Hardware fault information management method and device, electronic equipment and storage medium

Abstract

The invention discloses a hardware fault information management method and device, electronic equipment and a storage medium, and relates to the technical field of server hardware tests.The method comprises the steps that multi-dimensional feature extraction is conducted on original hardware fault information, mixed information is disassembled into feature vectors independent of all dimensions, and the structure of the mixed information is broken; the feature vectors of all dimensions are decoupled to obtain exclusive feature representation, so that the exclusive feature representation has statistical independence in a feature space, and separation of information of different dimensions on physical storage and logic analysis is realized. In this way, cross interference of multi-dimensional information is avoided, and features of all dimensions can be independently analyzed and retrieved. Original information and feature representation of all dimensions are stored in an associated mode, and a structured basis is provided for reuse of historical solutions. The feature representation of each dimension is independent and clear, so that historical cases can be quickly matched based on a specific dimension, and an adaptive solution can be accurately recommended.

Description

Technical Field

[0001] The present application relates to the technical field of server hardware testing, and in particular to a method, device, electronic device, and storage medium for managing hardware fault information. Background Art

[0002] In the server hardware testing process, the discovery and resolution of faults (bugs) is a core component of ensuring hardware product quality. Currently, companies generally manage hardware bugs through a unified bug management system to achieve standardized daily bug control. In practice, when test engineers submit bugs to the R&D stage, hardware developers typically rely on their own experience to explore solutions or manually search for similar hardware issues in the bug management system. However, this approach often consumes a significant amount of time and effort, resulting in inefficient responses to hardware issues and making it difficult to meet the demand for rapid troubleshooting and resolution.

[0003] Especially when dealing with complex hardware defects, related bug management systems suffer from severe information coupling issues. Critical information from multiple dimensions is mixed together for analysis, leading to significant dimensional confusion. This mixed analysis model results in significant discrepancies between the retrieved issues and the target issues, making it more difficult to pinpoint the root causes and effectively reuse historical solutions. This further restricts the efficiency of hardware R&D and testing processes, making it unable to meet the quality control requirements of complex hardware products. Summary of the Invention

[0004] The present application provides a method, device, electronic device and storage medium for managing hardware fault information, so as to at least solve the problem that when related technologies deal with complex hardware defects, there is multi-dimensional key information coupling leading to dimensional confusion, which in turn causes problem retrieval deviation, difficulty in root cause location, low reuse rate of historical solutions, and restricts R&D testing efficiency and quality control.

[0005] The present application provides a method for managing hardware fault information, including: obtaining hardware fault information; performing multi-dimensional feature extraction on the hardware fault information to obtain feature vectors of each dimension; decoupling the feature vectors of each dimension to obtain feature representations of each dimension, wherein the feature representations of each dimension have statistical independence in the feature space; and storing the hardware fault information and the feature representations of each dimension in a database.

[0006] This application also provides a device for managing hardware fault information, including: Acquisition module, used to obtain hardware fault information; Multi-dimensional feature extraction module, used to extract multi-dimensional features of hardware fault information and obtain feature vectors of each dimension; The feature decoupling module is used to decouple the feature vectors of each dimension to obtain the feature representation of each dimension. The feature representation of each dimension has statistical independence in the feature space; The storage module is used to store hardware fault information and feature representations of each dimension in the database.

[0007] The present application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for implementing the steps of the above-mentioned method for managing hardware fault information when executing the computer program.

[0008] The present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned method for managing hardware fault information are implemented.

[0009] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned hardware fault information management method when executed by a processor.

[0010] This application extracts multi-dimensional features from the original hardware fault information, decomposes the originally mixed hardware fault information into independent feature vectors of each dimension, and breaks up the originally mixed information structure; then, by decoupling the feature vectors of each dimension, a unique feature representation for each dimension is obtained. The feature representation of each dimension has statistical independence in the feature space, so that information of different dimensions can be separated from physical storage and logical analysis. This processing method avoids the cross-interference of multi-dimensional fault information, so that the features of each dimension can be analyzed and retrieved independently. By associating the original hardware fault information with the feature representation of each dimension for storage, a structured basis is provided for the reuse of historical solutions. Since the feature representation of each dimension is independent and clear, historical cases can be quickly matched based on specific dimensions, which is conducive to accurately recommending adapted solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0012] Figure 1 A schematic diagram of a specific hardware architecture on which the execution of a hardware fault information management method provided in an embodiment of the present application relies; Figure 2 A flowchart of a method for managing hardware fault information provided in an embodiment of the present application; Figure 3A schematic diagram of the structure of a device for managing hardware fault information provided in an embodiment of the present application; Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0013] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0014] It should be noted that, in the description of this application, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. The terms "first," "second," etc., in this application are used to distinguish similar objects, and are not used to describe a particular order or sequence.

