Industrial talent intelligent matching method and system based on AI-RAG

Abstract

The invention relates to the technical field of artificial intelligence and human resources, and provides an industrial talent intelligent matching method and system based on AI-RAG, and the method comprises the steps: generating a talent portrait according to post demand information inputted by a user, the portrait comprises a unit feature array sequence, feature similarity arrays arranged in a descending order corresponding to each array, and matching object information, matching objects and industrial data information are included; after a portrait is displayed in a matching window, determining a target unit feature array sequence in response to a selection operation, and grouping matching object information sets according to industrial data identifiers; and determining a target matching object information group and displaying the corresponding industrial data in a data display window, wherein the matching object is highlighted. The industrial data information comprises identification and position information, a feature similarity array is generated through cumulative similarity threshold screening, matching object information is generated through an attention value array of a core AI-RAG attention module, and target group updating and data display through switching operation are supported.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and human resources technology, and more specifically, to an AI-RAG-based industrial talent intelligent matching method and system. Background Art

[0002] In today's rapidly developing industrial landscape, with industry upgrades and technological innovations, companies' demands for industrial talent are becoming increasingly diverse and specialized. To meet companies' needs for precise recruitment and efficient talent placement, traditional industrial talent matching methods rely primarily on manual resume screening and simple keyword matching techniques. However, this approach has numerous limitations. On the one hand, manual screening is inefficient, making it difficult to quickly and accurately identify qualified candidates from a vast volume of resumes. On the other hand, simple keyword matching techniques fail to fully understand the complex semantic relationship between job requirements and talent skills, easily leading to inaccurate matches or missing potentially suitable candidates. Furthermore, most existing talent matching systems lack the ability to deeply mine industrial domain expertise and data, making it impossible to fully leverage industrial data to improve matching accuracy and efficiency.

[0003] With the continuous development of artificial intelligence (AI), technologies such as natural language processing and machine learning have gradually been introduced into the field of talent matching. These technologies can better understand and process textual information, thereby improving the accuracy and efficiency of matching. However, existing AI-based talent matching methods still have some shortcomings. For example, while some methods can provide a certain degree of semantic understanding of text, when dealing with complex industrial domain knowledge and data, they often fail to fully utilize the structured information of industrial data, resulting in inaccurate matching results. Other methods, while attempting to incorporate industrial data for matching, lack effective feature extraction and similarity calculation mechanisms, making them unable to accurately identify and match the key skills of talents with job requirements.

[0004] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the existing technology: the existing industrial talent matching methods cannot fully combine the professional knowledge and data in the industrial field, making it difficult to achieve accurate matching; the existing matching technology cannot accurately identify and match key features when dealing with complex industrial job requirements and talent skill descriptions, resulting in low matching efficiency and accuracy; the existing matching system lacks an effective user interaction mechanism, and cannot flexibly adjust the matching results according to user needs and provide intuitive display. Summary of the Invention

[0005] The present invention provides an AI-RAG-based intelligent matching method and system for industrial talents.

[0006] In a first aspect of the present invention, a method for intelligent matching of industrial talents based on AI-RAG is provided, comprising: Generate a corresponding talent profile based on the job requirement information input by the user. The talent profile includes a sequence of unit feature arrays, each unit feature array corresponds to a feature similarity array, each feature similarity in the feature similarity array corresponds to matching object information, the matching object information includes the matching object and corresponding industrial data information, and the industrial data information includes an industrial data identifier and the location information of the matching object in the industrial data corresponding to the industrial data identifier; Displaying the talent profile in a preset position corresponding to the job requirement information in the matching window; In response to detecting a selection operation acting on any feature array in the talent profile, determining at least one unit feature array included in the feature array as a target unit feature array sequence, the feature array including at least one unit feature array; For each target unit feature array in the target unit feature array sequence, determining each piece of matching object information corresponding to the target unit feature array as a matching object information set; According to each industrial data information included in each obtained matching object information set, the matching object information set is divided into matching object information group sets, each matching object information in each matching object information group corresponds to the same industrial data identifier; Determining a target matching object information group; The industrial data corresponding to the target matching object information group is displayed in the data display window corresponding to the matching window. The industrial data corresponds to the industrial data identifier corresponding to the target matching object information group. In the displayed industrial data, each matching object included in the target matching object information group is in a highlighted state.

[0007] Furthermore, the talent portrait includes text information, each unit feature array included in the feature array is a feature segment; and generating a talent portrait corresponding to the job requirement information according to the job requirement information input by the user includes: Generate a reference segment information set based on the job requirement information, wherein each reference segment information in the reference segment information set corresponds to an industrial data identifier and segment location information, and the reference segment information set includes reference segment information corresponding to a text type; Generating a talent profile corresponding to the job requirement information based on the job requirement information and the reference fragment information set; For each feature segment included in the talent profile, perform the following steps: In response to determining that the feature segment is not the first segment in the talent profile, generating an initial feature similarity array based on the job requirement information, the reference segment information set, and a previous feature segment of the feature segment in the talent profile, wherein matching objects corresponding to feature similarities in the initial feature similarity array include matching words; Sorting each feature similarity in the initial feature similarity array in descending order to obtain a descending feature similarity array; Determine the preset initial value as the cumulative feature similarity; Determining the first feature similarity in the descending feature similarity array as the target feature similarity; Based on the cumulative feature similarity and the target feature similarity, the following loop steps are performed: Determining the sum of the cumulative feature similarity and the target feature similarity as the cumulative feature similarity, so as to update the cumulative feature similarity; In response to determining that the updated cumulative feature similarity is greater than a preset similarity threshold, concatenating the feature similarities preceding the target feature similarity in the descending feature similarity array and the target feature similarity into a screening feature similarity array; In response to determining that the updated cumulative feature similarity is less than or equal to the preset similarity threshold, determining the next feature similarity of the target feature similarity in the descending feature similarity array as the target feature similarity to update the target feature similarity, and executing the looping step again based on the updated cumulative feature similarity and the updated target feature similarity; According to the spliced screening feature similarity array, a feature similarity array corresponding to the feature segment and each matching object information is generated.

