by explicit functions, the 2D contours are generated by calculating intersecting points between a slicing layer and triangle meshes. The determining regions includes dividing lattice structures or thin-walled structures into two regions including thin-walls and wall joints. Geometric characteristics of the two regions are defined by inclination angles and sheet thicknesses.

The invention discloses a three-dimensional shape generation method fusing image retrieval and sketch regulation, which comprises the following steps of: firstly, mapping an input sketch to a shared intermediate feature space by constructing a sketch-image mapping network, and retrieving a plurality of images which are closest to sketch semantics from an image library based on cosine similarity; secondly, a pre-trained CLIP visual encoder is adopted to carry out feature extraction on the retrieved image, fusion is carried out through a maximum pooling mode, and final image features are obtained; introducing a shape adapter with a cross attention mechanism into a de-noising network of a diffusion generation model, and fusing the geometric features of the input sketch into the de-noising network to obtain a three-dimensional latent vector consistent with the input sketch in structure and semantics; and finally, a decoder of the three-dimensional vector quantization variational auto-encoder is reduced into a three-dimensional symbol distance function for representation. According to the method, the quality of the three-dimensional shape generated by sketch driving in the aspects of structural consistency and local detail reduction is improved, and the method has high generalization ability and practical application value.

The invention discloses a profile steel section parameterization generation method based on Autolisp, and belongs to the technical field of mechanical design. Comprising the following steps: 1, establishing a database comprising different section steel section parameters; 2, representing drawing processes of different types of section steel sections through a drawing function; 3, setting keywords for obtaining form parameters in the drawing function; 4, programming by using a language of a DLC dialog box in CAD to form interactive interfaces corresponding to different types of section steel sections, and drawing the section steel sections by selecting the keywords in the step 3; 5, a main function is defined as a drawing command in CAD, different driving functions are selected through the main function, and the driving functions are used for calling interaction interfaces of different types of section steel sections in the step 4. According to the method, automatic drawing is achieved by programming section parameters of different types of section steel through Autolisp, manual consulting and section drawing operation is not needed any more, and drawing time and drawing energy are reduced.

The application discloses a kind of robust sparse bayesian two-dimensional boresight angle estimation method.It is characterized in that: the patent method is by introducing a new auxiliary angle, two-dimensional DOA estimation problem is converted into two one-dimensional DOA estimation problems.Through solving two sparse reconstruction problems containing amplitude and phase errors, the auxiliary angle and its corresponding pitch angle are estimated respectively, and according to the relationship between the three angles, the azimuth angle estimation and angle automatic matching process are completed simultaneously.For the sparse reconstruction problem containing amplitude and phase errors, the expectation maximum algorithm is used to derive the estimation expression of all unknown parameters, obtain a new spatial spectrum function representation method, and finally obtain the angle information.The patent method can effectively complete the angle automatic matching process, and improve the azimuth estimation accuracy and angle resolution capability, so it has high application value in practical engineering.

The invention relates to the technical field of image super-resolution reconstruction, and discloses a ridgelet-based image super-resolution reconstruction method, which comprises the following steps of: modeling an input low-resolution image into a continuous function, directly predicting a pixel value from a space coordinate through a multi-layer perceptron network, and obtaining continuous function representation of the whole image; periodic features of space coordinates are extracted, and the expression ability of a periodic structure is improved through a periodic coding enhancement network of Fourier transform; analyzing the fractal dimension of the ridgelet structure of the original image, and introducing a fractal dimension retention constraint according to the fractal dimension; constructing a high-order derivative regularization term near the ridgelet; network parameters are optimized through end-to-end training, and super-resolution images of any scale are generated; according to the invention, the image is represented as a continuous function, so that the image can be sampled under any resolution, and amplification of any scale is realized without retraining the model.

The present invention is directed to a computer implemented method for controlling a finite set of components which are fluidly connected to a common compressed air distribution system, the method comprising iteratively repeating the steps of: - receiving prediction data for said compressed air distribution system; - receiving characterising data for each component of said set of components; - determining one or more sets of continuously differentiable functions, wherein each of said sets of functions represents a unique sequence of operation of the components in said set of components; - selecting an optimal set of functions from said one or more sets of continuously differentiable functions; wherein the unique sequence of operation represented by said optimal set meets said prediction data; - deriving configuration data for said set of components from said optimal set of functions; - configuring each component of said set of components based on said configuration data.

