39 results about "Temporal complexity" patented technology

This invention discloses a fast beam alignment method based on the simplex method. It achieves rapid beam alignment between communication devices operating on highly mobile platforms through two stages: preprocessing and simplex processing. The preprocessing stage introduces a hierarchical search approach, providing a more reasonable initial solution for the subsequent beam alignment algorithm, effectively reducing the likelihood of the simplex algorithm getting stuck in local optima. Compared to previous beam alignment methods, this significantly reduces the overall algorithm's time complexity and implementation complexity, while maintaining the same order of magnitude of angular error, demonstrating superior performance.

The application provides a Petri net supervisory control method and a controller applied to a complex manufacturing system, and the method comprises the following steps: obtaining system information of the complex manufacturing system, and constructing a Petri net model of the complex manufacturing system according to the system information; constructing a reachable graph of the Petri net model according to the Petri net model; taking the current state and resources of each device as input, and predicting transition enabling values of each device in all reachable states through a trained neural network model; and performing supervisory control on the complex manufacturing system according to the transition enabling values of each device in all reachable states. For the complex manufacturing system, the application can avoid the state explosion problem caused by the calculation of the reachable graph of the complex manufacturing system, greatly reduce the time complexity and calculation complexity caused by the calculation of the reachable graph, and directly realize the deadlock control of the complex Petri net and has high behavior permissibility.

The application relates to a fine image classification method, which is characterized by the following steps: constructing a decomposition dictionary learning model, constructing a self-adaptive local order preserving constraint model, constructing a classifier model, constructing a target function, solving the target function and constructing a classification method model. The application can not only adaptively capture the local geometric structure information of atoms and the high-order rank information features of the neighborhood of each decomposition atom in the decomposition dictionary learning process, but also can measure the contribution of multi-view data to the representation of an object through the distance between atoms, enhance the discrimination performance of the decomposition coefficients and the decomposition dictionary. Furthermore, the decomposition coefficients and the class label matrix of the fine image are used to construct a classifier model, the classifier parameters are constrained by a norm, the robustness of the classifier learning model is enhanced, the time complexity of the algorithm is reduced, and the classification performance of the adaptive local order preserving constraint multi-view decomposition dictionary learning algorithm is improved.

The invention provides an aircraft cockpit man-machine interaction task complexity obtaining method and device. The method comprises the steps that man-machine interaction tasks needing complexity evaluation are determined; performing step decomposition on the task according to a flight manual; drawing a task action flow chart according to the specific steps of the task; determining input and output nodes, the number of the input and output nodes, an action type, a time type and a decision type of each node; calculating task step logic complexity; calculating task step scale complexity; calculating task step operation complexity; calculating task step time complexity; task step decision complexity is calculated; calculating a weight factor of each complexity in the total complexity; and calculating the total complexity. According to the method, the task complexity and the expected man-machine interaction behavior in the flight process can be quantitatively described, and the problem of relatively high subjectivity in the previous model workload evaluation work is solved.

This invention relates to the field of game data processing technology, specifically to a method and system for determining the validity of grids in a match-3 game. The method includes the following steps: parsing the acquired game layout configuration data to determine geometric boundary features; traversing each scan line using a scan line algorithm to extract continuous valid intervals; constructing a compact index list containing only valid units based on the valid intervals; and generating interactive objects based on this index list. This invention introduces a discrete topology mapping mechanism to directly locate valid interactive areas, eliminating redundant calculations for invalid empty areas in the traditional full matrix traversal mode. This significantly reduces the time complexity and memory space consumption of logic initialization, solves the technical problem of low efficiency in constructing irregularly shaped layouts, ensures rapid construction and accurate response of game logic in complex terrain scenes, and effectively improves the frame rate stability and user interaction smoothness.

The invention discloses a Chinese character braille conversion method and system oriented to embedded equipment. The method comprises the following steps: firstly, linearly storing a Chinese character pinyin library according to a GB2312 coding sequence, wherein each Chinese character pinyin data record adopts a fixed-length L byte; secondly, analyzing an area code byte Qh and a bit number byte Q1 from the GB2312 code; calculating a physical offset address of the Chinese character pinyin data according to the area code byte Qh and the bit number byte Q1, and adding the physical offset address with the initial physical address to obtain a final physical address; directly reading the Chinese character pinyin data with the fixed length of L from the final physical address; and finally, converting the Chinese character pinyin data into 6-bit braille codes, converting the 6-bit braille codes into 8-bit data through bit splicing and regrouping, and outputting the 8-bit data to an 8-bit data latch to drive braille dot matrix display. According to the method, traversal searching is replaced by direct address calculation, the algorithm time complexity is improved, and the method is suitable for an embedded environment and has the advantages of being high in efficiency and low in resource occupation.

