Information acquisition method and device, computer equipment and readable storage medium

Abstract

The invention relates to an information acquisition method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining talent information of candidate talents matched with an information retrieval request in an open domain in response to the information retrieval request; acquiring contact information of at least one associated talent corresponding to each candidate from a preset information base according to the talent information and the social relation information of each candidate; in response to selection of a target talent in the candidate talents, determining an association weight between the target talent and each corresponding associated talent according to an association relationship between the target talent and at least one corresponding associated talent; and sorting the contact information of the at least one associated talent according to the association weight, and determining the contact information of the target talent based on the sorted contact information. By adopting the method, the information acquisition efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular to an information acquisition method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] The sources of talent information mainly include candidate resumes collected by recruiters through websites, resumes submitted by candidates voluntarily through various channels, and resumes recommended by headhunting firms. However, information on high-end talent (such as senior managers and technical experts) has low visibility on recruitment websites, resulting in a lack of high-end talent information on recruitment platforms.

[0003] Traditionally, recruiters use a web-wide search approach, proactively mining information on high-end talent from news reports, technical papers, and author information. However, this data is typically massive and structurally diverse, making the collection and processing time-consuming and labor-intensive, resulting in low information acquisition efficiency. Summary of the Invention

[0004] Therefore, it is necessary to provide an information acquisition method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the efficiency of information acquisition in order to address the above-mentioned technical problems.

[0005] Firstly, this application provides an information acquisition method, including:

[0006] In response to an information retrieval request, retrieve talent information of candidates that match the information retrieval request in an open domain;

[0007] Based on the talent information and social relationship information of each candidate, retrieve the contact information of at least one related talent corresponding to each candidate from the preset information database;

[0008] In response to the selection of target talents among the candidates, the association weight between the target talent and each of the corresponding associated talents is determined based on the association relationship between the target talent and at least one associated talent.

[0009] The contact information of at least one related talent is sorted according to the association weight, and the contact information of the target talent is determined based on the sorted contact information.

[0010] The aforementioned information acquisition method, in response to an information retrieval request, obtains talent information of candidates matching the request from an open domain. Based on the talent information and social relationship information of each candidate, it retrieves the contact information of associated talents corresponding to each candidate from a pre-set information database. Since the talent information obtained from the open domain often does not include contact information, and associated talents are those who have a relationship with the candidate, when selecting a target talent from among the candidates, the contact information of the target talent can be determined through the contact information of at least one associated talent. This method of obtaining the target talent's contact information through the contact information of associated talents is beneficial to improving information acquisition efficiency. Since the probability of obtaining the target talent's contact information through associated talents with different relationships varies, different association weights are determined based on the association relationships. This allows the contact information of at least one associated talent to be sorted according to the association weights, and the contact information of the target talent is determined according to the sorted contact information, which is beneficial to further improving information acquisition efficiency.

[0011] In one embodiment, obtaining talent information of candidates matching the information retrieval request in the open domain includes:

[0012] Retrieve the search keywords from the information retrieval request;

[0013] Retrieve a summary page containing at least one search keyword in an open domain;

[0014] Talent information for candidates is determined based on text on the summary page that contains at least one search keyword.

[0015] In this embodiment, by using the search keywords in the information retrieval request, a summary page containing the search keywords is obtained, and then the talent information of the candidates is extracted from the summary page. This method of obtaining information by searching keywords is beneficial for quickly retrieving talent information of candidates that meet the user's search needs from the open domain.

[0016] In one embodiment, retrieving a summary page containing at least one search keyword in an open domain includes:

[0017] For any search keyword, at least one candidate keyword with a similar relationship to the current search keyword is obtained through a preset entity similarity model;

[0018] Retrieve a summary page containing at least one candidate keyword from the open domain;

[0019] Talent information for candidates is determined based on text on the summary page that contains at least one search keyword, including:

[0020] Talent information for candidates is determined based on text on the summary page that contains at least one candidate keyword.

