216 results about "Hierarchical clustering" patented technology

The invention relates to the technical field of artificial intelligence, in particular to a self-organizing knowledge base construction method, system and device based on a multi-agent system and a storage medium, and the method comprises the steps: S1, monitoring a dialogue information flow, calculating the self-reply confidence coefficient of an agent, and if the self-reply confidence coefficient exceeds a preset dynamic threshold value, judging that a potential knowledge value exists and triggering an extraction request; s2, in response to the request, performing semantic coding on a dialogue fragment, extracting knowledge elements, and generating a structured knowledge unit by utilizing graph neural network modeling and relation calibration; s3, through a pre-training semantic coding model, mapping the multi-dimensional semantic coding model into a multi-dimensional semantic vector; s4, carrying out hierarchical clustering comparison and combination, and if a threshold value is exceeded, establishing a new classification and adjusting a knowledge base index; and S5, knowledge fingerprints are generated and compared, and fusion updating is carried out based on the reputation scoring model when semantics conflict or redundancy occurs. According to the method, unstructured dialogue high-precision knowledge extraction is realized, so that the knowledge base structure is dynamically self-organized according to semantics, multi-source knowledge is coordinated, and the knowledge management automation level and quality are improved.

The invention discloses a layered densification Gaussian sputtering method based on visibility, and relates to the technical field of artificial intelligence and computer vision. According to the scheme, an initial three-dimensional Gaussian primitive set is generated based on sparse multi-view observation data, and initial scene representation is established by extracting spatial distribution parameters, morphological parameters and radiation parameters; performing fusion analysis on the Gaussian primitives based on the multi-dimensional observability parameter set to obtain comprehensive observability index data; executing hierarchical clustering according to the index data, and constructing a multi-layer Gaussian structure of a significant layer, a transition layer and a background layer; performing density enhancement, geometric continuity constraint and parameter update processing on different levels of Gaussian structures, and generating a rendered image of a target view angle under a volume light traveling and transparency hybrid mechanism; according to the method, continuous reconstruction of a scene structure and accurate expression of radiation characteristics can be realized under the sparse view condition, and the geometric fidelity and rendering consistency of new view angle synthesis are improved.

The invention relates to the technical field of document intelligent analysis, and discloses a document content self-adaptive analysis method and system based on a large model. The method comprises the following steps: acquiring original data flow of a to-be-analyzed document, wherein the original data flow comprises a text coding sequence, a layout structure mark and a multimedia embedding feature; the data stream is input into a pre-trained multi-modal large model, and a document semantic graph structure, a concept-containing node set, a relation edge weight matrix and a cross-modal alignment index are generated through context sensing analysis; performing dynamic hierarchical clustering on the semantic graph structure to obtain a hierarchical topic tree containing core topic branches, secondary topic branches and leaf node association strength; extracting a document logic framework containing chapter division suggestions, key information positioning coordinates and a cross reference mapping table according to the topic tree; a result is generated based on an adaptive analysis strategy optimization framework, and the strategy adjusts clustering granularity and relation mining depth according to document type features.

The invention relates to a three-dimensional point cloud registration method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring source point cloud data and target point cloud data; identifying the source point cloud data and the target point cloud data, and determining a target reference point; performing hierarchical clustering on the target reference points to obtain a reference point cluster; determining a global transformation matrix according to the first reference point cluster of the source point cloud data and the second reference point cluster of the target point cloud data; wherein the global transformation matrix represents a pose transformation relationship between the source point cloud data and the target point cloud data; and iterating the source point cloud data and the target point cloud data according to the global transformation matrix in combination with ICP fine registration until point cloud registration is completed. Thus, by extracting the target reference points and performing hierarchical clustering to obtain the reference point clusters, key local units with global structure significance can be extracted from the disordered point cloud, and compared with a traditional point-by-point matching point cloud registration method, redundant calculation and local structure misjudgment can be reduced.