[0015] In order to more clearly illustrate the embodiments of the present application, the following briefly introduces the technical terms used in the embodiments: Principal Component Analysis (PCA) is an unsupervised dimensionality reduction technique. Its core function is to convert high-dimensional data into a low-dimensional representation while retaining the main variability of the data and achieving decoupling between features.

[0016] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0017] In conjunction with the specific application environment architecture or specific hardware architecture on which the execution of the hardware fault information management method depends, the specific application environment architecture or specific hardware architecture is described herein.

[0018] like Figure 1 As shown, Figure 1 A schematic diagram of the specific hardware architecture that the hardware fault information management method relies on for execution. The hardware architecture includes a fault information collection module, a feature extraction module, a feature vector construction module, a feature splitting module, and a storage and retrieval module.

[0019] The fault information collection module is used to collect multi-dimensional raw data related to hardware faults, including hardware module characteristics, fault symptom descriptions, and environmental parameters. This data is obtained through multiple channels, such as hardware testing tools, system logs, and manufacturer documentation, providing the basis for subsequent analysis.

[0020] The feature extraction module extracts key features from data subsets across various dimensions, transforming unstructured information into quantifiable feature representations. It performs structured encoding of hardware module characteristics, extracting numerical features such as the number of CPU cores and memory frequency. It applies natural language processing techniques to fault symptom text to extract keywords and their weights. It also performs numerical analysis of environmental parameters to extract statistical features such as temperature extremes and humidity ranges.

[0021] The feature vector building module is used to quantify the extracted features into mathematical vectors.

[0022] The feature splitting module uses methods such as principal component analysis to decouple and reduce the dimensionality of feature vectors, eliminating redundancy and correlation between features. By calculating the covariance matrix and decomposing the eigenvalues, the original features are mapped to a new orthogonal space, resulting in a more independent and discriminative feature representation.

[0023] The storage and retrieval module is used to establish a dedicated index structure for each dimensional feature representation to achieve efficient fault information retrieval and matching.

[0024] An embodiment of the present application provides a method for managing hardware fault information, and the method is described in detail in conjunction with the execution flow of the method for managing hardware fault information.

[0025] like Figure 2 As shown, Figure 2 A flowchart of a method for managing hardware fault information provided in an embodiment of the present application is provided. The method includes the following steps S201 to S204: S201: Obtain hardware failure information.

[0026] Hardware fault information can be submitted by test engineers to the fault management system, which can also collect detailed information based on the submitted information.

[0027] In some embodiments, the hardware fault information includes a first data subset corresponding to a hardware module characteristic dimension, a second data subset corresponding to a fault phenomenon description text dimension, and a third data subset corresponding to an environmental parameter dimension.

[0028] Optionally, when obtaining the first data subset corresponding to the hardware module characteristic dimension, detailed information about each module in the server hardware is collected, including but not limited to core component information and electrical parameters. Core component information includes but is not limited to: processor model, number of cores, main frequency, cache size; memory system model, capacity, frequency, slot configuration, and bandwidth; storage device model, capacity, interface type, and read / write speed; motherboard model, chipset, expansion slot specifications, and Basic Input / Output System (BIOS) version; power module model, rated power, conversion efficiency, and output voltage rail; cooling system radiator model, fan specifications, and cooling method; and expansion card model, interface type, and performance parameters. Electrical parameters include but are not limited to: operating voltage range, maximum current demand, and peak current fluctuation range of each component; static power consumption, full-load power consumption, and power consumption distribution of each component; and thermal design indicators for each component.

[0029] This information can be directly read through hardware detection tools, remotely obtained through system management interfaces, or obtained from technical documentation provided by hardware manufacturers.