[0021] In this embodiment, at least one candidate keyword with a similar relationship to the current search keyword is obtained through a preset entity similarity model. The recall summary page can include not only web pages containing the search keyword, but also pages containing candidate keywords, which improves the recall rate of the summary page. In addition, the talent information of the candidate is determined based on the candidate keywords in the summary page, which helps to improve the recall rate of talent information.

[0022] In one embodiment, based on the talent information and social relationship information of each candidate, contact information of at least one associated talent corresponding to each candidate is obtained from a preset information database, including:

[0023] For each candidate, target information matching at least one piece of information from the preset information database, either social relationship information or talent information, is retrieved.

[0024] The talent corresponding to the target information is regarded as the associated talent with the current candidate talent, and the contact information of the associated talent is obtained from the target information.

[0025] In this embodiment, target information that matches at least one of the social relationship information or talent information of the current candidate is obtained from a preset information database. The talent corresponding to the target information is regarded as the associated talent corresponding to the current candidate. Since the target information can match not only talent information but also social relationship information, it is beneficial to improve the recall rate of information. In addition, obtaining the contact information of associated talents from the target information is beneficial to improving the accuracy and efficiency of talent information acquisition.

[0026] In one embodiment, the association weight between the target talent and each corresponding associated talent is determined based on the association relationship between the target talent and at least one associated talent, including:

[0027] Obtain the mapping relationship between association relationships and association weights;

[0028] Find the mapping relationship and use the association weight corresponding to the relationship as the association weight between the target talent and the corresponding related talent.

[0029] In this embodiment, the association weights corresponding to the association relationships are obtained by querying the preset stored mapping relationships. The mapping relationships can be flexibly set according to the actual situation to ensure that the mapping relationships can determine appropriate association weights, thereby improving the efficiency of information acquisition.

[0030] In one embodiment, obtaining talent information of candidates matching the information retrieval request in the open domain includes:

[0031] Determine whether there is talent information in the pre-constructed talent map that matches the information retrieval request;

[0032] If not found, retrieve the talent information of candidates matching the information retrieval request from the open domain.

[0033] In this embodiment, by first searching the talent map and then further searching the open domain if no talent information matching the information retrieval request is found in the talent map, the efficiency of information acquisition is improved.

[0034] In one embodiment, the information acquisition method further includes:

[0035] For each candidate, obtain the individual information types of each candidate's talent information;

[0036] If the information type does not include all the preset information types, the talent information of the current candidate is supplemented according to the relationship between the candidate and at least one corresponding related talent.

[0037] The talent map is updated based on the talent information supplemented by each candidate.

[0038] In this embodiment, by supplementing the talent information according to the information type to which each talent information belongs, and using the supplemented talent information to update the talent map, it is beneficial to improve the completeness of talent information in the talent map, thereby improving the efficiency of information acquisition.

[0039] In one embodiment, the information acquisition method further includes:

[0040] The talent map is updated based on the talent information of each candidate, the talent information of each related talent, contact information, relationship, and relationship weight.

[0041] In this embodiment, the talent graph includes talent information of candidates recalled from the open domain, as well as talent information, contact information, and contact information of related talents, enriching the information content of the talent graph and improving the efficiency of subsequent information retrieval from the talent graph. Furthermore, because the talent graph includes the relationships between talents and the corresponding association weights, it enhances the connections between talents. Information on both candidates and related talents can be provided to users as retrieved talent information, improving the information recall rate. Users can understand the relationships between talents, thus helping them select suitable target talents.

[0042] In one embodiment, the information acquisition method further includes:

[0043] The talent information of each candidate is corrected to obtain the corrected talent information; the correction includes at least one of entity alignment, information repair and data verification.

[0044] The talent map is updated based on the revised talent information of each candidate.

[0045] In this embodiment, the talent information extracted in real time is corrected, often through an asynchronous offline method, to ensure the accuracy of the talent information entering the talent map. This combination of real-time and offline methods allows for more detailed analysis and processing. Real-time extraction focuses on quickly retrieving and integrating the most relevant information, which is beneficial for ensuring information acquisition efficiency. To avoid missing important information from actual web pages, offline tasks are used to ensure that the talent information entering the talent map remains up-to-date and comprehensive.