A three-stage pipeline is used to create a data structure for efficiently producing grounded recommendations that guarantee that the recommended items are part of a set, D. In the first stage, each item in D is converted into a vector representation. In the second stage, a hierarchical clustering method is used to build a tree based on the vector representations. Each item is a leaf node of the tree. Each non-leaf node represents a group of items or a group of groups of items, and so on. In the third stage, an LLM is used to generate text that encapsulates the information of the group (or groups) of items represented by each node. The generated tree is recursively traversed to generate recommendations.

The invention belongs to the technical field of electronic search, and relates to a radio reconnaissance method and system based on multichannel parallel processing and intelligent clustering, and the method comprises the steps: carrying out the multi-phase filtering digital channelization processing of a broadband radio signal, so as to obtain adjacent channel sequences with overlapped frequency bands; performing square law detection and constant false alarm detection on each channel sequence to generate an initial pulse description word; establishing a candidate corresponding relationship in adjacent channels based on arrival time similarity and noise consistency, and determining a cross-channel merging relationship according to frequency range continuity and frequency-time relationship slope consistency to generate a complete pulse description word; performing hierarchical clustering sorting based on the arrival angle and the pulse width; and performing multi-stage differential analysis on a clustering result to extract a repetition frequency sequence and determine a repetition frequency type. According to the technical scheme, the stability and consistency of cross-channel pulse merging and sorting results can be kept under the broadband high-density condition, and the radio search processing precision under the complex electromagnetic environment is improved.

The invention belongs to the field of computer system structures and artificial intelligence accelerators, and discloses a software and hardware collaborative heterogeneous storage computing accelerator oriented to full-chip DLRM reasoning. A control domain composed of a static random access memory and an embedded dynamic random access memory, an on-chip storage domain, an in-memory computing domain and an on-chip computing domain are integrated in the same chip, an embedded table is compressed by introducing a hierarchical clustering quantization frame, and a dynamic storage management strategy of access frequency perception is combined. The embedded access path is shortened, and the bandwidth requirement is reduced. According to the method, the reasoning process of the whole set of DLRM can be independently completed, and the memory access overhead in a traditional terminal SoC is remarkably reduced by reducing cross-domain data migration between the calculation unit and the storage unit, so that the model reasoning performance and energy efficiency are improved.

The invention provides a content auditing method and device, electronic equipment and a storage medium, and relates to the technical field of content auditing, multi-level feature extraction is performed on an original text to generate text vectors, and a self-adaptive hierarchical clustering algorithm is adopted to cluster the text vectors, so that the content auditing efficiency is improved. And meanwhile, key text features are extracted based on a clustering result to generate hierarchical violation labels, and the violation classification model is updated by utilizing new violation labels through an incremental learning mechanism. The problems that in the prior art, text variants are difficult to recognize due to insufficient keyword filtering semantic understanding, traditional hierarchical clustering calculation is high in complexity and cannot adapt to large-scale data, a machine learning classification model lacks a dynamic updating mechanism, so that new illegal content is missed, and the model is difficult to continuously adapt to new auditing requirements can be solved. The technical effects of improving the semantic recognition accuracy of the violation content, improving the real-time performance of large-scale text auditing, automatically discovering new violation categories, reducing leak detection and enhancing the expansibility and dynamic adaptability of a content auditing system are achieved.

The invention discloses a semiconductor device aging test data analysis method and device and a medium, and relates to the technical field of data analysis, and the method comprises the steps: expanding a time sequence-parameter joint aging behavior matrix into a space-time degradation mapping matrix, generating a Hilbert space holding scanning path matched with the order of the space-time degradation mapping matrix, and carrying out the scanning of the Hilbert space holding scanning path; sequentially accessing each numerical value of the space-time degradation mapping matrix according to a Hilbert space maintaining scanning path, sequentially recording the numerical values corresponding to an access sequence to obtain an aging behavior one-dimensional sequence, and sequentially arranging the aging behavior one-dimensional sequences in batches according to device numbers to form a cross-device behavior sequence matrix; according to the method, the cross-device behavior sequence matrix is constructed to automatically identify the anomaly, and the hierarchical clustering and the independent extraction of the anomaly group are assisted to realize consistency grading, so that the sensitivity and specificity of anomaly identification and the objectivity and accuracy of batch consistency evaluation are remarkably improved.