[0030] Optionally, when obtaining the second data subset corresponding to the fault phenomenon description text dimension, the log text generated by the server when the fault occurs is obtained. This log text contains key content such as the time of the fault, real-time status information of related components, error codes reported by the system, and a detailed description of the abnormal phenomenon. This log text can be extracted from a wide range of sources, including the server's operating system logs, logs generated during application execution, and log information recorded by hardware monitoring systems.

[0031] Optionally, when obtaining the third data subset corresponding to the environmental parameter dimension, various environmental parameters of the server during operation in the BUG information are recorded, including but not limited to temperature parameters, humidity parameters, and voltage parameters. Temperature parameters include the ambient temperature in the computer room and the real-time surface temperature of the hardware equipment. Humidity parameters refer to the air humidity in the server operating space. Voltage parameters include the input voltage of the power supply system and the actual voltage of key components during operation.

[0032] S202, extracting features from the hardware fault information to obtain feature vectors of various dimensions; Hardware fault information involves different dimensions. When executing step S202, the hardware fault information can first be dimensionally separated according to at least one of the following dimensions: hardware module characteristics, fault phenomenon description text, and environmental parameters, to obtain at least one data subset. Feature extraction is then performed on the at least one data subset to obtain feature vectors for each dimension.

[0033] The above embodiment structuredly separates multi-dimensional hardware fault information, breaking down mixed fault information into independent subsets by dimension. This breaks the traditional mixed analysis model and addresses the dimensionality confusion issue at the source. Feature extraction is then performed on each data subset to obtain feature vectors for the corresponding dimension. This allows key information from different dimensions to be presented as structured features, avoiding cross-information interference.

[0034] From a technical perspective, dimensional separation eliminates information coupling, making the analysis of information in each dimension more targeted; structured feature vectors facilitate the system to accurately locate the root cause of the problem, while providing a clear dimensional basis for the classification, storage, and reuse of historical solutions, which is conducive to improving the efficiency of solution reuse.

[0035] In some embodiments, data subsets of different dimensions are preprocessed, including but not limited to missing value supplementation, deduplication, word segmentation, and correction or removal of outliers. For example, missing model numbers and repeated parameters in module characteristics are supplemented and deduplicated; fault phenomenon text is word segmented and stop words are removed; and outliers in environmental parameters, such as temperature values ​​outside a reasonable range, are corrected or removed.

[0036] In some embodiments, feature extraction is performed on a first data subset corresponding to a hardware module characteristic dimension to obtain a first feature vector corresponding to the hardware module characteristic dimension, including: performing structured coding on the first data subset to extract key module characteristic features, including but not limited to CPU features and memory features. Among them, CPU features include the number of cores, main frequency, cache size, etc. Memory features include capacity, frequency, type, etc. Then, the key module characteristic features are quantized to obtain a first feature vector X=(x1,x2,…x n ), where x i Indicates the specific quantified characteristic values ​​of each hardware module.

[0037] For example, discrete attributes are one-hot encoded. For example, if the module type is motherboard, it is encoded as 100, and if the module type is power supply, it is encoded as 010. Continuous parameters, such as interface transmission rate, are normalized and mapped to the interval [0, 1]. Alternatively, inter-module correlation features, such as the motherboard and memory compatibility parameters, are extracted to construct a module relationship matrix.

[0038] The above embodiment achieves effective stripping and quantification of key information by performing structured processing on the first data subset corresponding to the hardware module characteristic dimension. Specifically, the above embodiment first performs structured coding on the first data subset, focusing on key module characteristic features such as the number of CPU cores, main frequency, cache size, and memory capacity, frequency, and type, and then converts these features into a computable first feature vector X=(x1,x2,…x nThis processing method breaks the mixing of hardware module feature information and other dimensional information, avoiding the interference of dimensional confusion on the analysis process.

[0039] From a technical perspective, structured coding and quantization processing separate the hardware module characteristic information from complex mixed data, forming a standardized feature vector, providing a clear dimensional basis for subsequent precise retrieval and analysis, and reducing matching deviations caused by information coupling.

[0040] In some embodiments, feature extraction is performed on the second data subset corresponding to the fault phenomenon description text dimension to obtain a second feature vector corresponding to the fault phenomenon description text dimension, including: analyzing the second data subset based on natural language processing technology, extracting key fault phenomenon features through a keyword extraction algorithm. Then, these key fault phenomenon features are quantified to obtain the second feature vector Y corresponding to the fault phenomenon description text dimension = (y1, y2, ...y m ), where y j The second feature vector Y may include the weight value or the frequency of occurrence of each keyword feature.