The present application relates to the technical field of surveying and mapping geographic information, and discloses a method and system for detecting error continuous surfaces based on geographic areas, which comprises the following steps: obtaining a plurality of interest surface sets of the same category in a target geographic area; obtaining the minimum circumscribed rectangle of all interest surfaces in the interest surface set; performing a primary judgment of error continuous surfaces on the circumscribed rectangle by comparing the distance from the midpoint of the circumscribed rectangle to the corner point with the distance from the midpoint of the circumscribed rectangle to the origin point, extracting all error continuous surfaces to construct an initial error continuous surface set; performing a secondary judgment of error continuous surfaces according to the positional relationship between the circumscribed rectangles in the initial error continuous surface set, deleting non-error continuous surfaces in the secondary judgment, and completing the update of the error continuous surface set. The present application reduces the time complexity of the algorithm and can meet the demand for rapid detection of large-scale data sets.

Embodiments of the present application provide a target object tracking method and device, equipment and computer readable storage medium, relating to the technical field of artificial intelligence. The method comprises: receiving a target tracking request; in response to the target tracking request, performing feature extraction on an initial template, a current cumulative template and a to-be-tracked region corresponding to a to-be-tracked image respectively, and correspondingly obtaining initial feature embedding, current cumulative feature embedding and to-be-tracked feature embedding; performing cross-correlation processing on the template aggregated feature embedding and the to-be-tracked feature embedding obtained by feature extraction, to obtain a similarity between each sub-region and a target object; determining a tracking result of the target tracking request in at least two sub-regions according to the similarity, and outputting the tracking result. Through the embodiments of the present application, long-range changes in the appearance of the target object can be captured at a relatively low time complexity, and the tracking performance and accuracy of the target object are improved.

This invention discloses a method, apparatus, and medium for solving variable quantum linear problems in subspaces. The method includes: determining the system of linear equations to be solved, constructing a variable quantum circuit and a Krylov subspace, and using the generalized minimum residual method and the variable quantum circuit to calculate an approximate solution of the system of linear equations to be solved in the subspace. This addresses the shortcomings of existing technologies and can reduce the time complexity and computational load of solving linear problems.

This application discloses a retrieval-enhanced ship trajectory prediction method and system based on a frozen TrAISformer and hierarchical adaptive kNN, belonging to the field of intelligent shipping. Addressing the limitations of traditional deep learning models in generalization when facing new routes, new ship types, or seasonal changes, as well as the high cost of updating parameters for pre-trained large models and the difficulty in effectively utilizing historical similar trajectories, this method proposes a solution: acquiring AIS data and generating standardized sequences, inputting the encoded parameters of a frozen TrAISformer model; performing multi-attribute retrieval and aggregation scoring in a four-level hierarchical adaptive kNN library ("ship type-region-season-time"); and finally, outputting predictions by dynamically adjusting weights based on confidence and scenario complexity through an adaptive weighted fusion mechanism. This method avoids catastrophic forgetting, significantly reduces retrieval time complexity, supports online dynamic updates to adapt to new scenarios, and greatly improves prediction accuracy while reducing computational costs.

The invention relates to the technical field of air combat intelligent aid decision making, in particular to an aid decision making recommendation model adaptive method and system based on a Bayesian neural network, and is particularly suitable for man-machine collaborative decision making in a 2V2 air combat game scene. By designing a self-adaptive framework and an incremental learning algorithm, the model can be dynamically updated based on real-time data in the use process of a user, so that the recommendation accuracy and robustness are improved. The method comprises the following steps: constructing an incremental learning algorithm of a Bayesian neural network, analyzing the time complexity and the space complexity of the incremental learning algorithm, evaluating the real-time performance of the incremental learning algorithm in an online reasoning process, and researching the robustness performance under the condition of data uncertainty. The method can be widely applied to strategy selection in an intelligent recommendation system, an auxiliary decision-making platform and a multi-agent game.