[0008] Furthermore, the generating of feature similarity arrays corresponding to the feature segments and information of each matching object based on the spliced screening feature similarity arrays includes: Determining the number of each feature similarity included in the screening feature similarity array as the similarity number; In response to determining that the number of similarities is greater than a preset number, extracting the preset number of feature similarities from the screening feature similarity array as a feature similarity array corresponding to the feature segment according to a preset extraction method; For each feature similarity in the feature similarity array, perform the following steps: Determining matching objects corresponding to the feature similarity; determining the reference segment information to which the determined matching object belongs as target reference segment information; Determine the industrial data identifier of the industrial data corresponding to the target reference segment information; Determining location information of the determined matching object in the industrial data corresponding to the industrial data identifier; Determining the determined industrial data identifier and location information as industrial data information; The determined matching object and industrial data information are determined as matching object information corresponding to the feature similarity.

[0009] Furthermore, the talent profile is obtained by inputting the job requirement information into a pre-trained AI-RAG intelligent matching model, the AI-RAG intelligent matching model including various decoding levels, each decoding level including various AI-RAG attention modules, each AI-RAG attention module being used to generate an attention value array for a corresponding feature segment based on the job requirement information, a reference segment information set, and a feature segment; and generating an initial feature similarity array based on the job requirement information, the reference segment information set, and the previous feature segment of the feature segment in the talent profile, including: Determining decoding levels corresponding to a preset number of levels among the decoding levels as key decoding levels; Determine the AI-RAG attention modules included in the key decoding level as a core attention module set; An initial feature similarity array is generated based on each attention value array corresponding to the feature fragment generated by each core attention module in the core attention module set.

[0010] Furthermore, determining the target matching object information group includes: determining the first matching object information group in the matching object information group set as a target matching object information group; In response to detecting the industrial data switching operation, according to the target matching object information group, determining the matching object information group corresponding to the industrial data switching operation in the matching object information group set as the target matching object information group, so as to update the target matching object information group; The industrial data displayed in the data display window is switched to the industrial data corresponding to the updated target matching object information group, wherein in the switched industrial data, each matching object included in the updated target matching object information group is in a highlighted state.

[0011] Furthermore, displaying the industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window includes: Determining whether a matching object included in the target matching object information group exists in the current data display page displayed in the data display window; In response to determining that a matching object included in the target matching object information group exists in the current data display page displayed in the data display window, highlighting each matching object included in the target matching object information group existing in the current data display page, wherein a matching object switching control is also displayed in the current data display page; In response to detecting a selection operation corresponding to the matching object switching control, determining a switching matching object according to the target matching object information group and the currently selected matching object; Position the current data display page to the page position corresponding to the switching matching object.

[0012] In a second aspect of the present invention, an AI-RAG-based industrial talent intelligent matching system is provided, comprising: a generating unit, which generates a corresponding talent profile based on the job requirement information input by the user, wherein the talent profile includes a sequence of unit feature arrays, each unit feature array corresponds to a feature similarity array, each feature similarity in the feature similarity array corresponds to matching object information, the matching object information includes a matching object and corresponding industrial data information, and the industrial data information includes an industrial data identifier and location information of the matching object in the industrial data corresponding to the industrial data identifier; a first display unit, displaying the talent portrait at a preset position corresponding to the job requirement information in the matching window; a first determining unit, in response to detecting a selection operation acting on any feature array in the talent profile, determining at least one unit feature array included in the feature array as a target unit feature array sequence; a second determining unit, for each target unit feature array in the target unit feature array sequence, determining each piece of matching object information corresponding to the target unit feature array as a matching object information set; a dividing unit, which divides each matching object information set into matching object information group sets according to each industrial data information included in each matching object information set, wherein each matching object information in each matching object information group corresponds to the same industrial data identifier; A third determining unit determines a target matching object information group; The second display unit displays the industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window. The industrial data corresponds to the industrial data identifier corresponding to the target matching object information group. In the displayed industrial data, each matching object included in the target matching object information group is in a highlighted state.

[0013] In a third aspect of the present invention, an electronic device is provided, comprising: at least one processor, a memory, and an input / output unit; wherein the memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute any one of the methods described in the first aspect.

[0014] In a fourth aspect of the present invention, a computer-readable medium is provided, comprising instructions, which, when executed on a computer, causes the computer to execute the method according to any one of the first aspects.

[0015] In a fifth aspect of the present invention, a computer program product is provided, comprising a computer program, wherein the computer program implements the method according to any one of the first aspects when executed by a processor.

