51 results about "Numerical range" patented technology

The embodiment of the invention provides a model reinforcement learning method, device and equipment. The scheme comprises the following steps: in a sampling stage, using an inference engine and adopting a to-be-trained target model to generate an output sequence for an input sequence under a first strategy parameter, recording a first probability value of each lexical element in the output sequence, and calculating a dominant value of each lexical element, in a training stage, after a training engine is used for forward propagation to obtain a second probability value of each lexical element generated by a target model under a first strategy parameter, a first ratio of the second probability value to the first probability value of each lexical element can be calculated, and then the lexical elements with the first ratios within a preset numerical range are screened out to participate in calculation of a target function; and optimizing the target function to update the parameters of the target model.

The invention relates to a self-adaptive micro-liquid adding time intelligent optimization method, and belongs to the technical field of industrial process control. The method comprises the following steps: dynamically switching a rotating speed control mode according to a numerical range of an added target quantity, and calculating and outputting a target rotating speed instruction; pipeline pressure is monitored in real time, the pressure state is judged according to a graded pressure threshold system, a graded regulation and control strategy is executed, and a rotating speed fine adjustment instruction is generated; a time adjustment model is constructed based on the residual addition amount and the estimated actual flow, the addition time is dynamically estimated, and replanning of the target time and the rotating speed is triggered when the parameters are abnormal; and after adding is completed every time, operation data are automatically collected, mismatching items of the control parameters and the working conditions are analyzed and recognized through deviation traceability, and the rotating speed control parameters and the pressure regulation and control parameters are updated in a self-adaptive mode. According to the invention, intelligent optimization of the adding process is realized, and the adding precision and the system adaptability are effectively improved.

The invention relates to the technical field of flow field data processing, solves the technical problem that interpolation precision and reliability are reduced due to the fact that data distribution characteristics cannot be accurately captured in a traditional interpolation method, and particularly relates to a flow field data interpolation method based on an optimal transportation theory. Comprising the following steps: dividing flow field data into N sub-regions and carrying out normalization processing; randomly extracting two batches of data as two discrete data distributions, and constructing discrete distributions according to probability quality; establishing a cost function for measuring data direction information and data value information included in the flow field data; establishing an optimal transportation plan for converting a flow field data interpolation problem into a discrete distribution handling problem, and solving an optimal transportation matrix to obtain an interpolation result; and carrying out reverse normalization on the interpolation result to an original numerical range and carrying out smoothing processing. According to the method, through optimization of the optimal transportation theory, the real physical behavior of the flow field can be better reflected, and the accuracy and credibility of the interpolation result in practical application are improved.

According to the method for training the target model based on reinforcement learning, the target model is used for executing tampering detection on an input image, and the method comprises the steps that a training sample is obtained, the training sample comprises a target image and boundary information of a plurality of truth value boxes, and the truth value boxes are borders of a part of images marked in the target image and comprising tampering content; and inputting the target image into the target model, and outputting boundary information of the plurality of prediction frames. And calculating the overall overlap ratio of the plurality of prediction frames and the plurality of true value frames. Determining a first reward score based on a preset piecewise function according to a numerical range of the overall overlap ratio; the function corresponding to each numerical range of the piecewise function is a constant function, and the value of the corresponding constant function is increased along with the improvement of the overall overlap ratio. And based on a reinforcement learning algorithm, according to the first reward score, updating the target model.

The present invention discloses an automated classification method for regression test error simulation waveforms based on a Mamba neural network architecture, belonging to the field of integrated circuits. The method comprises parsing a simulation waveform file obtained from a regression test, converting a key signal waveform into a discrete sequence of integer waveforms; normalizing the numerical range of the waveform sequence to the 0-1 interval, intercepting a fixed length and inputting it into a Mamba neural network model, with the network outputting two parameters: block size and block starting position; dividing the waveform using the two characteristic parameters output by the network, calculating the thickness, degree of fluctuation, and mean characteristics of each waveform block, and calculating the overall characteristics; performing a fast Fourier transform on the waveform using integer multiples of the waveform block size as a window, and calculating frequency domain features; and merging the extracted features into a multidimensional feature vector and inputting it into a support vector machine to output a final classification result. The present invention can effectively improve the accuracy of diagnosing the cause of regression test failure, replace the manual triage process, and improve verification efficiency.