The invention discloses a micro-grid cluster dimension reduction method and system based on scene self-adaption and topology maintenance. Aiming at different types of energy, constructing a steady-state characteristic index system and decoupling the steady-state characteristic index system into a plurality of groups of characteristics; the method comprises the following steps: clustering system time series data, constructing an operation state manifold diffusion matrix, clustering by using a fuzzy clustering method based on diffusion distance, and outputting an optimal scene division result; constructing a Wasserstein distance matrix by using the steady-state characteristic index of each type of energy, defining a scene objective function and solving an optimal index weight; and calculating a scene membership degree and a fusion index vector weight according to the real-time operation state, weighting the steady-state characteristic index system, and outputting a weighted characteristic matrix. And performing multi-scale topology analysis on the weighted feature matrix, constructing a topology stability constraint, and aggregating the power grid equipment by adopting a hierarchical clustering method based on the topology stability constraint. According to the technical scheme, the internal evolution rule of the operation state can be accurately captured, so that the aggregation result better meets the actual operation requirement.

The invention discloses a bit-level CAN (Controller Area Network) data load reversing method and device based on hierarchical clustering. The method comprises the following steps: acquiring original CAN message data in a vehicle running process through a CAN data acquisition tool, and processing the CAN message data by utilizing a frequency matrix, a probability matrix, a bit flipping rate and a hierarchical clustering method to obtain CAN message signal characteristics; performing reverse analysis on the CAN message signal by analyzing the characteristics of the CAN message signal and combining the signal type of the automobile to obtain a reverse analysis result; and controlling the functions of the automobile based on the reverse analysis result. According to the method and the device, the problem of dependence on special equipment and a database (DBC file) private by a manufacturer in the prior art is solved, and the technical problem of more accurate and finer-grained signal automatic reverse analysis is realized.

The application provides a data hierarchical clustering method and device, which comprises the following steps: determining a plurality of feature vector sets, wherein the i-th feature vector set in the plurality of feature vector sets corresponds to the i-th confidence and the i-th activity; wherein the activity represents the number of occurrences of a target within a preset time length, and the confidence represents the confidence of the data corresponding to the target; the i-th feature vector set comprises at least one feature vector, the at least one feature vector corresponds to at least one target in a one-to-one manner, the confidence of each feature vector in the at least one feature vector is the i-th confidence, and the activity of each feature vector in the at least one feature vector is the i-th activity; wherein i is a positive integer. The plurality of feature vector sets are sorted according to the confidence and the activity corresponding to the plurality of feature vector sets. The above method is used to filter the data in layers, so that the data can be quickly clustered.

The invention relates to a voice tag generation method and device, electronic equipment and a storage medium, and belongs to the technical field of computers, and the method comprises the steps: carrying out the feature matrix decomposition of a plurality of to-be-processed first voice feature matrixes, and obtaining a decomposed first matrix and a decomposed second matrix; performing hierarchical clustering on column vectors in the first matrix to obtain a tag tree of the first matrix; performing projection processing on the second voice according to the second matrix to obtain a voice feature representation vector of which the dimension is the same as the column number of the first matrix; querying a tag tree, and determining a feature tag of the second voice; wherein the feature tag of the second voice is determined according to the clustering tag in the tag tree mapped by the voice feature representation vector. According to the embodiment of the invention, the construction and maintenance cost of a voice feature tag generation system can be remarkably reduced, and the automation degree and robustness of tag generation are improved.