[0016] The above embodiments of the present invention have at least the following beneficial effects: (1) This invention achieves accurate matching of industrial talents and job requirements based on the AI-RAG model. By generating a talent portrait containing a sequence of unit feature arrays and combining feature similarity calculation with industrial data association analysis, the intelligent level of matching is improved. (2) The present invention adopts a descending feature similarity array and a cumulative similarity threshold screening mechanism to ensure the accuracy and reliability of the matching results. At the same time, through the association mapping of industrial data identification and location information, a deep association between talent characteristics and industrial data can be established; (3) The present invention realizes the dynamic interactive display of talent portraits and industrial data. Through the linkage design of the matching window and the data display window, the matching results can be presented intuitively and interactive operations can be supported; (4) In the present invention, users can trigger the matching process by selecting a feature array. The system automatically groups and filters the target matching object information group and highlights the matching object in the data display window. It also supports industrial data switching and positioning functions, which can improve the efficiency of talent assessment and decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily apparent by reading the following detailed description with reference to the accompanying drawings, in which several embodiments of the present invention are shown by way of example and not limitation, in which: Figure 1 A flowchart of an AI-RAG-based industrial talent intelligent matching method provided in one embodiment of the present invention; Figure 2 A schematic diagram of the structure of an AI-RAG-based industrial talent intelligent matching system provided in one embodiment of the present invention; Figure 3The figure schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided solely to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. Rather, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0019] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software.

[0020] It should be noted that any number of elements in the drawings is for illustration only and not for limitation, and any naming is only for distinction and does not have any limiting meaning.

[0021] Reference below Figure 1 , Figure 1 The following is a flow chart of an AI-RAG-based industrial talent intelligent matching method provided by one embodiment of the present invention. Figure 1 As shown, an AI-RAG-based industrial talent intelligent matching method includes: S1 generates a talent profile corresponding to the job requirement information input by the user, wherein the talent profile includes a sequence of unit feature arrays, each unit feature array corresponds to a feature similarity array, each feature similarity in the feature similarity array corresponds to matching object information, the matching object information includes a matching object and industrial data information corresponding to the matching object, the industrial data information includes an industrial data identifier and location information of the matching object in the industrial data corresponding to the industrial data identifier, and each feature similarity in the feature similarity array is arranged in descending order; S2 displays the talent profile in a preset position corresponding to the job requirement information in the matching window; S3, in response to detecting a selection operation acting on any feature array in the talent profile, determining at least one unit feature array included in the feature array as a target unit feature array sequence, wherein the feature array includes at least one unit feature array; S4, for each target unit feature array in the target unit feature array sequence, determining each piece of matching object information corresponding to the target unit feature array as a matching object information set; S5: dividing each matching object information set into matching object information group sets according to each industrial data information included in each matching object information set, wherein each matching object information in each matching object information group corresponds to the same industrial data identifier; S6 determines the target matching object information group; S7 displays the industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window, wherein the industrial data corresponds to the industrial data identifier corresponding to the target matching object information group, and in the displayed industrial data, each matching object included in the target matching object information group is in a highlighted state.

[0022] It should be noted that the generation of a talent profile corresponding to the job requirement information input by the user in the embodiments of the present invention refers to the system receiving a detailed description of the requirements for a specific position input by the user and using a pre-designed algorithm or model to generate a profile that represents the talent characteristics required for that position. A talent profile is a structure consisting of a sequence of multiple unit feature arrays, each of which corresponds to a feature similarity array, which is used to measure the degree of match between that feature and the job requirement. Each feature similarity in the feature similarity array corresponds to matching object information, which includes the matching object and the corresponding industrial data information of the matching object. Industrial data information refers to data identifiers related to the industrial field and the location of the matching object within that data. The feature similarities in the feature similarity array are arranged in descending order, meaning that the system sorts features based on their similarity to more intuitively display the features that best match the job requirement. After the talent profile is generated, the system displays it in a preset position in the matching window for easy user viewing and operation.

[0023] Specifically, job requirement information refers to the specific requirements for a job entered by the user, such as the job title, job description, and required skills. A talent profile is a complex structure. The unit feature array sequence decomposes the job requirements into multiple feature units and organizes them in an array format. Each unit feature array corresponds to a feature similarity array, where each element represents the similarity between that feature and the job requirement. Matching object information refers to the candidate talent information corresponding to that feature similarity, including the matching object (i.e., the candidate talent) and its related information in the industrial data. The industrial data identifier in the industrial data information is an identifier used to uniquely identify industrial data, and location information refers to the candidate's specific location within that data, such as their role in an industrial project or their index position within a dataset. When the system displays the talent profile in the matching window, it positions it appropriately according to pre-set display rules, allowing the user to clearly see each component of the talent profile.

[0024] Preferably, the process of generating talent portraits can be further refined. For example, for the step of generating a talent portrait corresponding to the job requirement information input by the user, a pre-trained AI-RAG intelligent matching model can be used to implement it. The model generates a feature array sequence that is highly relevant to the job requirement by inputting the job requirement information and combining it with the knowledge base and data set in the industrial field. When generating the feature similarity array, the model calculates the similarity between each feature and the job requirement and arranges them in descending order so as to prioritize the features that best match the job requirement. For the generation of matching object information, the system retrieves matching candidate talents from the industrial data based on the similarity of each feature in the feature similarity array, and obtains its location information in the industrial data, thereby constructing complete matching object information. When displaying talent portraits, the system can set different display parameters, such as font size, color coding, etc., so that users can more intuitively identify the importance of different features and matching objects.