[0046] Secondly, this application also provides an information acquisition device, comprising:

[0047] The first acquisition module is used to respond to information retrieval requests and acquire talent information of candidates that match the information retrieval requests in the open domain;

[0048] The second acquisition module is used to acquire the contact information of at least one related talent corresponding to each candidate from a preset information database based on the talent information and social relationship information of each candidate.

[0049] The first determining module is used to determine the association weight between the target talent and each corresponding associated talent in response to the selection of the target talent among the candidates.

[0050] The second determining module is used to sort the contact information of at least one related talent according to the association weight, and determine the contact information of the target talent based on the sorted contact information.

[0051] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0052] In response to an information retrieval request, retrieve talent information of candidates that match the information retrieval request in an open domain;

[0053] Based on the talent information and social relationship information of each candidate, retrieve the contact information of at least one related talent corresponding to each candidate from the preset information database;

[0054] In response to the selection of target talents among the candidates, the association weight between the target talent and each of the corresponding associated talents is determined based on the association relationship between the target talent and at least one associated talent.

[0055] The contact information of at least one related talent is sorted according to the association weight, and the contact information of the target talent is determined based on the sorted contact information.

[0056] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0057] In response to an information retrieval request, retrieve talent information of candidates that match the information retrieval request in an open domain;

[0058] Based on the talent information and social relationship information of each candidate, retrieve the contact information of at least one related talent corresponding to each candidate from the preset information database;

[0059] In response to the selection of target talents among the candidates, the association weight between the target talent and each of the corresponding associated talents is determined based on the association relationship between the target talent and at least one associated talent.

[0060] The contact information of at least one related talent is sorted according to the association weight, and the contact information of the target talent is determined based on the sorted contact information.

[0061] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0062] In response to an information retrieval request, retrieve talent information of candidates that match the information retrieval request in an open domain;

[0063] Based on the talent information and social relationship information of each candidate, retrieve the contact information of at least one related talent corresponding to each candidate from the preset information database;

[0064] In response to the selection of target talents among the candidates, the association weight between the target talent and each of the corresponding associated talents is determined based on the association relationship between the target talent and at least one associated talent.

[0065] The contact information of at least one related talent is sorted according to the association weight, and the contact information of the target talent is determined based on the sorted contact information. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0067] Figure 1 This is an application environment diagram of the information acquisition method in one embodiment;

[0068] Figure 2 This is a flowchart illustrating an information acquisition method in one embodiment;

[0069] Figure 3 This is a schematic diagram illustrating the correlation weights between talents in one embodiment;

[0070] Figure 4 This is a schematic diagram of the process for constructing a talent graph in one embodiment;

[0071] Figure 5 This is a flowchart illustrating the information acquisition method in another embodiment;

[0072] Figure 6 This is a structural block diagram of an information acquisition device in one embodiment;

[0073] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0075] The sources of talent information mainly include candidate resumes collected by recruiters through websites, resumes submitted by candidates voluntarily through various channels, and resumes recommended by headhunting firms. This talent information is centrally stored in a talent pool, enabling recruiters to quickly locate and obtain specific information about target candidates through information retrieval functions. However, information on high-end talent (e.g., senior managers and technical experts) has low visibility on recruitment websites, resulting in a lack of high-end talent information on recruitment platforms.

[0076] Traditional methods for acquiring high-end talent involve recruiters using a comprehensive online search approach, proactively extracting information from news reports, technical papers, and author information. However, this data is typically massive and structurally diverse, and the collection and processing of this data is time-consuming and labor-intensive, resulting in low information acquisition efficiency.

[0077] To improve information retrieval efficiency, traditional methods typically employ open-domain data collection to automatically crawl relevant information from the internet and store it in a talent database for subsequent querying and use. The storage and retrieval of talent information is generally based on MySQL. MySQL can store structured data such as name, gender, contact number, and email address, but it only supports conditional searches and does not support full-text keyword matching. Furthermore, its efficiency is low when dealing with large amounts of data.

