13 results about "Scalar Value" patented technology

The invention discloses a GRU-based vehicle-mounted CAN intrusion detection method and system, a storage medium and a computer system. According to the method, an automatic encoder (AE) is introduced to deepen the understanding of a model on input sequence characteristics, a sliding window is used for selecting batch CAN data to be preprocessed to obtain 13-dimensional time sequence data, and a scalar value within the range of [0, 1] is obtained through processing of the encoder, a GRU, a decoder, a full connection layer and a sigmoid activation function and used for classification of abnormal data. The Conv1D is used as a hidden layer, and compared with two-dimensional convolution, the one-dimensional convolution parameter quantity is smaller, and the calculation is simpler and more convenient. An attack message and a normal message can be completely distinguished, the precision and the accuracy rate reach 100%, and the precision in Fuzz detection is 0.9983; compared with the prior art, the method has high accuracy and reliability in the aspect of intrusion behavior detection, can effectively identify most intrusion events, and can keep a relatively low overall error rate, so that good balance between safety and availability is realized.

The application provides an image processing method and a processing method and device of an image processing model, relates to the field of artificial intelligence, and particularly relates to the field of computer vision. The image processing method comprises the following steps: generating a scalar map based on the difference between the distribution of reconstruction errors and the distribution of noise between a first image and a second image, the scalar map comprising one or more regions, and the scalar value corresponding to each region in the one or more regions being used for indicating the quality of one or more image blocks in the second image. The application can intuitively reflect the quality of the image, and is beneficial to guaranteeing the effect of image processing.

A computer-implemented system for graph-based analysis of institutional succession coverage, the system comprising the following: a storage unit that stores structured data records, including • a role matrix consisting of role identifier values, organizational unit identifiers, hierarchy level parameters, strategic impact parameters, operational impact parameters, replacement difficulty parameters, disruption risk parameters, and role dependency identifiers; • a person attribute matrix comprising person identifiers linked to role identifiers and generative attribute parameters, including parameters for contribution to the legacy system, parameters for mentoring contribution, parameters for innovation contribution, parameters for contribution to the community, and parameters for institutional alignment; • an interaction event matrix comprising relational interaction data records, including originating person identifier, targeting person identifier, interaction type identifier, interaction frequency parameters, interaction intensity parameters, interaction start timestamp, interaction end timestamp, and knowledge criticality parameters; and • a risk parameter matrix that includes person role risk parameters including exit probability parameters, proximity to retirement parameters, contractual risk parameters and external market risk parameters; a matrix preprocessing unit connected to the storage unit and configured to convert heterogeneous parameter scales stored in the person attribute matrix into normalized scalar values ​​within a predefined reference interval; a vector generation unit coupled to the matrix preprocessing unit and configured to create a multidimensional generative vector for each person identifier by arranging the normalized scalar values ​​of the generative attribute parameters into an ordered numeric vector; a graph construction unit coupled to the interaction event matrix and configured to generate an institutional interaction graph represented by an adjacency structure in which the nodes correspond to person identifiers and the edges to relational interaction datasets weighted by the interaction frequency parameter, the interaction intensity parameter, and the knowledge criticality parameter; a unit coupled to and configured with the role matrix for calculating role criticality, which calculates a role criticality index for each role identifier using a weighted aggregation of the strategic impact parameter, the operational impact parameter, the difficulty of replacement parameter, and the disruption risk parameter stored in the role matrix; a network metric calculation unit coupled with the graph construction unit and the vector generation unit, configured to calculate node relevance indicators, including graph degree values, mediation values, and proximity values ​​derived from the adjacency structure, and subsequently combines the node relevance indicators with the multidimensional generative vectors to generate generative centrality values ​​associated with the nodes; and A successor selection control unit coupled with the role criticality calculation unit, the network metric calculation unit, and the risk parameter matrix is ​​configured to identify successor candidate sets for roles with role criticality index values ​​exceeding a threshold by applying coverage constraints defined in a coverage parameter matrix, including minimum successor number parameters and readiness level constraints.