[0025] In some embodiments, the talent profile includes text information, each unit feature array included in the feature array is a feature segment; and generating a talent profile corresponding to the job requirement information input by the user includes: Generate a reference segment information set based on the job requirement information, wherein each reference segment information in the reference segment information set corresponds to an industrial data identifier and segment location information, and the reference segment information set includes reference segment information corresponding to a text type; Generating a talent profile corresponding to the job requirement information based on the job requirement information and the reference fragment information set; For each feature segment included in the talent profile, perform the following steps: In response to determining that the feature segment is not the first segment in the talent profile, generating an initial feature similarity array based on the job requirement information, the reference segment information set, and a previous feature segment of the feature segment in the talent profile, wherein matching objects corresponding to feature similarities in the initial feature similarity array include matching words; Sorting each feature similarity in the initial feature similarity array in descending order to obtain a descending feature similarity array; Determine the preset initial value as the cumulative feature similarity; Determining the first feature similarity in the descending feature similarity array as the target feature similarity; Based on the cumulative feature similarity and the target feature similarity, the following loop steps are performed: Determining the sum of the cumulative feature similarity and the target feature similarity as the cumulative feature similarity, so as to update the cumulative feature similarity; In response to determining that the updated cumulative feature similarity is greater than a preset similarity threshold, concatenating the feature similarities preceding the target feature similarity in the descending feature similarity array and the target feature similarity into a screening feature similarity array; In response to determining that the updated cumulative feature similarity is less than or equal to the preset similarity threshold, determining the next feature similarity of the target feature similarity in the descending feature similarity array as the target feature similarity to update the target feature similarity, and executing the looping step again based on the updated cumulative feature similarity and the updated target feature similarity; According to the spliced screening feature similarity array, a feature similarity array corresponding to the feature segment and each matching object information is generated.

[0026] It should be noted that the present invention pays special attention to the semantic matching between job requirement information and talent characteristics when generating talent portraits. To this end, the system will generate a reference fragment information set based on the job requirement information, and each reference fragment information in these reference fragment information sets corresponds to an industrial data identifier and fragment location information. The reference fragment information set here refers to a representative set of fragments extracted from the job requirement text, and these fragments can reflect the key requirements of the position. The industrial data identifier and fragment location information are used to locate the specific positions of these fragments in the industrial data, so that talent data matching these fragments can be accurately found later. In the process of generating talent portraits, the system will combine the job requirement information with the reference fragment information set to generate a talent portrait containing feature fragments. Feature fragments refer to the parts of the talent portrait that can reflect the key features of the job requirements. By processing these feature fragments, the system can more accurately match talents that meet the job requirements.

[0027] Specifically, job requirement information refers to the specific requirements for a job entered by the user, typically presented in text form. The reference segment information set is a collection of key segments extracted from the job requirement information. These segments reflect the core requirements of the job. Each reference segment contains an industrial data identifier and segment location information. The industrial data identifier uniquely identifies the relevant information within the industrial data, while the segment location indicates the specific location of the segment within the industrial data. When generating a talent profile, the system compares and analyzes the job requirement information with the reference segment information set to generate a talent profile containing feature segments. Feature segments are the basic units of a talent profile. Each feature segment corresponds to a feature similarity array, which measures the degree of match between the feature and the job requirement. When generating the feature similarity array, the system generates an initial feature similarity array based on the job requirement information, the reference segment information set, and the previous feature segment of the feature segment in the talent profile. The matching objects corresponding to the feature similarities in the initial feature similarity array include matching terms. These matching terms are generated by the system based on the job requirement and the reference segment information to initially match the talent's key skills or experience. Then, the system will sort the feature similarities in the initial feature similarity array in descending order to obtain a descending feature similarity array, and through a series of accumulation and filtering operations, it will finally generate a filtered feature similarity array, thereby determining the feature similarity array and matching object information of each feature fragment.

[0028] More specifically, the process of generating a feature similarity array can be achieved through the following steps: First, the system generates a set of reference segment information based on the job requirement information. This step can be accomplished using natural language processing techniques, such as using a text segmentation algorithm to segment the job requirement text into multiple segments and select key segments based on semantic relevance. Each reference segment is associated with an industrial data identifier and segment location information, which can be obtained from a pre-established industrial data index. When generating the initial feature similarity array, the system combines the job requirement information, the reference segment information set, and the previous feature segment of the feature segment in the talent profile. Using a pre-trained model (such as the AI-RAG model), the system calculates the similarity between each feature segment and the job requirement. The AI-RAG model is an AI-based retrieval-enhanced generative model that combines retrieved relevant information to generate high-quality feature similarity arrays. The model's input parameters include the job requirement information, the reference segment information set, and the contextual information of the feature segment (i.e., the previous feature segment). Using its internal encoder and decoder architecture and attention mechanism, the model calculates the initial similarity between each feature segment and the job requirement and generates the initial feature similarity array. The system then sorts the initial feature similarity array in descending order and, by accumulating feature similarities, selects those that meet the preset similarity threshold. During this process, the system continuously updates the accumulated feature similarities and determines the final filtered feature similarity array based on the preset similarity threshold. Finally, based on the filtered feature similarity array, the system generates a feature similarity array and matching object information for each feature segment, thus completing the creation of the talent profile.

[0029] In some embodiments, generating a feature similarity array corresponding to the feature segment and information about each matching object based on the spliced screening feature similarity array includes: Determining the number of each feature similarity included in the screening feature similarity array as the similarity number; In response to determining that the number of similarities is greater than a preset number, extracting the preset number of feature similarities from the screening feature similarity array as a feature similarity array corresponding to the feature segment according to a preset extraction method; For each feature similarity in the feature similarity array, perform the following steps: Determining matching objects corresponding to the feature similarity; determining the reference segment information to which the determined matching object belongs as target reference segment information; Determine the industrial data identifier of the industrial data corresponding to the target reference segment information; Determining location information of the determined matching object in the industrial data corresponding to the industrial data identifier; Determining the determined industrial data identifier and location information as industrial data information; The determined matching object and industrial data information are determined as matching object information corresponding to the feature similarity.