A program according to one embodiment of the present disclosure causes a computer to execute: a first input procedure for inputting a first nonterminal node and a second nonterminal node on a zero-suppressed binary decision diagram; a second input procedure for inputting at least designation of a first function representing a condition to be satisfied by a set pair of a set included in a set group represented by the first nonterminal node and a set included in a set group represented by the second nonterminal node; and a calculation procedure for executing a set group calculation of the set group represented by the first nonterminal node and the set group represented by the second nonterminal node, on the basis of the first nonterminal node, the second nonterminal node, the first function, and a template of an algorithm for achieving the set group calculation.

The application relates to a trajectory prediction method and device, electronic equipment and a computer readable storage medium, and belongs to the technical field of computers. The method comprises the following steps: acquiring a semantic map corresponding to a prediction target and a historical observation trajectory sequence corresponding to the prediction target, the historical observation trajectory sequence representing the trajectory position of the prediction target at each historical moment; then, the historical observation trajectory sequence corresponding to the prediction target is represented by sequence coding to obtain sequence coding features; a function representation corresponding to the semantic map is generated based on the semantic map corresponding to the prediction target, the function representation corresponding to the semantic map representing the mapping relationship between semantics and trajectory positions; then, trajectory prediction is performed based on the sequence coding features and by using the function representation corresponding to the semantic map to obtain a future prediction trajectory sequence. The application relates to a trajectory prediction method and device, electronic equipment and a computer readable storage medium, which can improve the vehicle trajectory prediction accuracy.

The invention discloses a cross-scale feature fusion method based on quantum inspiration, and relates to the technical field of target detection, and the method comprises the steps: carrying out the up-sampling operation of a cross-scale feature map; mapping the original spatial domain features to a quantum inspiration domain through amplitude coding and phase coding; performing feature preliminary fusion by simulating quantum interference; predicting a spatial offset based on the similarity; performing resampling operation on the original feature map by using the space offset to realize feature alignment; using a quantum observation mechanism to convert the complex value features into real number representation; and performing channel dimension splicing on the aligned original feature map and the feature map after quantum interference fusion, and recovering the number of channels through convolution operation to obtain a fused cross-scale aggregation representation feature map. According to the invention, through the complex value quantum wave function representation and interference fusion formula, the feature fusion which is finer and more accordant with semantic logic is realized, semantic dislocation and cross-scale intra-class difference are reduced, and the detection precision under a small target and a complex scene is effectively improved.

The application discloses a table structuring method, related equipment and program products, and relates to the technical field of computer vision and table recognition. The application inputs a table image to be structured into a configured control parameter prediction model, obtains a group of control parameters corresponding to each row and column line in the table image, generates a curve function corresponding to each row and column line based on the group of control parameters corresponding to each row and column line and a preset curve function containing a group of unknown control parameters, and generates a structured table of the table image according to the curve function corresponding to each row and column line. The application directly predicts the function expression of each row and column line, eliminates the additional error caused by the fitting process, and thus improves the accuracy of the structured table generated by the curve function of the row and column line.

The invention provides a chromatographic data processing method and a chromatographic data processing system, which can shorten the time for separating peaks derived from a plurality of components overlapped on a chromatogram. In the chromatogram data processing method and system, first data and second data are prepared, and information relating to a first retention time range in which a target peak is located and a waveform of the target peak is acquired from a chromatogram created on the basis of the first data. The number of component peaks within a second retention time range corresponding to the first retention time range in a three-dimensional graph created on the basis of the second data is acquired as the number of measured component peaks. A function including a peak width variable is prepared as a model function for measuring the number of component peaks, the model function is changed by adjusting the value of the peak width variable of the model function for measuring the number of component peaks, and the waveform of the model peak represented by the model function for measuring the number of component peaks is compared with the shape of the waveform of the target peak. When the similarity is equal to or greater than a threshold value, it is determined that the change of the model function is ended.

The invention discloses a design method, system and device of a high-power microwave transition waveguide and a medium, and relates to the technical field of transition waveguides, and the method comprises the steps: constructing an initial contour function of the transition waveguide, the initial contour function being the superposition of a specific function and an arbitrary function; any function is converted and expressed as a first function expression form through a preset complete orthogonal basis function; sequentially performing origin center symmetry processing and simplification processing on the first function representation form to obtain a second function representation form; solving a group of coefficients in the second function representation form through a multi-parameter optimization algorithm, and obtaining a third function representation form based on the solving result of the group of coefficients; a target transition function of the transition waveguide is constructed through the third function representation form and the specific function, and the transition waveguide is designed and manufactured based on the target transition function; the problem that a traditional large-caliber transition section is large in length when high transmission efficiency is guaranteed, and compactness of a microwave system is not facilitated is solved.