This disclosure provides a data verification method, apparatus, computer program product, non-transitory computer-readable storage medium, and electronic device. The method involves obtaining a target judgment result and a root hash value. The root hash value is generated by the device being verified according to a preset algorithm and leaf hash values ​​generated based on at least one judgment result. Each leaf hash value corresponds to the judgment result of the target data under a certain judgment condition, and the target judgment result corresponds to the target leaf hash value among the at least one leaf hash value. A corresponding verification hash value is calculated according to the preset algorithm based on the target judgment result. The verification hash value is compared with the root hash value to complete the verification of the target data. Embodiments of this disclosure provide an efficient and secure means of verifying numerical ranges without relying on a centralized system.

The invention provides a meteorological data compression method and device, computer equipment and a storage medium, and belongs to the technical field of meteorological data processing, and the method specifically comprises the steps: dynamically adjusting a storage space according to the actual numerical range of original meteorological data; performing precision control on the original meteorological data to obtain loss data; predicting the loss data by adopting a self-adaptive prediction algorithm, and converting the loss data into a residual sequence in combination with prediction data; and performing entropy coding on the residual sequence to obtain compressed data for transmission. Through the processing scheme provided by the invention, the characteristics of high real-time requirement and large data volume of the traditional meteorological data are fully considered, the meteorological data are compressed, the repeatability of the data is reduced, the redundancy of the data is reduced, and the size of the data is reduced to the greatest extent.

The present invention provides a controller parameter adjustment method and device, which relate to the field of automatic control technology. The method comprises: if it is determined that the control loop of the control system is in a steady-state condition, obtaining the overshoot time series of the control system; calculating the relative damping index according to the overshoot time series, determining the controller parameter adjustment strategy according to the numerical range of the relative damping index and the preset corresponding relationship; adjusting the controller parameters according to the controller parameter adjustment strategy. The device executes the above method. The controller parameter adjustment method and device provided by the embodiment of the present invention do not require precise modeling of the control system, but only need to collect the control system operation data without affecting the normal operation of the system to achieve online index evaluation, and are suitable for online performance evaluation of control systems in industrial sites.

This invention provides a parameter prediction model prediction method, device, and medium based on gradient decision trees, belonging to the field of integrated circuit computer-aided design technology. The method includes: acquiring the topological features of the target network; wherein, the topological features include: fan-out number, number of resistor segments, and number of branches; inputting the topological features into a pre-trained prediction model to obtain the predicted value of total parasitic capacitance and the intermediate predicted value of total parasitic resistance; performing an inverse exponential transformation on the intermediate predicted value of total parasitic resistance to obtain the predicted value of total parasitic resistance. This invention, by performing a logarithmic transformation preprocessing on the target variable of total parasitic resistance and performing an inverse exponential transformation after prediction, can effectively compress the numerical range of total parasitic resistance and optimize the data distribution, thereby accurately restoring the true physical resistance value.

When an arithmetic control unit according to the present invention is to extract high-bit data larger than a bit width for output based on a decimal point position, from a cumulative memory of a fixed bit width in which a decimal point position and a division position are set, the cumulative memory storing a cumulative addition result of a multiply-accumulate operation result between input data and a kernel, in an arithmetic circuit that divides data into higher-order data and lower-order data on the basis of the division position and extracts and outputs the higher-order data and the lower-order data, the number of times that each of the higher-order data and the lower-order data exceed an upper limit and fall below a lower limit of a given numerical range is acquired from a counter, and on the basis of the acquired lower limit count for the lower-order side, and the upper limit count for the lower-order side or the upper limit count for the upper-order side, a decimal point position for calculating the next input data is determined and set in the arithmetic circuit, and on the basis of the lower limit count for the higher-order side, a data distribution of the higher-order data and the lower-order data, and a data distribution of the kernel, the division position for calculating the next input data is determined and set in the arithmetic circuit.