[0030] It should be noted that, in the process of generating feature similarity arrays and matching object information, the present invention pays special attention to the screening of feature similarity arrays and the detailed construction of matching object information. Specifically, the system will generate feature similarity arrays and each matching object information for the corresponding feature segments based on the screened feature similarity arrays. The screened feature similarity array here refers to a feature similarity set that is highly relevant to job requirements and is obtained after a series of screening operations, while the matching object information includes the matching object and its specific location information in the industrial data. In this way, the system can more accurately generate a corresponding feature similarity array for each feature segment, and find the corresponding matching object for each similarity, thereby providing a more accurate basis for subsequent talent matching.

[0031] Specifically, the feature similarity array refers to a collection of similarity values generated by the system based on the degree of match between the feature segments and the job requirements. The similarity count refers to the total number of similarity values contained in the array. The system determines whether the filtered feature similarity array requires further processing based on a preset similarity count threshold. If the similarity count exceeds the preset number, the system extracts a preset number of feature similarity values from the filtered feature similarity array according to a preset extraction method to generate the final feature similarity array. For each feature similarity value, the system performs a series of operations to determine the matching object information. First, the matching object corresponding to the feature similarity is determined, and then the reference segment information to which the matching object belongs is determined as the target reference segment information. Next, the system determines the industrial data identifier corresponding to the target reference segment information, as well as the location information of the matching object in the industrial data corresponding to the industrial data identifier. This information together constitutes the complete matching object information, which is used for subsequent talent matching and display.

[0032] More specifically, the process of generating a feature similarity array and matching object information can be achieved through the following steps: First, the system determines whether the number of similarities in the filtered feature similarity array exceeds a preset similarity threshold. If so, the system uses a preset extraction method, such as extracting a preset number of feature similarity values in descending order of similarity. For each extracted feature similarity value, the system uses a matching algorithm to determine the corresponding matching object. The matching object refers to the talent or skill information corresponding to the feature similarity value. The system retrieves target reference segment information related to the matching object from the industrial data. The target reference segment information is the original segment information containing the matching object. The system extracts the industrial data identifier and location information from the target reference segment information. The industrial data identifier uniquely identifies the relevant information in the industrial data, while the location information indicates the specific location of the matching object within the industrial data. Finally, the system combines the matching object, its corresponding industrial data identifier, and location information into matching object information and stores this information along with the feature similarity array for subsequent display and manipulation in the data display window.

[0033] In some embodiments, the talent profile is obtained by inputting the job requirement information into a pre-trained AI-RAG intelligent matching model, the AI-RAG intelligent matching model including various decoding levels, each decoding level including various AI-RAG attention modules, each AI-RAG attention module being used to generate an attention value array for a corresponding feature segment based on the job requirement information, a reference segment information set, and a feature segment; and generating an initial feature similarity array based on the job requirement information, the reference segment information set, and the previous feature segment of the feature segment in the talent profile, including: Determining decoding levels corresponding to a preset number of levels among the decoding levels as key decoding levels; Determine the AI-RAG attention modules included in the key decoding level as a core attention module set; An initial feature similarity array is generated based on each attention value array corresponding to the feature fragment generated by each core attention module in the core attention module set.

[0034] It should be noted that the talent image mentioned in the present invention is obtained after inputting the job requirement information into the pre-trained AI-RAG intelligent matching model, which means that the system uses an artificial intelligence-based retrieval-augmented generation (RAG) model to generate talent portraits. The AI-RAG model is a deep learning model that combines retrieval and generation capabilities. It can retrieve relevant industrial data based on the input job requirement information and generate talent portraits that are highly matched with the job requirements. The model implements complex talent feature extraction and matching through multiple decoding levels, and each decoding level contains multiple AI-RAG attention modules. These attention modules can generate corresponding attention value arrays based on job requirement information, reference fragment information sets, and feature fragments, and then generate initial feature similarity arrays. In this way, the model can effectively capture the complex semantic relationship between job requirements and talent characteristics, thereby achieving accurate matching.

[0035] Specifically, the AI-RAG intelligent matching model is a deep learning model whose core approach is to combine retrieval and generation capabilities to handle complex text matching tasks. The model consists of multiple decoding layers, each composed of multiple AI-RAG attention modules. A decoding layer is a hierarchical structure within the model that progressively generates outputs. Each layer processes a portion of information and passes it on to the next layer. The AI-RAG attention module is the basic unit of the model, computing an attention value array for input information (such as job requirements, reference fragments, and feature fragments). The attention value array is a set of numerical values that represents the importance weights of various components of the input information. These weights enable the model to focus on the components most relevant to the job requirements. When generating the initial feature similarity array, the system selects a key decoding layer from the multiple decoding layers, which is considered to have the greatest impact on the final output. The core attention module set within this key decoding layer generates attention value arrays associated with the feature fragments, which are further processed to generate the initial feature similarity array. Each similarity value in the initial feature similarity array corresponds to a matching object. These matching objects include keywords or phrases related to job requirements and are used for subsequent feature similarity calculations and talent matching.