[0078] In summary, traditional methods have the following technical problems:

[0079] 1. Low efficiency in processing massive amounts of real-time and complex data: Open domain data collection and rule processing can only handle structured data. However, open domain talent information mostly comes from long text data such as news and information, including a large amount of unstructured data. It is highly diverse, has different data structures, large data volume, and is not updated regularly. Information acquisition is time-consuming, laborious, and inefficient.

[0080] 2. Lack of a mechanism for real-time dynamic updates and data linking: Traditional methods incrementally write new information into the talent pool, failing to achieve real-time entity linking, i.e., dynamically integrating information such as names, work unit names, and job titles. This generates a large amount of unnecessary data and can lead to the loss of key information.

[0081] 3. Low retrieval recall and precision: Due to the diversity and complexity of open-domain data, traditional keyword retrieval methods struggle to accommodate data sources with different structures and cannot effectively recall semantically similar information. Furthermore, individuals closely related to the target candidate and possessing similar experience and abilities are also potential candidates for recruitment; traditional retrieval solutions also fail to capture these potential target groups, resulting in unsatisfactory retrieval results.

[0082] 4. Lack of a talent relationship discovery mechanism leads to poor data connectivity and difficulty in obtaining candidate contact information: In open domains, candidate contact information is often not publicly available and needs to be obtained indirectly, such as through colleague relationships. Traditional methods often require significant manual intervention to organize and analyze the relationship chains between candidates, which is not only inefficient but also prone to errors. The long and complex path to obtaining contact information makes it difficult to acquire, impacting recruitment efficiency and success rates.

[0083] Based on this, embodiments of this application propose an information acquisition method, apparatus, computer device, computer-readable storage medium, and computer program product. In response to an information retrieval request, it acquires talent information of candidates matching the request in an open domain. Based on the talent information and social relationship information of each candidate, it retrieves contact information of associated talents corresponding to each candidate from a preset information database. Since the talent information acquired in the open domain often does not include contact information, and associated talents are those related to the candidate, when selecting a target talent from among the candidates, the target talent's contact information can be determined through the contact information of at least one associated talent. This method of acquiring the target talent's contact information through associated talent's contact information is beneficial for improving information acquisition efficiency. Since the probability of obtaining the target talent's contact information through associated talents with different relationships varies, different association weights are determined based on the association relationships. This allows for the sorting of the contact information of at least one associated talent according to the association weights, and the determination of the target talent's contact information based on the sorted contact information, further improving information acquisition efficiency.

[0084] The information acquisition method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated on server 104 or placed on the cloud or other network servers. This embodiment can be applied to server 104, terminal 102, and systems including terminal 102 and server 104, and is implemented through the interaction between terminal 102 and server 104. In response to an information retrieval request, server 104 obtains talent information of candidates matching the information retrieval request in an open domain; based on the talent information and social relationship information of each candidate, it obtains contact information of at least one associated talent corresponding to each candidate from a preset information database; in response to the selection of a target talent among the candidates, it determines the association weight between the target talent and each corresponding associated talent based on the association relationship between the target talent and at least one associated talent; it sorts the contact information of at least one associated talent according to the association weight, and determines the contact information of the target talent based on the sorted contact information. The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle systems, and projection devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. The server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0085] In one exemplary embodiment, such as Figure 2 As shown, an information acquisition method is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 202 to 208. Wherein:

[0086] Step 202: In response to the information retrieval request, obtain the talent information of the candidates that match the information retrieval request in the open domain.

[0087] An information retrieval request refers to a request initiated by a user on a server to retrieve talent information. Users can be recruiters, etc. Information retrieval requests often contain at least one search keyword; for example, the search keywords in an information retrieval request could be: Tsinghua University, scientists, etc.

[0088] Open domains refer to the various web pages within search engines. The information in open domains is vast and diverse, encompassing both structured and unstructured information. Structured information refers to information that, after analysis, can be broken down into multiple interconnected components; for example, structured information includes names, genders, and occupations. Unstructured information refers to information with no fixed format, such as long text information in news reports. In some embodiments, information in open domains may include news reports, technical papers, and article authors.