The present invention addresses the problem of calculating the root of a power function. An arithmetic device according to the present invention calculates a root Y of a power function (Y = AX, wherein A is a constant, and X is a prescribed numerical range). The arithmetic device comprises: an input unit into which input data indicating an exponent X is input; a first arithmetic circuit that executes a first arithmetic process on the basis of a linear approximation equation of a graph of a power function in a first numerical region of a numerical range of the exponent X; a second arithmetic circuit that executes a second arithmetic process on the basis of a linear approximation equation of a graph of a power function in a second numerical region of the numerical range, the second numerical region differing from the first numerical region; and a selector that outputs output data indicating the root Y calculated by the first arithmetic circuit when if exponent X indicated by the input data inputted into the input unit belongs to the first numerical region, and outputs output data indicating the root Y calculated by the second arithmetic circuit if the exponent X indicated by the input data inputted into the input unit belongs to the second numerical region.

To provide a data processing method and device which can suppress narrowing of a predictable numerical range, while improving prediction accuracy.SOLUTION: A data processing method includes: an outlier detection step of detecting an outlier from a plurality of data included in training data 31; a predicted value calculation step of creating a first regression model 36 using first training data 35 excluding the outlier from the training data 31 as teaching data, and determining a first predicted value which is a predicted value of an objective variable corresponding to a value of an explanatory variable of each of the excluded outliers, using the first regression model 36; and a data substitution step of substituting the excluded outlier with a value based on the first predicted value, to create substituted outlier training data 38 from the training data 31.SELECTED DRAWING: Figure 1

This invention provides a quantization method for the EltwiseAdd operator in neural networks. The method quantizes the EltwiseAdd operator to obtain the EltwiseAddQuant operator, which can directly perform element-wise addition calculations using fixed-point input activation and directly output fixed-point results. The entire calculation process only involves fixed-point addition, avoiding floating-point multiplication and rounding operations. The method includes: S1. Recording the operator activation range; S2. Updating the numerical range of the EltwiseAdd operator activation and updating the EltwiseAdd quantization calculation method; S3. Performing the overall quantization. The advantages of quantizing the EltwiseAdd operator proposed in this invention include reduced computational resource consumption, faster computation speed, lower power consumption, enhanced compatibility, and improved memory efficiency. These advantages make quantization an effective means of model compression and acceleration, especially suitable for resource-constrained environments and applications requiring rapid response.

This disclosure relates to a codeword determination method, apparatus, and storage medium, including: generating a codeword based on a first precoding vector and a second precoding vector, wherein the first precoding vector and the second precoding vector are orthogonal, the codeword is used to precode data, and the difference between a first parameter of the first precoding vector and a second parameter of the second precoding vector is within a first numerical range. In the above embodiments, codewords are determined by the difference between the parameters of two orthogonally related precoding vectors, ensuring that the precoding vectors included in the generated codeword are orthogonal, providing a method for selecting precoding vectors with multiple orthogonal options, thereby ensuring the accuracy of the determined codeword, the accuracy of data precoding based on the codeword, and the reliability of communication.

This invention provides a computer program that can reduce the burden of creating lookup tables for determining the flux linkage of electromagnetic devices having windings. [Solution] A computer program that causes a computer to perform a process to create a lookup table for simulating the operation of an electromagnetic device having multiple windings, based on an analysis model of the electromagnetic device, wherein the program transforms the component axes of multiple currents flowing through each winding so as to broaden the numerical range in which the change in flux linkage for the currents flowing through each winding is gradual and narrow the numerical range in which the change in flux linkage is steep, performs a magnetic field analysis based on the analysis model for the multiple currents whose component axes have been transformed, and causes the computer to perform a process to associate the flux linkage calculated by the magnetic field analysis with the currents.