[0036] More specifically, the construction process of the AI-RAG intelligent matching model can be divided into the following steps: First, a large amount of job requirement text and corresponding talent data is collected as training data. After preprocessing, this data is input into the model for training. The model's input parameters include the job requirement information, a set of reference fragment information, and contextual information about the feature fragments. During training, the model adjusts its internal parameters by optimizing an objective function (such as minimizing prediction error) to ensure that the generated talent profile better matches the job requirement. When generating the initial feature similarity array, the system constructs the model architecture based on a preset number of decoding layers and attention modules. For example, assume the model contains five decoding layers, each containing eight AI-RAG attention modules. The system experimentally determines which decoding layers and attention modules contribute most to the generated results and designates these as the key decoding layers and core attention module sets. During actual runtime, the system inputs the job requirement information into the model. The model uses a retrieval module to retrieve fragment information related to the job requirement from industrial data. The attention modules in the decoding layers then generate an array of attention values. These attention value arrays are weighted and summed to generate an initial feature similarity array. Ultimately, the system sorts and filters matching objects based on the similarity values in the initial feature similarity array to generate accurate talent profiles.

[0037] In some embodiments, determining the target matching object information group includes: determining the first matching object information group in the matching object information group set as a target matching object information group; In response to detecting the industrial data switching operation, according to the target matching object information group, determining the matching object information group corresponding to the industrial data switching operation in the matching object information group set as the target matching object information group, so as to update the target matching object information group; The industrial data displayed in the data display window is switched to the industrial data corresponding to the updated target matching object information group, wherein in the switched industrial data, each matching object included in the updated target matching object information group is in a highlighted state.

[0038] It should be noted that determining the target matching object information group mentioned in this invention refers to selecting one or more matching object information groups from the divided set of matching object information groups as the target matching object information group for subsequent display of the corresponding industrial data. This process is a key step in the talent matching system, used to filter and organize the matched talent information according to specific rules, thereby providing users with talent data that best meets job requirements. In actual operation, the system will dynamically update the target matching object information group based on user needs and operations to meet user query requirements in different scenarios.

[0039] Specifically, a matching object information group set refers to a collection of multiple information groups obtained by dividing all matching object information according to industrial data identifiers. The matching object information in each matching object information group corresponds to the same industrial data identifier, meaning that these matching object information have the same source or background in the industrial data. A target matching object information group is one or more information groups selected from the matching object information group set to display the corresponding industrial data. When determining the target matching object information group, the system first uses the first matching object information group in the matching object information group set as the initial target matching object information group. Subsequently, the system dynamically updates the target matching object information group based on user operations, such as industrial data switching. An industrial data switching operation refers to a user requesting to view matching object information groups corresponding to different industrial data identifiers through interface controls or other interactive methods. Upon detecting such an operation, the system finds the corresponding matching object information group from the matching object information group set based on the user-specified industrial data identifier and updates it as the target matching object information group.

[0040] More specifically, the process of determining the target matching object information group can be achieved through the following steps: First, after completing the division of the matching object information set, the system will use the first matching object information group in the matching object information group set as the initial target matching object information group. This operation ensures that the matching object information related to the first industrial data identifier can be immediately displayed when the system starts. Subsequently, the system monitors user operations, such as when a user clicks on the industrial data switch control. When a switch operation is detected, the system analyzes the user's operation intention and determines the industrial data identifier that the user wishes to view. The system then searches the matching object information group set for the industrial data identifier and sets it as the new target matching object information group. To achieve this function, the system needs to pre-define the mapping relationship between industrial data identifiers and matching object information groups, and quickly query and update the target matching object information group when the user operates. In addition, the system can also provide automatic switching or recommendation functions based on the user's operation history or preferences to further enhance the user experience.

[0041] In some embodiments, displaying the industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window includes: Determining whether a matching object included in the target matching object information group exists in the current data display page displayed in the data display window; In response to determining that a matching object included in the target matching object information group exists in the current data display page displayed in the data display window, highlighting each matching object included in the target matching object information group existing in the current data display page, wherein a matching object switching control is also displayed in the current data display page; In response to detecting a selection operation corresponding to the matching object switching control, determining a switching matching object according to the target matching object information group and the currently selected matching object; Position the current data display page to the page position corresponding to the switching matching object.

[0042] It should be noted that the display of industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window mentioned in the present invention refers to displaying industrial data related to the target matching object information group in a specific area of the user interface (i.e., the data display window). This process involves not only displaying the data but also highlighting the matching objects, allowing users to quickly identify and focus on the talent information most relevant to the job requirements. In this way, the system can provide users with intuitive and targeted data display, improving user experience and operational efficiency.

[0043] Specifically, the data display window refers to a specific area in the user interface used to display industrial data. It is associated with the matching window and is used to present talent data that matches job requirements. The target matching object information group refers to a collection of matching object information that has been screened and determined. The matching objects in this information group all correspond to the same industrial data identifier. When displaying the industrial data corresponding to the target matching object information group, the system first checks whether the matching objects in the target matching object information group exist on the current data display page. If so, the system highlights these matching objects so that users can quickly identify them. At the same time, the data display page also displays a matching object switch control, which allows users to select different matching objects for viewing. When the user clicks the matching object switch control, the system determines the next matching object to switch to based on the target matching object information group and the currently selected matching object, and positions the data display page to the page position corresponding to that matching object.