[0041] For example, a bag-of-words model or a Term Frequency-Inverse Document Frequency (TF-IDF) algorithm is used to convert keywords such as "black screen" and "abnormal noise" into weight vectors; semantic features are extracted through a pre-trained language model to capture the synonymous relationship between "unable to start" and "boot failure"; and fault phenomenon classification labels (such as "startup category" and "heat dissipation category") are constructed as a dimensionality reduction representation of high-dimensional features.

[0042] The above embodiment uses natural language processing technology to perform special processing on the second data subset corresponding to the text dimension of the fault phenomenon description, thereby achieving the structuring and quantification of text information. Specifically, the above embodiment first analyzes the second data subset based on natural language processing technology, captures the fault phenomenon features with the help of keyword extraction algorithm, and then quantifies these features into a second feature vector Y=(y1,y2,…y m This processing method converts the originally vague text description into computable numerical features, completely decoupling the fault phenomenon text from other dimensional information such as hardware module characteristics and environmental parameters.

[0043] From a technical perspective, keyword extraction and quantification allow the fault phenomenon description to be separated from mixed information, avoiding analysis bias caused by confusion between text description and other dimensional information, and enabling the system to accurately focus on the characteristics of the fault phenomenon itself.

[0044] In some embodiments, extracting features from the third data subset corresponding to the environmental parameter dimension to obtain a third feature vector corresponding to the environmental parameter dimension includes: performing numerical analysis on the third data subset to extract key environmental features, including but not limited to the average value, maximum value, minimum value of temperature, range of humidity, and voltage fluctuation. Then, quantizing these key environmental features to obtain a third feature vector Z corresponding to the environmental parameter dimension = (z1, z2, ... z p ), where z k Indicates the characteristic values ​​of each specific quantified environmental parameter.

[0045] For example, parameter distribution characteristics are calculated, such as the temperature mean and voltage standard deviation; abnormal fluctuation characteristics are extracted, such as the time difference between the voltage surge and the fault occurrence time; and correlation analysis is performed on multi-parameter combinations, such as the synergistic influence coefficient of high temperature and high humidity.

[0046] The above embodiment achieves accurate stripping and quantification of environmental information by performing special processing on the third data subset corresponding to the environmental parameter dimension. Specifically, the above embodiment first performs numerical analysis on the third data subset to extract key environmental characteristics such as the average value, maximum value, minimum value of temperature, range of humidity, voltage fluctuation, etc., and then quantifies these characteristics into a third eigenvector Z=(z1,z2,…z p This processing method breaks the coupling between environmental parameters and information such as hardware module characteristics and fault phenomenon descriptions, avoiding the interference of dimensional confusion on the analysis process.

[0047] From a technical perspective, the extraction and quantification of key environmental features allows environmental parameters to be separated from complex mixed data, forming standardized feature vectors. This provides a clear basis for subsequent precise analysis of environmental factors and reduces problem matching deviations caused by information congestion.

[0048] In some embodiments, variance selection, mutual information methods, and the like can be used to eliminate low-contribution features and retain core features. For example, module redundancy parameters unrelated to the fault can be eliminated while retaining core features such as power supply voltage fluctuations and overheat protection triggering.

[0049] S203: Decouple the feature vector to obtain feature representations of each dimension.

[0050] The feature representations of each dimension are statistically independent in the feature space. Feature representations of different dimensions are linearly independent and do not interfere with each other in the feature space. Each feature representation is low-dimensional and standardized. Optionally, a feature decoupling algorithm can be used to decouple the feature vectors to obtain feature representations of each dimension. This makes the dimensions independent and decoupled, preserving the key content of hardware fault information.

[0051] In some embodiments, the principal component analysis method is used to decouple the eigenvectors of each dimension. The eigenvectors of each dimension are first normalized, and the covariance matrix of the eigenvectors of each dimension after normalization is calculated; then the covariance matrix is ​​decomposed to obtain the eigenvalues ​​and the columns of the projection matrix; further, the projection matrix is ​​constructed based on the eigenvalues ​​and the columns of the projection matrix; finally, the eigenvectors are mapped to a new feature space through a projection transformation to obtain the feature representation of each dimension.