This application relates to a method, apparatus, and computer program product for determining cost data based on machine learning. The method includes: acquiring part data of a target part; constructing a feature vector corresponding to the target part using the part data; determining the similarity between the target part and historical parts by using the distance between the feature vector of the target part and the feature vectors of pre-stored historical parts; constructing learning data using the feature vectors of historical parts and historical cost data, and training a machine learning model using the learning data; and, when the similarity meets a preset numerical range, using the machine learning model to predict the cost using the feature vector of the target part, thereby obtaining the target cost data of the target part. This method can utilize historical parts that are moderately similar to the target part to construct learning data to train the machine learning model, ensuring the reliability of the model's prediction results and thus improving the accuracy of the target cost data.

The invention discloses a multi-scene diffusion calculation method for a chemical pollution source. The method comprises the following steps: (1) parameter input and verification: inputting pollution source parameters, meteorological parameters, calculation region parameters and pollutant types; performing numerical range, data type and logic consistency verification on the input parameters, filling default values when non-key parameters are missing, and prompting correction when the parameters are abnormal; (2) coordinate conversion and caching; (3) calculating diffusion parameters; (4) single-source concentration calculation; (5) multi-source concentration superposition; and (6) result output: outputting a calculation result in the forms of a data table, a scatter diagram and a JSON file. According to the method, multi-scene unified calculation can be provided, an instantaneous point source, a continuous point source, an instantaneous multi-point source and a moving line source are supported, dynamic weather, terrain and pollutant characteristic correction are introduced to improve precision, calculation logic is optimized, and emergency efficiency is ensured.

The invention relates to the technical field of bearing texture optimization, in particular to a sliding bearing surface texture optimization method and system. According to the technical scheme, the load curve of the load of the rotating shaft in the sliding bearing changing along with the operation time and the rotating speed curve of the rotating speed changing along with the operation time are obtained; determining characteristic working conditions of the sliding bearing in the main operation cycle according to the load curve and the rotating speed curve; carrying out feasible region analysis on each texture parameter of the sliding bearing according to the characteristic working condition, taking a numerical interval of each texture parameter as friction performance prediction input of a BP neural network model, and carrying out friction performance optimization in the numerical interval in combination with a genetic algorithm; and determining the parameter combination corresponding to the optimal friction performance as the target texture parameter of the sliding bearing. According to the technical scheme, the optimization efficiency and accuracy are improved when global optimization is carried out on the local texture of the sliding bearing.

This invention provides a method for testing the vibration response of wind tunnel equipment, belonging to the field of wind tunnel technology. The invention involves installing multiple triaxial accelerometers at the main structural nodes of the wind tunnel equipment and establishing a vibration response game-theoretic optimization model. A vibration signal enhancement model is used to process the original vibration signal, constructing a micro-amplitude vibration feature matrix, a signal superposition weight matrix, and a spurious vibration index matrix for signal analysis. A comprehensive vibration evaluation index is calculated using a vibration response evaluation function, and corresponding optimization strategies are adopted to adjust the input size parameters according to different numerical ranges, achieving adaptive signal processing and temperature compensation. Finally, a test report is output, containing vibration response data, spurious signal identification results, and structural safety assessment conclusions. This invention solves the technical problem in existing technologies where it is impossible to accurately distinguish between the actual structural vibration of wind tunnel equipment and spurious vibration signals caused by temperature changes.

A betting amount proportion display system for roulette calculates and obtains a substantial betting point of game results respectively through a game processing unit, and performs a numerical range classification on numerical values of the points of the game results, so that the points are arranged from small to large numerical values and classified to belong to range groups Pi, i=one of 1, 2, . . . , N, N is greater than or equal to 3. The roulette image displays the game results arranged in a ring, and the roulette image displays bars at positions adjacent to the game results, the bars correspondingly show the points of the adjacent game results, so that betting amounts of the points are displayed proportionally, a length of each bar correspondingly shows the range group Pi, i=1, 2, . . . , N the points are classified to belong to.