The application relates to a least square fast image matching method, device, equipment and medium. The method comprises the following steps: simplifying an image transformation model between a real-time image and a reference image according to the characteristics that a translation movement only comprises two translation parameters in a high-precision displacement measurement scene, reducing six affine parameters of a least square matching method into two translation parameters, directly representing a least square target function as a cost function of two-dimensional displacement parameters, directly solving the cost function, and determining positioning information of a displacement table. The target function is represented as a function of transformation parameters, can be directly solved by numerical value, avoids multiple resampling of images, can obtain a global optimal solution, avoids convergence to a local optimal solution, improves the working efficiency of the least square image matching method, improves accuracy and reliability, and has the advantages of high execution efficiency, good real-time performance, simplicity and reliability.

The invention discloses a high-level comprehensive power consumption prediction method and system for composite data stream application, and the method comprises the steps: obtaining a high-level comprehensive function set, constructing a function representation graph for any function in the function set, and determining an anchor node list based on the memory access mode of each function; performing cross-function graph fusion processing on the function representation graph of each function and the anchor node list to obtain a function integration graph; calculating effective delay based on the total delay of single execution, the function starting interval and the number of calling times; counting the total number of bit flips based on each data stream edge in the function integration graph, and obtaining standardized switching activity intensity based on the total number of bit flips and the effective delay; and constructing an initial power consumption prediction model, training the initial power consumption prediction model based on the standardized switching activity intensity and the function integration graph to obtain a power consumption prediction model, and predicting high-level comprehensive power consumption based on the power consumption prediction model.

A method for modeling a wafer profile by a function is provided in which the function is used for calculating a displacement z in a thickness direction of a wafer and is a sum of plural functions. The first function g(r) has a distance r from the center of the wafer as a variable. The second function Ar×h(Nθ) indicates multiplying a sine or cosine function h(Nθ), with a first angle θ with reference to a predetermined position in a circumferential direction of the wafer as a variable and an integer N as a constant, by a coefficient A with the distance r. The third function Br×i(M(θ-φ)) indicates multiplying a sine or cosine function i(M(θ-φ)), with the first angle θ as a variable, a second angle φ with reference to the predetermined position as a constant, and an integer M as a constant, by a coefficient B and the distance r.

The invention discloses a continuous function representation obtaining method and device, a storage medium and an electronic device.The method comprises the steps that first waveform data of a wind power system are obtained, and the first waveform data are discrete waveform data of a draught fan in the wind power system; and inputting the first waveform data into a first implicit representation model obtained by pre-training, so as to obtain continuous function representation corresponding to the first waveform data through the first implicit representation model. The problem that in the related technology, a wind power system waveform data processing technology cannot provide compact and high-precision continuous function representation for waveform data and capable of capturing complex transient characteristics is solved.

A computer-implemented method in a model-based reinforcement learning (RL) system with logic states includes switching an agent between a first mode and a second mode, the first mode being a probabilistic planning mode and the second mode being an information gathering mode. In response to the agent being in the probabilistic planning mode, the agent computes a predictive state representation, given a history of observations and actions taken, and the agent scores action candidates based on planning with the predictive state representation so that actions with resolved plans with confidence to achieve a goal state are scored higher than actions without resolved plans. In response to the agent being in the information gathering mode, the agent scores action candidates based on a Q function of a value of expected information to be gathered from a given pair of state and action.

The application discloses a kind of control flow chart-oriented cross-platform binary function representation method and device, the method includes: obtaining the instruction set of binary code and statistics feature;Obtain the feature vector of each instruction feature;Based on the feature vector of the instruction feature and LSTM model, the basic block embedding is carried out to each basic block in each CFG, and the feature vector of each basic block is obtained;Filter invalid node in CFG, and obtain the feature vector set corresponding to the filtered basic block;Function representation is carried out using graph neural network model;In the graph neural network model, each node obtains the information of its corresponding k-order neighbor;The k-order neighbor of a node is the node in the graph neural network model, which is connected with the node with less than or equal to k connection number.The method has strong data understanding ability and cognitive ability, and the data size of k-order subgraph is controlled by random sampling of neighbor nodes, to save the cost.