[0052] For example, assume that the original feature vector is V = (v1, v2, ... v q ), the feature after feature splitting is expressed as V`=(v1`,v2`,…v q `), V`=W T V, where W is the projection matrix, which is obtained by solving the eigenvector of the covariance matrix so that the split eigenvector has the maximum variance in the new space.

[0053] Specifically, the eigenvectors of each dimension are normalized first, and the covariance matrix C of the eigenvector V of each dimension after normalization is calculated. Then, the covariance matrix C is decomposed to obtain the eigenvalue λ and the column ω of the projection matrix, where the eigenvalue λ represents the variance in the direction of the corresponding eigenvector. Based on the eigenvalue λ and the column ω of the projection matrix, the projection matrix W is constructed. Further, according to V`=W T V performs a projection transformation on the feature vector V to obtain the feature representation V' of each dimension.

[0054] The above embodiment first normalizes the feature vectors of various dimensions such as hardware module characteristics, fault phenomenon description, and environmental parameters to eliminate the interference of features of different magnitudes, and then calculates the covariance matrix to quantify the correlation between features. Then, the eigenvalues ​​and columns of the projection matrix are obtained by decomposing the covariance matrix, and the projection matrix W is constructed based on this. Finally, V`=W T The projection transformation of V maps the original feature vectors into a new feature space, obtaining the feature representations V' for each dimension. This process removes the implicit coupling between features by retaining the features in the direction of maximum variance, making the separated features more independent in the new space. For example, the feature vectors of hardware module characteristics and environmental parameters no longer interfere with each other in the new space, fundamentally resolving the underlying problem of dimensional confusion.

[0055] From a technical perspective, principal component analysis effectively reduces the redundancy of feature dimensions, making the physical meaning of features in each dimension clearer, avoiding analytical biases caused by implicit associations, and improving the accuracy of problem retrieval and matching; the feature representation in the new space retains the key information of the original features and is easier to calculate. Low-dimensional feature representation facilitates rapid fault retrieval and reduces the search space. This eliminates the need to traverse the entire data during retrieval, and instead allows for quick location of relevant subsets through dimensional indexing, improving retrieval efficiency by orders of magnitude. Standardized feature representation makes similar faults easier to identify. Decoupled feature representation enables the system to more accurately recommend historical solutions, improve reuse rates, and reduce manual troubleshooting time.

[0056] S204: Store the hardware fault information and the feature representations of each dimension in a database in correspondence.

[0057] For example, in the database, the feature representations "specific motherboard" and "high temperature environment" are stored correspondingly with the hardware fault information "power failure." The feature representations "memory error" and "temperature anomaly" are stored correspondingly with the fault information "heat dissipation problem."

[0058] In some embodiments, independent indexes are established for the feature representations of each dimension: a hash index is established for the first feature representation of the hardware module feature; an inverted index is established for the second feature representation of the fault phenomenon description text; and a range index is established for the third feature representation of the environmental parameter.

[0059] The above embodiment establishes a hash index for the first feature representation of the hardware module feature, and uses the fast mapping capability of the hash algorithm to improve the efficiency of hardware parameter matching; establishes an inverted index for the second feature representation of the fault phenomenon description text, and enhances the accuracy of fault phenomenon retrieval through the association mapping between keywords and text; establishes a range index for the third feature representation of the environmental parameter, and adapts to the interval characteristics of parameters such as temperature, humidity, and voltage, so as to facilitate the rapid location of environmental data that meets specific range conditions. This method of customizing indexes by dimension breaks the situation where different types of features in the original mixed index interfere with each other, and allows the retrieval of features of each dimension to be carried out efficiently based on its own attributes. Exclusive index structures designed for different feature types are conducive to improving retrieval speed.

[0060] In some embodiments, the method further includes obtaining query conditions input by the user, and determining target hardware fault information and target feature representation that match the query conditions from a database based on the query conditions; and then determining a solution strategy based on the target hardware fault information and target feature representation.

[0061] Specifically, the system parses the query conditions entered by the user into multi-dimensional feature vectors, including hardware module characteristics, fault symptom descriptions, and environmental parameters. It then uses independent indexes to search the database in parallel. It then calculates the similarity between the user's query feature vector and the target feature representation in the database. It then sorts candidate cases by similarity and extracts corresponding solutions based on the top N target cases with the highest similarity.