The invention discloses a high-precision softmax hardware implementation method, and belongs to the field of neural network hardware acceleration. According to the method, hardware storage and computing resource consumption are reduced by increasing memory access times and two lookup tables; non-spilling the exponent result by subtracting the maximum of the calculated vector elements; by utilizing the format characteristics of floating-point numbers, the index and mantissa parts are processed separately, the numerical range of an input lookup table is reduced, and a high-precision softmax hardware implementation result with an absolute error controlled within 3e-5 can be achieved by using a small number of lookup table storage resources; the processing efficiency is improved by avoiding data conflicts through pipeline parallelism.

The invention relates to a fast phase unwrapping method combining DeBruijn and a half-cycle Gray code. The fast phase unwrapping method comprises the following steps: step 1, generating four phase shift patterns, and a sum; step 2, generating a DeBruijn pattern with a period of T and a DeBruijn pattern with a period of T / 2, and generating a Gray code pattern with a period of T / 2 at the same time; step 3, determining a projection pattern as,..., and projecting the projection pattern to a plane; step 4, calculating a normal fourth-step phase shift; step 5, obtaining a wrapped phase; step 6, searching a corresponding position in the original code element list to respectively obtain a phase order sum; and step 7, defining boundary conditions of phase intervals, introducing a tiny correction value, and selecting and calculating absolute phases of all the intervals through a formula according to a numerical range wrapping the phases. The method has the beneficial effects that the problem of edge jump is solved, and the unwrapping time is shortened.

The application relates to the technical field of artificial intelligence, and provides a weight compression method, a decompression method, a unit and equipment with bit width perception, wherein the compression method comprises the following steps: dividing weight parameters of an artificial intelligence model into a plurality of weight groups, analyzing a numerical range in each weight group, and calculating minimum effective bit widths required for covering the numerical range; generating a corresponding non-zero value bitmap for each weight group; the non-zero value bitmap is used for recording the position distribution of non-zero weight values and zero values in the weight group; and the non-zero weight data in each weight group is arranged according to the corresponding minimum effective bit width, so as to form compressed weight data. The application can adaptively match the variable bit width demand caused by the change of repeated high-bit symbols in different weight parameters; by generating a non-zero value bitmap for each weight group, the all-zero numerical overhead in the weight matrix is accurately identified and directly removed while the bit width is dynamically compressed, so that the storage density is improved, and the external memory overhead and system energy consumption are reduced.

The invention discloses an error code prediction model construction method for LVDS link dynamic characteristics, and belongs to the field of chip test algorithms. Inputting the fault type experiment data with different characteristics into the dynamic model, and extracting and processing the data characteristics by the dynamic model; dividing the time sequence data into an input group and an output group by adopting a sliding window method; the bit error rate is placed in an interval of [0, 1], and the influence of numerical range difference on model training is eliminated; a moving average filter is adopted to carry out de-noising processing on error code noise of experimental data, and the weight of a real-time updating model is designed. In order to verify the actual effect of the dynamic error code prediction model and algorithm, various changing scenes are simulated, and prediction precision and adaptive capacity tests are carried out on the model. Experiments show that the prediction model improves the accuracy of error code prediction, and has obvious advantages compared with a static model.

The invention provides a data processing method and a data processing device, which can improve prediction accuracy and restrain a condition that a predictable numerical range is narrowed. This data processing method is provided with: an outlier detection step for detecting outliers from a plurality of pieces of data included in learning data (31); a predicted value calculation step for generating a first regression model (36) using, as training data, first learning data (35) from which the outliers have been removed from the learning data (31), and obtaining, using the first regression model (36), a first predicted value, which is a predicted value of the target variable corresponding to the value of the explanation variable for each of the removed outliers; and a data replacement step for replacing the removed outliers from the learning data (31) with values based on the first predicted values to generate outlier-replaced learning data (38).