The invention provides a child teaching interaction system and platform for an intelligent robot, and belongs to the technical field of intelligent interaction.The child teaching interaction system comprises the steps that an interaction collection module collects an element trigger array of child output behaviors and an effective state sequence of the interaction collection module, and an interaction control matrix is constructed; the vector extraction module associates the response component with the matrix row vector; the sensitivity judgment module comprehensively judges the sensitivity of the core sensing element through hierarchical clustering and multi-index and performs targeted adjustment; and the interaction control module acquires new data based on the adjusted elements, and compares the new data with a standard matrix to generate multi-modal teaching feedback. According to the invention, component abnormity accurate identification and rectification are realized, the interaction data acquisition accuracy and the teaching feedback personalized matching degree are improved, and the teaching interaction experience of children is optimized.

The invention discloses an aero-engine measurement parameter selection method based on a clustering method, and belongs to the technical field of engine state monitoring and fault diagnosis, and the method comprises the steps: building an aero-engine model, and generating a gas path fault sample; obtaining a sample and performing normalization processing to enable the sample to only reflect the percentage of the measurement parameter change in the fault mode; hierarchical clustering is executed under different clustering numbers, and a Davies-Bouldin index is calculated; the clustering number corresponding to the minimum Davies-Bouldin and the clustering result of the minimum Davies-Bouldin are selected as the optimal division; and aligning the optimal division result with a real label, and calculating a consistency index. According to the method, on the premise that key fault information is reserved, the number of required measurement parameters is remarkably reduced, the optimized parameter combination still keeps a high fault recognition rate under the condition of limited sensors, and the efficiency and practicability of state monitoring and fault diagnosis are improved.

The application relates to the technical field of security protection, can be applied to a business system platform such as financial technology and medical health, and discloses a Byzantine attack defense method, device and equipment based on federated learning and a medium. The method comprises the following steps: acquiring local models and sample data sets of a plurality of target clients, performing forward propagation on the sample data sets by using the local models, constructing a plurality of probability matrices, determining the Euclidean distance between any two probability matrices, constructing a distance matrix, performing hierarchical clustering on the target clients, obtaining a plurality of client clusters, acquiring a probability standard matrix of a standard model, determining the similarity between the probability matrices and the probability standard matrix one by one, marking the types of the to-be-marked matrices according to the similarity, updating the information of the client clusters, obtaining updated clusters, aggregating all the updated clusters, and obtaining a final defense model. The application can effectively improve the defense efficiency and defense scene coverage rate for the Byzantine attack.

The purpose of this invention is to address the issue of accuracy in processing spinning time-series data streams, which are characterized by high noise levels and distinct time-domain features generated during the operation of spinning workshops, by employing a fuzzy hierarchical clustering method that integrates time-domain characteristics. This method reduces the impact of noise during the classification process and considers both time-domain features and noise effects. The technical solution of this invention is to provide a fuzzy hierarchical clustering analysis method for spinning time-series data that integrates time-domain characteristics. This invention proposes a fuzzy hierarchical clustering analysis method for spinning time-series data that integrates time-domain features. It iterates between the DTS feature matrix and the MTS feature matrix, considering the time-domain characteristics and noise effects in the spinning time-series data. Without increasing time complexity, it incorporates the time-frequency characteristics and noise effects in the spinning time-series data during the iteration process. Compared with the latest methods, this invention can more accurately process newly generated time-series data in spinning manufacturing.