[0062] For example, suppose a user enters the query "CPU temperature overheating causes system crash." The analysis yields multi-dimensional feature vectors, including hardware module characteristics (CPU model, temperature threshold), fault symptom description (system crash, error code), and environmental parameters (computer room temperature, voltage fluctuation). A hash index is used to quickly filter historical cases that meet the CPU model, temperature threshold, and other criteria. An inverted index is used to match keywords and semantic similarity. A range index is used to locate cases within a range of parameters such as temperature and voltage. This allows for the identification of solutions, such as replacing the cooling fan or adjusting memory timings.

[0063] The above embodiment separates the search paths for hardware, fault symptoms, and environmental parameters through independent indexing, avoiding the interference of different dimensional information in traditional mixed indexing. Parallel index querying improves search accuracy and efficiency. By combining multi-dimensional indexing with feature vector matching, an efficient fault solution recommendation system is constructed, which not only solves the dimensional confusion problem of traditional bug management systems but also significantly improves search efficiency and solution reuse.

[0064] In summary, the embodiment of the present application provides a method for managing hardware fault information, which extracts multi-dimensional features from the original hardware fault information, decomposes the originally mixed hardware fault information into independent feature vectors of each dimension, and breaks up the originally mixed information structure; then, by decoupling the feature vectors of each dimension, a unique feature representation of each dimension is obtained, and the feature representation of each dimension has statistical independence in the feature space, so that information of different dimensions can be separated from physical storage and logical analysis. This processing method avoids the cross-interference of multi-dimensional fault information, so that the features of each dimension can be analyzed and retrieved independently. By associating the original hardware fault information with the feature representation of each dimension for storage, a structured basis is provided for the reuse of historical solutions. Since the feature representation of each dimension is independent and clear, historical cases can be quickly matched based on a specific dimension, which is conducive to accurately recommending adapted solutions.

[0065] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0066] like Figure 3As shown, an embodiment of the present application further provides a device for managing hardware fault information, the device comprising: Acquisition module 301, used to obtain hardware fault information; A multi-dimensional feature extraction module 302 is used to extract multi-dimensional features from hardware fault information to obtain feature vectors of each dimension; A feature decoupling module 303 is used to decouple the feature vectors of each dimension to obtain feature representations of each dimension, where the feature representations of each dimension have statistical independence in the feature space; The storage module 304 is used to store the hardware fault information and the feature representations of each dimension in a database.

[0067] As an optional implementation provided in an embodiment of the present application, the multi-dimensional feature extraction module 302 is specifically used to: separate the hardware fault information according to at least one dimension of hardware module characteristics, fault phenomenon description text and environmental parameters to obtain at least one data subset; perform feature extraction on at least one data subset to obtain feature vectors of each dimension.

[0068] As an optional implementation provided in an embodiment of the present application, at least one data subset includes a first data subset corresponding to the hardware module characteristic dimension; the multi-dimensional feature extraction module 302 is specifically used to: perform structured encoding on the first data subset to extract key module characteristic features; and quantify the key module characteristic features to obtain a first feature vector of the hardware module characteristic dimension.

[0069] As an optional implementation method provided in an embodiment of the present application, the hardware fault information includes a second data subset corresponding to the fault phenomenon description text dimension; the multi-dimensional feature extraction module 302 is specifically used to: analyze the second data subset based on natural language processing technology to extract key fault phenomenon features; quantify the key fault phenomenon features to obtain a second feature vector of the fault phenomenon description text dimension.

[0070] As an optional implementation method provided in an embodiment of the present application, the hardware fault information includes a third data subset corresponding to the environmental parameter dimension; the multi-dimensional feature extraction module 302 is specifically used to: perform numerical analysis on the third data subset to obtain key environmental features; and quantify the key environmental features to obtain a third feature vector of the environmental parameter dimension.

[0071] As an optional implementation provided in an embodiment of the present application, the feature decoupling module 303 is specifically used to: normalize the eigenvectors of each dimension; calculate the covariance matrix of the eigenvectors of each dimension after normalization; decompose the covariance matrix to obtain eigenvalues ​​and columns of the projection matrix; construct a projection matrix based on the eigenvalues ​​and columns of the projection matrix; and perform a projection transformation on the eigenvectors of each dimension based on the projection matrix to obtain feature representations of each dimension.