The invention discloses an underground data center energy efficiency grade determination method based on multi-index evaluation. The invention provides a comprehensive evaluation method integrating multiple energy efficiency evaluation indexes for solving the problem of multi-dimensional index difference of an underground data center in energy efficiency evaluation. Firstly, a system comprising five key energy efficiency evaluation indexes including PUE, pPUE, PLF, CLF and WUE is constructed, and a standard range and a good and bad threshold value of each index are determined. In view of the difference between the dimension and the numerical range of each index, the numerical value of each index is converted into a unified score by adopting a mathematical conversion method. The conversion method comprises but is not limited to a linear interpolation method, and a weight is set for each index to ensure the scientificity of an evaluation result. Finally, according to the converted score and weight, the energy efficiency level of the underground data center is obtained through comprehensive calculation, the limitation of traditional single index evaluation is solved, and the overall energy efficiency performance of the underground data center is scientifically reflected.

The invention relates to the technical field of forging data processing, in particular to a forging process parameter automatic writing method and system.The method comprises the steps that in response to target process parameters generated by a client side, local process parameters stored by execution equipment at the current moment are obtained; judging whether the parameter items of the same category between the target process parameters and the local process parameters are the same or not; in response to different parameter items of at least one category between the target process parameters and the local process parameters, judging whether each parameter item of the target process parameters is within a preset numerical range or not; and in response to the condition that each parameter item of the target process parameter is within a preset numerical range, updating a local process parameter stored by the execution equipment at the current moment according to the target process parameter. According to the method, the problem that whether the written process parameters are reasonable or not is misjudged possibly due to huge difference of different parameter dimensions is solved.

The application discloses a kind of based on numerical range multi-converter system stability evaluation method.Method includes: whether the open loop stability condition is satisfied by the equivalent admittance matrix of the admittance matrix transfer function of the converter side of multi-converter system;The numerical range of the equivalent admittance matrix of the admittance matrix transfer function of the network side and the converter side of multi-converter system is established;When meeting the open loop stability condition, then whether the closed loop stability condition is satisfied by the numerical range of multi-converter system, if meeting, then evaluate multi-converter system stability.The application is based on numerical range analysis framework, without global modeling, can realize the fast evaluation of small signal stability of multi-converter system.The application shows stronger criterion applicability and engineering reliability in large-scale, high proportion converter access power system, provides more effective technical means and theoretical basis for the design, planning, operation and control of multi-converter power system.

The present invention relates to a meter state adaptive estimation method supporting data incomplete completion. The method collects continuous measurement data of the meter under a unified time base to form an original measurement data sequence. Based on the original measurement data sequence, a state parameter set for describing the current working state is generated. The original measurement data sequence is divided into multiple data windows, and the expected measurement value in each window is predicted based on the change amplitude and trend direction of the data collected in adjacent time periods. The predicted measurement value is compared with the actual measurement value to identify the position where the difference exceeds a set threshold and form an incomplete position index set. For the incomplete position index set, the completion value is searched within a predetermined reasonable numerical range in combination with the change trend and continuity of adjacent data to generate a completed measurement data sequence. The present invention improves the integrity and accuracy of meter data in the presence of loss or abnormality.

The present disclosure provides a data format conversion method and device, and a matrix processing method and device, and relates to the technical field of computers, in particular to the technical field of artificial intelligence, deep learning, chips and the like. The specific implementation scheme of the data format conversion method is as follows: determining a maximum value of a target matrix according to each element in the target matrix; determining an offset according to a numerical value distribution interval of each element; determining a plurality of continuous numerical value intervals according to a numerical digit width of an index bit of a target data format, the offset and the maximum value of the target matrix; and converting an original data format of the target matrix into the target data format according to a corresponding relationship between each element and the plurality of numerical value intervals and a corresponding relationship between the plurality of numerical value intervals and the target data format. According to the scheme of the present disclosure, the operation amount of data format conversion can be reduced, and the quantization precision of each element in the target matrix can be improved while considering the numerical value range that can be represented by each element.