The invention relates to the technical field of urban rainfall pattern recognition, and aims to solve the problems that an existing method depends on an empirical formula, rainfall classification is unstable, and interpretability is insufficient. The method comprises the following steps: acquiring multi-site long-time actually measured rainfall data, and identifying an independent rainfall event based on a rainfall interval threshold value; multi-dimensional features such as rainfall duration and peak rainfall intensity are extracted and standardized; performing hierarchical clustering by adopting a Ward minimum variance method, and determining an optimal clustering number by combining a Clinski-Harabasz index and the like to obtain a rainfall mode; training a linear discriminant analysis (LDA) model by taking a clustering result as a label, and verifying and quickly classifying a new rainfall event; the result is used for flood control scheduling, flood risk assessment and the like. According to the method, rationality is improved based on measured data, classification stability is improved by fusing clustering and discriminant analysis, expandability is high, and intelligent disaster prevention is supported.

The invention provides a nearest neighbor search method and device, equipment and a storage medium, and the method comprises the steps: obtaining a retrieval data set, carrying out clustering processing, carrying out clustering again in a first-layer cluster obtained through clustering, generating a hierarchical clustering result comprising a second-layer cluster, and carrying out distance statistics based on the clustering result; obtaining a priori candidate set based on a clustering result; determining an initial answer set and a query radius based on the priori candidate set; performing query pruning processing by using a triangular inequality rule, performing batch pruning based on the first-layer cluster closest to a query point, sequentially performing hierarchical judgment pruning on all the first-layer clusters in a distance increasing order, entering a corresponding second-layer cluster for re-judgment if the first-layer cluster cannot be completely pruned, and returning to the second-layer cluster for re-judgment if the second-layer cluster cannot be completely pruned. Continuously updating the initial answer set in the pruning process; and taking the finally updated answer set as a nearest neighbor search result. According to the method, through combination of hierarchical segmentation and the triangular inequality, the calculation efficiency is remarkably improved.

The application belongs to the technical field of data processing, and discloses a hierarchical clustering method combining distance features, shape features and time sequence features, which comprises: in an initial stage, each time sequence process is regarded as an independent cluster, and the distances between different features of each time sequence process are calculated; a composite similarity index DST of the combined distance, shape and time sequence features is constructed according to the contribution degrees of the distances; the distances between different clusters are calculated according to the constructed composite similarity index and the determined inter-cluster connection mode, and the nearest clusters are continuously iterated and merged; when all the clusters are merged into a large cluster, the clustering is completed, and the optimal clustering number is determined according to the contour coefficients after each merging. The composite similarity index DST is constructed, the weights are automatically calculated according to the contribution degrees of different features to the current time sequence process, the multi-dimensional differences between the time sequence processes can be more comprehensively reflected, and the clustering result is more in line with the actual dynamic evolution law of the time sequence process.

The invention discloses an AI-assisted insomnia typing and personalized intervention system, which belongs to the field of medical artificial intelligence, and comprises a multi-modal data acquisition module used for synchronously acquiring psychological assessment data, physiological index data, clinical diagnosis data and dynamic behavior data of an insomnia patient, the psychological assessment data comprises a DCPR semi-structured interview text, a PSSS scale score, a PSQI scale score, an AIS scale score, an ESS scale score, an HAMA scale score and an HAMD scale score; according to the method, the DCPR semi-structured interview text, the EEG / PSG physiological signal and the wearable device behavior data are subjected to time-space synchronization integration, the semantic features of the psychological text are analyzed through BiLSTM + attention mechanism, and precise classification of healthy anxiety type, persistent psychosomatic type and other psychosomatic syndrome subtypes is realized in combination with hierarchical clustering and random forest algorithms, so that the accuracy of psychosomatic syndrome subtypes is improved, and the accuracy of psychosomatic syndrome subtypes is improved. The limitation of traditional ICSD-3 / DSM-5 single-dimension diagnosis is broken through, and the clinical significance of subtype distinguishing is improved.