[0072] As an optional implementation provided in an embodiment of the present application, the hardware fault information management device also includes a retrieval module, which is used to: obtain query conditions input by the user; determine the target hardware fault information and target feature representation that match the query conditions from the database based on the query conditions; and determine a solution strategy based on the target hardware fault information and target feature representation.

[0073] For the description of the features in the embodiment corresponding to the hardware fault information management device, please refer to the relevant description of the embodiment corresponding to the hardware fault information management method, which will not be repeated here.

[0074] like Figure 4 As shown, an embodiment of the present application also provides an electronic device, including a memory 401 and a processor 402, wherein the memory 401 stores a computer program, and the processor 402 is configured to run the computer program to execute the steps in any of the above-mentioned hardware fault information management method embodiments.

[0075] An embodiment of the present application further provides a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps of any of the above-mentioned hardware fault information management method embodiments when running.

[0076] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0077] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of any of the above-mentioned hardware fault information management method embodiments are implemented.

[0078] An embodiment of the present application also provides another computer program product, including a non-volatile computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps in any of the above-mentioned hardware fault information management method embodiments.

[0079] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0080] The above is a detailed introduction to a method, device, electronic device and storage medium for managing hardware fault information provided by the present application. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A method for managing hardware fault information, characterized in that: include: Get hardware failure information; Performing multi-dimensional feature extraction on the hardware fault information to obtain feature vectors of each dimension; Decoupling the feature vectors of each dimension to obtain feature representations of each dimension, wherein the feature representations of each dimension have statistical independence in the feature space; The hardware fault information and the feature representations of each dimension are correspondingly stored in a database.

2. The method according to claim 1, characterized in that The multi-dimensional feature extraction of the hardware fault information to obtain feature vectors of each dimension includes: Separate the hardware fault information according to at least one dimension of hardware module characteristics, fault phenomenon description text, and environmental parameters to obtain at least one data subset; Feature extraction is performed on at least one data subset to obtain feature vectors of each dimension.

3. The method according to claim 2, characterized in that The at least one data subset includes a first data subset corresponding to a hardware module characteristic dimension; The extracting features from at least one data subset to obtain the feature vectors of each dimension includes: Performing structured coding on the first data subset to extract key module characteristic features; The key module characteristic features are quantified to obtain a first characteristic vector of the hardware module characteristic dimension.

4. The method according to claim 2, characterized in that The hardware fault information includes a second data subset corresponding to a fault phenomenon description text dimension; The extracting features from at least one data subset to obtain the feature vectors of each dimension includes: Analyzing the second data subset based on natural language processing technology to extract key fault phenomenon features; The key fault phenomenon features are quantified to obtain a second feature vector of the fault phenomenon description text dimension.

5. The method according to claim 2, characterized in that The hardware fault information includes a third data subset corresponding to the environmental parameter dimension; The extracting features from at least one data subset to obtain the feature vectors of each dimension includes: performing numerical analysis on the third data subset to obtain key environmental features; The key environmental features are quantified to obtain a third eigenvector of the environmental parameter dimension.

6. The method according to claim 1, characterized in that Decoupling the feature vectors of each dimension to obtain feature representations of each dimension includes: Normalizing the feature vectors of each dimension; Calculating the covariance matrix of the normalized eigenvectors of each dimension; Decomposing the covariance matrix to obtain eigenvalues ​​and columns of a projection matrix; constructing a projection matrix according to the eigenvalues ​​and the columns of the projection matrix; The feature vectors of each dimension are projected and transformed according to the projection matrix to obtain feature representations of each dimension.

7. The method according to claim 1, characterized in that The method further comprises: Get the query conditions entered by the user; According to the query condition, determining target hardware fault information and target feature representation matching the query condition from the database; A solution strategy is determined based on the target hardware fault information and target feature representation.

8. A device for managing hardware fault information, characterized in that: include: Acquisition module, used to obtain hardware fault information; A multi-dimensional feature extraction module is used to extract multi-dimensional features of the hardware fault information to obtain feature vectors of each dimension; A feature decoupling module is used to decouple the feature vectors of each dimension to obtain feature representations of each dimension, wherein the feature representations of each dimension have statistical independence in the feature space; The storage module is used to store the hardware fault information and the feature representations of each dimension in a database.

9. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the method for managing hardware fault information according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the method for managing hardware fault information according to any one of claims 1 to 7 are implemented.

Images