Approaches presented herein provide for the use of implicit functions to simulate differentiable object elasticity. An implicit continuous function, such as a signed distance function (SDF), can be used to approximate the surface of an object by providing scalar values from a set of vertices of a regular grid in which the object representation is to be generated. Interpolation can be applied to determine an approximate surface location and shape within each boundary cell. A trained neural network, such as a multilayer perceptron (MLP), can be used to determine appropriate quadrature points that fall within the volume of the object. A finite element analysis can integrate over these quadrature points, using both continuous and discrete settings, as a basis for performing efficient differentiable elasticity simulations including the deformable object.

This application provides an intelligent answering method, apparatus, device, and medium based on interactive follow-up questioning, applicable to the financial or healthcare fields. The method includes: receiving a question to be answered and generating multiple candidate answers; calculating the overall uncertainty scalar value of all candidate answers based on a preset uncertainty identification model, and determining whether the uncertainty scalar value is lower than a preset threshold; if not lower, determining the uncertainty type based on the preset uncertainty identification model; generating a target follow-up question based on the question to be answered, multiple candidate answers, and uncertainty type based on a preset follow-up question generation model, and sending it to a client; receiving the target answer returned by the client and updating the question to be answered until the overall uncertainty scalar value of the regenerated candidate answers is lower than the preset threshold, then selecting the optimal answer from among the multiple regenerated candidate answers according to a preset answer selection rule and sending it to the client. This addresses the problem of model overconfidence.

An information processing device according to one embodiment comprises: an acquisition unit that acquires observation data which is composed of the number of data points and input value data at each of the data points; an inverse M-matrix processing unit that, on the basis of the observation data, selects, for a matrix which is calculated from the data points and the input value data at any data point, a non-negative scalar value and a positive definite kernel function in which all off-diagonal elements of an inverse matrix of the matrix becomes negative or zero; an estimation model learning unit that learns a parameter of an estimation model by solving a minimization problem using the observation data, the positive definite kernel function, and the non-negative scalar value; and an output control unit that outputs the positive definite kernel function, the non-negative scalar value, the learned parameter, and the input value data.

The application discloses a prefix sum and calculation device, method, chip, board card and electronic equipment, and relates to the technical field of prefix sum and calculation, in particular to a prefix sum and calculation device, method, chip, board card and electronic equipment.The device comprises a data segmentation unit, which is used for grouping vector data to be processed according to a preset segmentation granularity, splicing each vector obtained by grouping in sequence to form a matrix to be processed; a matrix operation unit, which is used for batch calculating prefix sums of each row data in the matrix to be processed to obtain an output matrix; a vector operation unit, which is used for performing addition operation on each row data in the output matrix and a corresponding scalar value to obtain a result vector corresponding to each row data, wherein the corresponding scalar value is equal to the accumulated value of the last data of each row before the row; and a data splicing unit, which is used for splicing each result vector in sequence to obtain a prefix sum vector corresponding to the vector data to be processed.The application can fully utilize the matrix operation resources and vector operation resources inside a chip, improve the overall processing efficiency of prefix sum calculation while ensuring the calculation accuracy.

The invention discloses a power cable partial discharge positioning method and system based on Transform and end-to-end regression. The method comprises the following steps: acquiring and preprocessing a cable partial discharge original time domain signal; calculating multi-order cumulative sum features of the preprocessed signal, and splicing the preprocessed signal with each order of cumulative sum feature in a channel dimension to construct a multi-channel input tensor; inputting the multi-channel input tensor into a pre-trained multi-task deep learning model, and synchronously outputting partial discharge pulse and reflection pulse position probability distribution and a continuous scalar value representing the physical distance of a discharge point through one-time forward propagation based on a Transform architecture; and outputting the continuous scalar value as a positioning distance, and verifying a positioning result based on the probability distribution. According to the invention, end-to-end accurate mapping from the original waveform to the physical position is realized, the positioning precision is high, the anti-noise capability is strong, and the automation degree is high.