[0044] More specifically, to implement the display and switching of industrial data corresponding to the target matching object information group in the data display window, the system can take the following steps: First, the system loads the industrial data corresponding to the target matching object information group into the data display window. During the loading process, the system checks whether there are matching objects in the target matching object information group on the current page. If so, the system highlights these matching objects through highlighting or other visual effects. For example, highlighting can be achieved by changing the font color or background color or adding a border. Simultaneously, the system embeds a matching object switching control, such as a button or drop-down menu, in the data display page. When the user clicks the switching control, the system determines the next matching object to be displayed based on the order in the target matching object information group. The system then obtains the location of the matching object in the industrial data and scrolls or jumps the data display page to that location. To enhance the user experience, the system can also add animation effects during the switching process to ensure smoother page scrolling or jumping. Furthermore, the system can automatically remember the user's last viewed location based on their operating habits or preferences, allowing them to navigate directly to that location the next time the page is loaded.

[0045] The aforementioned embodiments of the present invention have the following beneficial effects: This method can intelligently match industrial talent with job requirements based on AI-RAG technology. By generating a talent profile containing a feature array sequence and combining feature similarity calculation with industrial data association analysis, matching accuracy is improved. This method utilizes a descending feature similarity array and a cumulative similarity threshold screening mechanism to ensure the quality of matching results. Furthermore, by mapping industrial data identifiers with location information, a deep correlation between talent characteristics and industrial data can be established, providing a reliable basis for decision-making.

[0046] like Figure 2 As shown, some embodiments provide an AI-RAG-based industrial talent intelligent matching system, the system comprising: The generating unit 201 is configured to generate a talent profile corresponding to the job requirement information input by the user, wherein the talent profile includes a sequence of unit feature arrays, each unit feature array corresponds to a feature similarity array, each feature similarity in the feature similarity array corresponds to matching object information, the matching object information includes a matching object and industrial data information corresponding to the matching object, the industrial data information includes an industrial data identifier and location information of the matching object in the industrial data corresponding to the industrial data identifier, and the feature similarities in the feature similarity array are arranged in descending order; The first display unit 202 is configured to display the talent portrait at a preset position corresponding to the job requirement information in the matching window; A first determining unit 203 is configured to, in response to detecting a selection operation acting on any feature array in the talent profile, determine at least one unit feature array included in the feature array as a target unit feature array sequence, wherein the feature array includes at least one unit feature array; The second determining unit 204 is configured to determine, for each target unit feature array in the target unit feature array sequence, each piece of matching object information corresponding to the target unit feature array as a matching object information set; The division unit 205 is configured to divide each matching object information set into matching object information group sets according to each industrial data information included in each matching object information set, wherein each matching object information in each matching object information group corresponds to the same industrial data identifier; The third determining unit 206 is configured to determine a target matching object information group; The second display unit 207 is configured to display the industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window, wherein the industrial data corresponds to the industrial data identifier corresponding to the target matching object information group, and in the displayed industrial data, each matching object included in the target matching object information group is in a highlighted state.

[0047] It is understandable that the modules and references recorded in the AI-RAG-based industrial talent intelligent matching system Figure 1 Therefore, the operations, features, and beneficial effects described above for the AI-RAG-based industrial talent intelligent matching method are also applicable to the AI-RAG-based industrial talent intelligent matching system and the modules contained therein, and will not be repeated here.

[0048] Reference below Figure 3 , which shows a schematic structural diagram of an electronic device 300 suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 3 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0049] like Figure 3As shown, electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 302 or programs loaded from a storage device 308 into a random access memory (RAM) 303. RAM 303 also stores various programs and data required for the operation of electronic device 300. Processing device 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to bus 304.

[0050] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as needed.

[0051] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or a terminal device such as a computer, server, mobile phone, or tablet.

[0052] Some embodiments of the present disclosure also provide a computer program product, including a computer program, which implements any of the above-mentioned AI-RAG-based industrial talent intelligent matching methods when executed by a processor.

[0053] The above descriptions merely illustrate some preferred embodiments of the present invention and the underlying technical principles. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. An AI-RAG-based industrial talent intelligent matching method, characterized by: include: Generate a corresponding talent profile based on the job requirement information input by the user. The talent profile includes a sequence of unit feature arrays, each unit feature array corresponds to a feature similarity array, each feature similarity in the feature similarity array corresponds to matching object information, the matching object information includes the matching object and corresponding industrial data information, and the industrial data information includes an industrial data identifier and the location information of the matching object in the industrial data corresponding to the industrial data identifier; Displaying the talent profile in a preset position corresponding to the job requirement information in the matching window; In response to detecting a selection operation acting on any feature array in the talent profile, determining at least one unit feature array included in the feature array as a target unit feature array sequence, the feature array including at least one unit feature array; For each target unit feature array in the target unit feature array sequence, determining each piece of matching object information corresponding to the target unit feature array as a matching object information set; According to each industrial data information included in each obtained matching object information set, the matching object information set is divided into matching object information group sets, each matching object information in each matching object information group corresponds to the same industrial data identifier; Determining a target matching object information group; The industrial data corresponding to the target matching object information group is displayed in the data display window corresponding to the matching window. The industrial data corresponds to the industrial data identifier corresponding to the target matching object information group. In the displayed industrial data, each matching object included in the target matching object information group is in a highlighted state.