The application relates to the technical field of artificial intelligence, and provides an article recommendation method and device, computer equipment and a storage medium, the method comprises the following steps: obtaining historical click data containing user clicks on articles; calculating the similarity between each two clicked articles according to the historical click data; each two clicked articles comprise a first article and a second article; optimizing the articles with each similarity through a preset optimization algorithm, and screening the second articles corresponding to the first article and meeting a preset similarity; obtaining click data corresponding to the first article selected by a user, and recommending the second articles corresponding to the first article to the user according to the click data. The application can reduce the time complexity and space complexity in the recall process, and improve the operation efficiency of article recommendation.

The invention discloses a power station equipment and measurement data association method based on a BiLSTM-CRF algorithm, and relates to the technical field of power station equipment data processing.The method comprises the steps that multi-source data of power station equipment is obtained in an acquisition and input mode, preprocessing including abnormal value and missing value processing, format standardization and space-time alignment is carried out, and the data of the power station equipment is obtained; obtaining a standardized intermediate data set; inputting the standardized intermediate data set into a power enhancement language model, identifying an equipment entity, a measurement type and a semantic relationship, and outputting an entity tag sequence; according to the method, a four-dimensional knowledge graph is constructed on the basis of an association relationship between entity tag sequence identification equipment and measurement points, and time complexity is optimized from a square level to a linear logarithm level through an optimization recursive algorithm, so that automatic association of primary equipment, secondary equipment and the measurement points of a plant station is realized. According to the invention, the problems of long time consumption and easy mismatching and mismatching omission caused by dependence on manual configuration of power station equipment and measurement data association are solved.

The invention relates to the technical field of surveying and mapping geographic information, and discloses an error continuous surface detection method and system based on a geographic region, and the method comprises the steps: obtaining a plurality of interest surface sets of the same category in a target geographic region; obtaining a minimum bounding rectangle of all interest surfaces in the interest surface set; comparing the distance from the midpoint of the external rectangle to the angular point with the distance from the midpoint of the external rectangle to the original point, performing initial judgment on the error continuous surfaces of the external rectangle, and extracting all the error continuous surfaces to construct an initial error continuous surface set; according to the position relation between the external rectangles in the initial error continuous surface set, secondary judgment of the error continuous surfaces is carried out, non-error continuous surfaces in the secondary judgment are deleted, and updating of the error continuous surface set is completed. According to the method, the algorithm time complexity is reduced, and the requirement of large-scale data set rapid detection can be met.

The application provides a delay calibration method and device in an infectious disease space collection population model, and belongs to the technical field of infectious disease transmission modeling and risk assessment; the application adopts a delay time discretization method to convert continuous distribution into a discrete probability vector; an infected person allocation algorithm based on matrix operation and a time axis translation algorithm based on matrix operation are designed; the efficiency is significantly improved, and the time complexity is reduced from to; experimental verification shows that the processing speed of a 100-million-population city is at least improved by 7 times, and the processing speed of a single 100-million-population city can be improved by more than 10 times; the whole algorithm is based on matrix operation under a determined dimension variable, so it is easier to use GPU to accelerate operation, and then the algorithm can be extended to a real-time epidemic monitoring and early warning system in the future.

The invention provides an sEMG hand joint motion continuous estimation method based on sparse attention and distillation, and relates to the technical field of motion continuous estimation methods, comprising the following steps: data acquisition and preprocessing; inputting projection and position codes; projecting the sEMG signal subjected to law normalization from 12 dimensions to 128 dimensions through a linear layer, adding the sEMG signal with a position coding vector, and endowing the sEMG signal with position information; the projected data are processed on the basis of a model architecture fusing a ProbSparse self-attention mechanism, attention distillation and copy stacking, and continuous angle values of hand key joints are output; 150-dimensional data obtained after copy stacking is processed through a regression head, and the 150-dimensional data is mapped into 22 pieces of hand joint angle data. By adopting the method, the time complexity is reduced from O (L) to O (L log L), the space complexity is optimized to O (L log L), the calculation efficiency during long sequence sEMG data processing is improved by more than 20%, and the hardware resource consumption is reduced by 30%.

This application relates to a file query method, apparatus, computer device, storage medium, and program product. The method includes: obtaining a file identifier to be queried; determining a query hash value corresponding to the file to be queried based on the file identifier and at least one preset hash function; and determining a query result for the file to be queried based on preset element values ​​at each bit of a preset bit array and the query hash value. The preset element values ​​include a first element value and a second element value, where the first element value is the element value corresponding to a whitelisted file. This allows for the determination of whether a file to be queried is a whitelisted file using a fixed number of hash calculations and bit array accesses, reducing the time complexity of file queries and improving file query efficiency. Furthermore, using a bit array to store the element values ​​corresponding to whitelisted files also saves storage space.