According to one aspect of the present invention, provided is an information processing system, comprising at least one processor, wherein the processor is configured to execute a program for executing the following steps: an acquisition step in which a model function V1, an allowable error δ required with respect to a piecewise polynomial expression V2 which approximately represents the model function V1, and an order set {p} which is a set of candidate orders p for the piecewise polynomial expression V2 are acquired, where the model function V1 outputs a scalar value with respect to input and is defined by at least one continuous interval; a division step in which, on the basis of one specified order p1 selected from the order set {p} and a division parameter m representing an upper limit for a number of continuous interval divisions, a continuous interval for which an upper bound function representing the upper bound of the maximum value for the difference between the model function V1 and the piecewise polynomial V2 in said continuous interval is the allowable error δ or less is divided into 2m or fewer divided intervals, where the upper bound function does not depend on a variable for determining the piecewise polynomial V2, and the allowable error δ is stipulated on the basis of the division parameter m and a norm of a derivative V1(p1+1) of the (p1+1)th order differentiation of the model function V1 corresponding to the specified order p1; an approximate expression calculation step in which the piecewise polynomial expression V2 is calculated by using prescribed spline interpolation to optimize a p1th-order polynomial expression corresponding to the specified order p1 to the model function V1 per each divided interval; and in a circuit generation step, on the basis of the calculated piecewise polynomial expression V2, a quantum circuit representing a diagonalized unitary operator corresponding to the piecewise polynomial expression V2 is generated, where the quantum circuit is configured to act on n computation quantum bits encoded so as to represent position in the continuous interval, and the operation amount of a quantum gate operation included in the quantum circuit is determined on the basis of a coefficient of the piecewise polynomial expression V2.

The invention discloses a smooth curved surface generation method based on sparse control points, and relates to the technical field of curved surface generation, and the method comprises the steps: S1, receiving a two-dimensional control point matrix arr, the elements of the matrix being scalar values, and the scale of the matrix being m rows and n columns; through a row-column step-by-step interpolation structure, a global system does not need to be solved or complex boundary and derivative conditions do not need to be set, the algorithm structure is simplified, the calculation cost is reduced, a smooth transition function is adopted, hard folding and angle sudden rotation of linear interpolation at nodes are avoided, soft acceleration-deceleration change of an interpolation transition section is achieved, adjustable interpolation density parameters are set, and the method is suitable for large-scale popularization and application. A user or a system can flexibly set according to performance and visual requirements, balance performance overhead and a smoothing effect, guarantee that the value of an original control point in an output grid is not changed through boundary fidelity processing, support block processing, cache calculation and a progressive refinement mode, and is suitable for a large-scale matrix and a real-time visual scene; and the requirements of multiple fields on smooth curved surface generation are met.

Techniques for encoding and decoding audio signals are described. A decoder (10) configured to generate an audio signal (16) from an encoded signal (3) representing the audio signal may comprise: an encoded signal reader (560) configured to read the encoded signal (3), thereby providing a plurality of indices (556); a scalar dequantization module (500) comprising: a plurality of quantization index converters (355), each quantization index converter (555) configured to convert an index (556) of the plurality of indexes to a corresponding potential scalar value (551) such that the plurality of potential scalar values (551) form a first potential audio signal representation (550) of the audio signal; and a first learnable section (540) for providing a second potential representation (530) from the first potential audio signal representation (550); a second learnable section (520) comprising at least one learnable layer and configured to generate an audio signal (16) from a second potential audio signal representation (530).

An example machine learning system for processing data associated with transactions is described. The machine learning system has a first processing stage that includes a recurrent neural network architecture. The recurrent neural network architecture has a forget gate to modify state data of a previous iteration based on data representing a time difference between a proposed transaction and a prior transaction. The machine learning system also has a second processing stage that has an attention neural network architecture communicatively coupled with the first processing stage. The machine learning system is configured to map output data from the second processing stage to a scalar value representing a likelihood that the proposed transaction presents an anomaly in a series of actions. The scalar value is used to determine whether to approve or reject the proposed transaction.