2. The method according to claim 1, characterized in that The talent portrait includes text information, and each unit feature array included in the feature array is a feature segment; And generating a corresponding talent profile based on the job requirement information input by the user, including: Generate a reference segment information set based on the job requirement information, wherein each reference segment information in the reference segment information set corresponds to an industrial data identifier and segment location information, and the reference segment information set includes reference segment information corresponding to a text type; Generating a talent profile corresponding to the job requirement information based on the job requirement information and the reference fragment information set; For each feature segment included in the talent profile, perform the following steps: In response to determining that the feature segment is not the first segment in the talent profile, generating an initial feature similarity array based on the job requirement information, the reference segment information set, and a previous feature segment of the feature segment in the talent profile, wherein matching objects corresponding to feature similarities in the initial feature similarity array include matching words; Sorting each feature similarity in the initial feature similarity array in descending order to obtain a descending feature similarity array; Determine the preset initial value as the cumulative feature similarity; Determining the first feature similarity in the descending feature similarity array as the target feature similarity; Based on the cumulative feature similarity and the target feature similarity, the following loop steps are performed: Determining the sum of the cumulative feature similarity and the target feature similarity as the cumulative feature similarity, so as to update the cumulative feature similarity; In response to determining that the updated cumulative feature similarity is greater than a preset similarity threshold, concatenating the feature similarities preceding the target feature similarity in the descending feature similarity array and the target feature similarity into a screening feature similarity array; In response to determining that the updated cumulative feature similarity is less than or equal to the preset similarity threshold, determining the next feature similarity of the target feature similarity in the descending feature similarity array as the target feature similarity to update the target feature similarity, and executing the looping step again based on the updated cumulative feature similarity and the updated target feature similarity; According to the spliced screening feature similarity array, a feature similarity array corresponding to the feature segment and each matching object information is generated.

3. The method according to claim 2, characterized in that The step of generating a feature similarity array corresponding to the feature segment and information about each matching object based on the spliced screening feature similarity array includes: Determining the number of each feature similarity included in the screening feature similarity array as the similarity number; In response to determining that the number of similarities is greater than a preset number, extracting the preset number of feature similarities from the screening feature similarity array as a feature similarity array corresponding to the feature segment according to a preset extraction method; For each feature similarity in the feature similarity array, perform the following steps: Determining matching objects corresponding to the feature similarity; determining the reference segment information to which the determined matching object belongs as target reference segment information; Determine the industrial data identifier of the industrial data corresponding to the target reference segment information; Determining location information of the determined matching object in the industrial data corresponding to the industrial data identifier; Determining the determined industrial data identifier and location information as industrial data information; The determined matching object and industrial data information are determined as matching object information corresponding to the feature similarity.

4. The method according to claim 2, characterized in that The talent profile is obtained by inputting the job requirement information into a pre-trained AI-RAG intelligent matching model. The AI-RAG intelligent matching model includes various decoding levels, each decoding level includes various AI-RAG attention modules, and each AI-RAG attention module is used to generate an attention value array for a corresponding feature segment based on the job requirement information, a reference segment information set, and a feature segment; and the initial feature similarity array is generated based on the job requirement information, the reference segment information set, and the previous feature segment of the feature segment in the talent profile, including: Determining decoding levels corresponding to a preset number of levels among the decoding levels as key decoding levels; Determine the AI-RAG attention modules included in the key decoding level as a core attention module set; An initial feature similarity array is generated based on each attention value array corresponding to the feature fragment generated by each core attention module in the core attention module set.

5. The method according to claim 1, characterized in that The step of determining the target matching object information group includes: determining the first matching object information group in the matching object information group set as a target matching object information group; In response to detecting the industrial data switching operation, according to the target matching object information group, determining the matching object information group corresponding to the industrial data switching operation in the matching object information group set as the target matching object information group, so as to update the target matching object information group; The industrial data displayed in the data display window is switched to the industrial data corresponding to the updated target matching object information group, wherein in the switched industrial data, each matching object included in the updated target matching object information group is in a highlighted state.

6. The method according to claim 1, characterized in that The displaying of the industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window includes: Determining whether a matching object included in the target matching object information group exists in the current data display page displayed in the data display window; In response to determining that a matching object included in the target matching object information group exists in the current data display page displayed in the data display window, highlighting each matching object included in the target matching object information group existing in the current data display page, wherein a matching object switching control is also displayed in the current data display page; In response to detecting a selection operation corresponding to the matching object switching control, determining a switching matching object according to the target matching object information group and the currently selected matching object; Position the current data display page to the page position corresponding to the switching matching object.

7. An AI-RAG-based industrial talent intelligent matching system, characterized by: include: a generating unit, which generates a corresponding talent profile based on the job requirement information input by the user, wherein the talent profile includes a sequence of unit feature arrays, each unit feature array corresponds to a feature similarity array, each feature similarity in the feature similarity array corresponds to matching object information, the matching object information includes a matching object and corresponding industrial data information, and the industrial data information includes an industrial data identifier and location information of the matching object in the industrial data corresponding to the industrial data identifier; a first display unit, displaying the talent portrait at a preset position corresponding to the job requirement information in the matching window; a first determining unit, in response to detecting a selection operation acting on any feature array in the talent profile, determining at least one unit feature array included in the feature array as a target unit feature array sequence; a second determining unit, for each target unit feature array in the target unit feature array sequence, determining each piece of matching object information corresponding to the target unit feature array as a matching object information set; a dividing unit, which divides each matching object information set into matching object information group sets according to each industrial data information included in each matching object information set, wherein each matching object information in each matching object information group corresponds to the same industrial data identifier; A third determining unit determines a target matching object information group; The second display unit displays the industrial data corresponding to the target matching object information group in the data display window corresponding to the matching window. The industrial data corresponds to the industrial data identifier corresponding to the target matching object information group. In the displayed industrial data, each matching object included in the target matching object information group is in a highlighted state.

8. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

9. A computer-readable medium having a computer program stored thereon, wherein: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images