The application discloses a SQL injection prediction model training method and system, a storage medium, and a SQL injection prediction model training method and system, and a storage medium, wherein a new SQL injection loss function is constructed, a SQL type balance factor, a SQL difficult sample focusing parameter and a smooth hinge loss are coupled, a first-order partial derivative and a second-order partial derivative of the loss function are calculated, the first-order partial derivative and the second-order partial derivative of the loss function are substituted into information gain, a decision tree structure is constructed, a decision tree model is trained, optimal SQL injection features and optimal split nodes are obtained, and an optimal decision tree structure and a corresponding SQL injection prediction model are constructed. The application can reduce the time complexity of model training, avoid overfitting or underfitting of the trained model, and improve the accuracy of machine learning SQL injection detection.

The invention provides a dynamic evolution-based network abnormal user identification method and device, equipment and a storage medium, and the method comprises the steps: constructing a'basic feature-content feature-social feature 'multi-dimensional quantitative portrait through the refined preprocessing of the full-dimensional data of a multi-platform user; based on an improved XGBoost algorithm, fusing the multi-dimensional features to output a user comprehensive credible value; constructing an attribute social knowledge graph, introducing five dynamic evolution factors including a trend factor, a periodic factor, a random factor, an autocorrelation factor and an external variable factor, and realizing dynamic behavior capture in combination with a hierarchical time complexity optimization strategy; the individual credible value is calibrated through confidence propagation and social association fusion, the propagation probability is calculated in combination with total time complexity attenuation, and finally, accurate identification of individual and group anomalies is realized through an anomaly score.

This invention belongs to the field of protocol reverse engineering technology and proposes a method and apparatus for reverse analysis of unknown industrial control protocols. It involves forming a set of message fragments based on network packets; the packets and message fragment sets constitute a dataset; a heterogeneous message graph is constructed based on the dataset; the heterogeneous message graph is input into a neural network model for message graph feature extraction to perform message clustering and train the neural network model; the dataset to be clustered is input into the trained neural network model for clustering; within the same message cluster, the Needleman-Wunsch algorithm is used to infer the syntax format and mark field boundaries. This invention reduces the time complexity of message clustering; it jointly optimizes feature extraction and clustering; and it provides the model with finer-grained and more reasonable analysis units.

The application relates to an abnormality identification and examination and evaluation method and system based on multi-source data fusion, which comprises the following steps: acquiring a to-be-checked object set; taking a globally unique identifier as an index to acquire scattered feature data, and cutting the to-be-checked object set into multiple parameterized sub-sets; taking the globally unique identifier as an aggregation connection key to perform in-memory level stream pipeline recombination on personnel library feature data and activity library feature data to obtain a high-dimensional state feature stream; calculating a real examination basis based on the high-dimensional state feature stream; calculating an integral quantization score based on the real examination basis and a legal exemption time set; further determining an abnormality preliminary circled number of people; calculating a corresponding abnormality maximum quota number based on the to-be-checked object set; comparing the abnormality preliminary circled number of people and the abnormality maximum quota number to determine an abnormality outlier set for object examination and evaluation. Compared with the prior art, the application has the advantages of low algorithm time complexity.

The application provides a network abnormal user identification method and device based on dynamic evolution, equipment and storage medium, which constructs a multi-dimensional quantitative portrait of "basic features-content features-social features" through fine preprocessing of multi-platform user full-dimensional data; outputs a user comprehensive credibility value based on the improved XGBoost algorithm and the fusion of multi-dimensional features; constructs an attributed social knowledge graph, introduces five types of dynamic evolution factors, namely trend factor, cycle factor, random factor, autocorrelation factor and external variable factor, and realizes dynamic behavior capture combined with hierarchical time complexity optimization strategy; calibrates individual credibility value by fusing social correlation through confidence propagation, calculates propagation probability combined with total time complexity decay, and finally realizes accurate identification of individual and group anomalies through abnormal score.

The invention discloses an SQL (Structured Query Language) injection prediction model training method and system and a storage medium, and the method comprises the steps: constructing a new SQL injection loss function, coupling an SQL type balance factor, an SQL difficult sample focusing parameter and smooth hinge loss, and calculating to obtain a first-order partial derivative and a second-order partial derivative of the loss function; substituting the calculated first-order partial derivative and second-order partial derivative of the loss function into information gain; and constructing a decision tree structure, training a decision tree model to obtain an optimal SQL injection feature and an optimal split node, and constructing an optimal decision tree structure and a corresponding SQL injection prediction model. The time complexity of model training can be reduced, over-fitting or under-fitting of the trained model is avoided, and meanwhile, the accuracy of machine learning SQL injection detection is improved.