The application discloses a method for identifying original tea tree varieties by using characteristic metabolites, and particularly relates to the field of identifying tea tree varieties, which comprises the steps of sample preparation, metabolite extraction, liquid chromatography-mass spectrometry detection, data correction processing, characteristic metabolite construction and multivariate statistical clustering determination, etc.; in view of the internal metabolic characteristics formed in the long-term natural domestication process of the floating Liang chestnut leaf population, a high-dimensional characteristic metabolite vector is constructed by using the stable numerical distribution mode of the fresh leaves of the floating Liang chestnut on the sixteen characteristic metabolites; after the peak intensity data of each sample is corrected, filtered and structured, the obtained characteristic vector is input into the Euclidean distance hierarchical clustering model, so that the metabolic composition difference between the samples to be identified and the standard floating Liang chestnut leaf population is unsupervised statistically distinguished.

According to the heterogeneous resource-oriented virtual power plant hierarchical clustering-cooperative game aggregation and trusted interaction system provided by the invention, refined and structured representation of a strong heterogeneous resource set is realized by constructing a dynamic digital twinborn archive and adaptive clustering of resources, and a hybrid prediction model fusing physical rules and data driving is adopted, so that the reliability of the system is improved, and the reliability of the system is improved. Physical rationality of behavior prediction and quantitative capture of uncertainty are ensured; through a hierarchical distributed optimization architecture of a master-slave-alliance collaborative game, a global complex optimization problem is efficiently decoupled, and real-time collaborative optimal scheduling under large-scale resource access is realized while privacy and autonomy of a resource main body are protected; through an online rolling credibility regulation capability interval evaluation mechanism, a diversified capability vector with a confidence level is output, so that the virtual power plant becomes a reliable transaction object which is transparent to a power grid and has a risk known, and the interaction credibility and credit rating of the virtual power plant in the power market are fundamentally enhanced.

The invention discloses an electric energy meter misalignment evaluation method based on an entropy weight method and related equipment, and relates to the field of intelligent measurement. Comprising the following steps: acquiring periodic verification and field operation data, and constructing feature vectors such as a multi-load / power factor error sequence, temperature / time drift and harmonic sensitivity; hierarchical clustering is adopted, DTW is used for time sequence features, and Mahalanobis distance is used for non-time sequence features to achieve misalignment grade layering; objective weights are obtained according to an entropy weight method, and time-varying adjustment is carried out according to working conditions, service life and environmental factors; establishing a feature vector-precision grade mapping matrix M for grade judgment, and degrading according to quantity and amplitude when a super-threshold component exists; and outputting a recheck instruction and recharging and updating based on a recheck result.

The invention belongs to the technical field of electric automobiles, and particularly relates to a single-pedal driving deceleration intention determination method. Comprising the following steps: S1, acquiring a deceleration and speed data set of a driver; s2, modeling is carried out on the common maximum deceleration of the driver at each vehicle speed; s3, constructing a deceleration statistical feature space of the driver; s4, carrying out clustering analysis on the deceleration statistical characteristics of all test drivers; s5, linear reference mapping of the pedal opening degree and the expected stable vehicle speed in the single pedal mode is established; s6, designing a dynamic critical position of a single-pedal driving deceleration mode; and S7, determining a multi-style single-pedal driving deceleration intention. According to the method, the drivers are scientifically divided into three typical styles of aggressive, common and prudent styles by using a weighted hierarchical clustering method, so that the defect that a single dynamic partition cannot adapt to individual differences in the prior art is overcome, and deceleration intention recognition according with real operation habits of different drivers is realized.

The application discloses a molecular conformation generation method and device based on an evolutionary clustering algorithm. The method optimizes an initial molecular structure input through a local optimization strategy, obtains an optimized initial molecular structure, and generates an initial population based on the optimized initial molecular structure. Clustering analysis is performed through a hierarchical clustering strategy, and a plurality of structure clusters are obtained. Abnormal structure clusters and normal structure clusters are obtained by judging the maximum distance in the clusters. For the abnormal structure clusters, the abnormal structures in the clusters are taken as parents in a set proportion to perform population iteration. For the normal structure clusters, a search strategy is optimized through an adaptive evolution strategy to perform population iteration. An algorithm termination condition is generated based on energy convergence criteria and structure convergence criteria. When the population iteration meets the algorithm termination condition or reaches a maximum iteration number, the iteration is stopped. A molecular conformation meeting energy criteria in the iteration process is output, and a final molecular conformation is obtained. The application can improve the efficiency and intelligent level of structure search.