The invention discloses a hexagonal pixel image conversion method capable of effectively keeping edges, which constructs a space geometry and gray feature double-weight influence model, designs a novel interpolation algorithm and is applied to hexagonal image pyramid generation. A contrast experiment result shows that compared with a traditional method, the method has the advantages that the capability of keeping image edge information is higher; in the synthesized image, the optimal average quality factors of 92.76% and 57.26% are respectively obtained under the high signal-to-noise ratio of 100 and the low signal-to-noise ratio of 20; in terms of calculation efficiency, the algorithm time complexity is close to that of a traditional linear interpolation algorithm and is generally superior to that of other edge preserving algorithms. The research result provides an effective technical support for the practical application of hexagonal image processing, and particularly shows a good application prospect in the aspects of high-precision geographic space analysis, multi-scale feature extraction and the like.

The application provides a controllable text generation method and device, comprising: obtaining a to-be-processed task; inputting the to-be-processed task into a pre-constructed controllable text generation model to obtain a task reply; the controllable text generation model is obtained by training a pre-stored language model using a training sample data set; the training sample data set comprises a data set composed of comparison samples; the comparison samples at least include negative example reply samples, positive example reply samples and to-be-processed task samples; the positive example reply samples are reply samples without undesirable behaviors, and the negative example reply samples are reply samples with undesirable behaviors. The application does not need to modify the model architecture or introduce additional auxiliary models, simultaneously performs language modeling and comparison learning training using the training sample data set to obtain the controllable text generation model, the trained model is ready for use, and the application has higher universality, lower time complexity and more stable reply fluency.

The application provides a multi-modal file real-time recommendation method and system based on cooperative intention perception, constructs a sparse-dense double-path model to unify semantic representation, breaks the cross-modal feature space isolation through a streaming topic tracking mechanism, realizes comprehensive coverage of explicit, implicit and reference intentions through a three-layer funnel model, processes high-frequency requests at a very low cost through a rule layer, realizes low-overhead implicit intention reasoning through a semantic layer by reusing existing semantic vectors, solves the long-tail reference resolution problem through a state machine in a context layer, realizes safe skipping of a large number of irrelevant data blocks through a geometric block index and a dynamic pruning strategy, compresses the range of accurate calculation to a small candidate space, significantly reduces the query time complexity, meets the real-time requirements of online cooperation, and realizes highly personalized accurate recommendation through a dynamic intention perception rearrangement network through a Transformer self-attention mechanism.

The present application relates to a kind of general sketch-based high-speed network per-flow cardinality and distribution function estimation method.The method utilizes hierarchical sketch structure design, while storing per-flow cardinality index using sketch method, for the super-point flow of large cardinality, exclusive stateful flow table space is allocated, the corresponding flow identifier and cardinality value are stored, and the per-flow cardinality estimation accuracy is improved.In the hierarchical structure, the sub-sketch of each layer combines an exact data structure and a sketch structure, a set of in-and-out mechanism between the two based on minimum heap is introduced, and the noise size at the time of moving in is recorded through snapshot, and the hash collision is relieved.Meanwhile, the method can calculate the arbitrary moment of per-flow cardinality, and reconstruct the distribution function of per-flow cardinality.The present application realizes the accurate estimation of per-flow cardinality value, flow cardinality high-order moment and distribution function and other related statistical indicators in high-speed network traffic without significantly improving the time complexity of algorithm.

This invention discloses a target recognition method, system, device, and medium based on binocular vision, relating to the field of target recognition technology. The method includes: performing grayscale processing on preprocessed left and right views; extracting multiple corresponding feature points from the left and right grayscale images obtained through grayscale processing using the FAST corner detection method; obtaining descriptors for each feature point using the BEBLID algorithm, and calculating the Hamming distance between each feature point in the left grayscale image and each feature point in the right grayscale image using the descriptors; generating a feature point group using at least two feature points whose Hamming distance does not exceed a Hamming distance threshold; removing feature points from the feature point group that do not satisfy a cosine similarity threshold to obtain feature point sets for the left and right views; reducing mismatches of feature points to reduce computational time complexity, thereby improving recognition speed and real-time performance of object recognition.