The application discloses a kind of based on spatial intelligence's active fault three-dimensional model automatic construction method, belong to geological exploration and earthquake engineering technical field;Method includes: collection data, analysis and extract minimum complete subdirectory;Through adaptive threshold hierarchical clustering combined with improved RANSAC algorithm, the exclusive small earthquake cluster of each fault is automatically identified;Based on local weighted regression, three-dimensional automatic slice is made to small earthquake cluster, and each profile fault interpretation line is fitted by moving least square method;Finally, active fault three-dimensional fine model is constructed;Test model rationality and output model file;The application realizes the automation, quantification of fault modeling whole process, can be fused multi-source data and realize multi-element constraint modeling, adapt complex fault zone modeling, and the fine three-dimensional model constructed can provide key data support for fault present-day deformation inversion, three-dimensional potential source research, earthquake geological disaster assessment, with higher engineering application value.

The present application relates to the technical field of jitter analysis, in particular to a high-speed oscilloscope eye diagram real-time jitter detection method and system, comprising the following steps: high-speed acquisition waveform zero-crossing detection obtains timestamp calculation offset, grouping obtains cross-period edge time sequence, according to sequence hierarchical clustering merges jump cluster, calculates deviation, obtains drift trend vector, based on trend vector, assigns compensation weight, weightedly corrects time coordinate, calls corrected coordinate to detect period boundary, calculates difference, triggers correction when difference exceeds threshold, obtains period-corrected edge coordinate, and according to correction coordinate, Fourier analysis is carried out to extract jitter frequency and amplitude to generate eye diagram jitter detection result, in the present application, phase correlation is established based on cross-period sequence, and a resolvable offset mode is formed by identifying the distribution structure through clustering, the cumulative deviation is suppressed through drift trend and weight correction, the abnormal period is calibrated through period dynamic discrimination, and the jitter structure and amplitude relationship are separated through frequency analysis, so that the overall stability and integrity are strengthened.

The application discloses a kind of based on big language model's evolvable software vulnerability detection method, comprising: by regular pattern matching identification Source sentence, based on call graph traversal and data dependence analysis execution function level inter-process slice, build cross-function code context, input the big language model of parameter efficient fine-tuning, output the vulnerability propagation path from Source to Sink;When new vulnerability type needs to be extended, the parameter variation characteristics of old data are extracted by multi-step fine-tuning, and the representative core set is selected by random projection dimension reduction and hybrid distance hierarchical clustering, and the training is played back by mixing new data to alleviate catastrophic forgetting;In the actual use process of tool, the false alarm and the false alarm confirmed by user are collected as feedback signal, the core set is clustered and layered filtered and refined based on perplexity and error prediction analysis, the harmful old knowledge that leads to false alarm and false alarm is removed, and verified new mode is supplemented at the same time, to realize the closed-loop evolution of self-improvement.

This invention relates to a quantum-enhanced clustering method based on spheres, belonging to the field of quantum computing technology. This method addresses the problems of high computational complexity, difficulty in capturing intrinsic geometric features, and sensitivity to noise in traditional clustering algorithms when processing large-scale, high-dimensional, nonlinear data. The technical solution includes obtaining a compressed set of spheres from the original dataset and extracting the center vector; encoding the center vector into a quantum state through quantum feature mapping; constructing a quantum kernel matrix using quantum circuits to calculate fidelity; calculating cohesion based on the quantum kernel matrix to identify and remove noisy spheres; performing hierarchical clustering on the core sphere set; and mapping the results back to the original data points. This invention improves clustering accuracy, robustness, and the ability to handle complex data structures.