[0089] Candidates refer to talents who match the information retrieval request. For example, if the search keywords in the information retrieval request are Tsinghua University and scientists, then candidates are talents with the attributes of Tsinghua University and scientists retrieved from the open domain.

[0090] Talent information refers to information related to a talent's profession, such as name, employer, and job title.

[0091] When the server receives an information retrieval request from a user, it responds by retrieving the talent information of candidates that match the information retrieval request from the open domain.

[0092] Step 204: Based on the talent information and social relationship information of each candidate, obtain the contact information of at least one related talent corresponding to each candidate from the preset information database.

[0093] Among them, social relationship information refers to information related to the social relationships of talents, including the schools they have attended, the employers they have worked for, and the social activities or forums they have attended.

[0094] Contact information refers to information used to contact the relevant talent, such as phone number, email address, home address, etc.

[0095] The default information database can be a database on a server, or a database connected to a server. The default information database stores relatively complete information about the talent, such as their name, contact information, and social relationships.

[0096] In some embodiments, the social relationship information of each candidate can be obtained from a preset information database. The server searches the preset information database for information that overlaps with the candidate's talent information and social relationship information, and identifies the talent corresponding to the found information as the candidate's associated talent. That is, associated talent refers to talent whose talent information and social relationship information overlap with the candidate's, and each candidate has at least one associated talent. For example, by using the work unit to which the talent belongs in the talent information and the position indicated by the social relationship information, talent holding the same position in the same work unit as the candidate can be obtained from the preset information database, and that talent can be identified as the candidate's associated talent.

[0097] After identifying at least one associated talent for each candidate, the server further retrieves the contact information of each associated talent from a pre-set database.

[0098] Step 206: In response to the selection of target talent among the candidates, determine the association weight between the target talent and each corresponding associated talent based on the association relationship between the target talent and at least one associated talent.

[0099] The server includes a human-computer interaction interface. After retrieving candidate information based on an information retrieval request, the server displays the candidate information on the interface, allowing users to quickly find candidates matching their search criteria. The target candidate refers to the candidate selected by the user from the available candidates. Users can further select the target candidate from the candidate information displayed on the interface.

[0100] Since the target talent is selected by the user from among various candidates, the target talent has at least one associated talent. The associated relationship refers to the relationship between the target talent and the corresponding associated talent; for example, it can be a colleague relationship, an alumni relationship, or a shared activity relationship. A colleague relationship refers to working in the same company; an alumni relationship refers to attending the same school; and a shared activity relationship refers to attending the same social event or forum.

[0101] The relationship information between target talent and corresponding related talent can be determined based on their respective talent information or social relationship information. For example, the server can determine that target talent and related talent working in the same workplace have a colleague relationship. The server can determine that target talent and related talent attending the same social event have an event-related relationship.

[0102] The degree of intimacy between the target talent and the associated talent varies depending on the type of relationship, and the likelihood of obtaining the target talent's contact information through the associated talent's contact information also differs. For example, the same associated talent corresponding to the target talent may have multiple relationships with the target talent, or a single associated talent may have the same relationship with the target talent. Therefore, the server determines the association weight between the target talent and each corresponding associated talent based on each relationship. The association weight can be used to characterize the degree of intimacy between the target talent and its corresponding associated talent, or to characterize the likelihood of obtaining the target talent's contact information through the associated talent's contact information.

[0103] For example, the association weight of colleague relationships is affected by whether they were employed at the same time and the similarity of their positions. For instance, the association weight of colleagues with overlapping employment periods within the same workplace is greater than that of colleagues with non-overlapping employment periods within the same workplace. Furthermore, the higher the similarity of positions within the same workplace, the greater the association weight of the colleague relationship.

[0104] Step 208: Sort the contact information of at least one related talent according to the association weight, and determine the contact information of the target talent based on the sorted contact information.

[0105] Since the related talent and the target talent have a connection, such as a colleague relationship, alumni relationship, or relationship based on the same activities, the server can further determine the contact information of the target talent based on the contact information of the related talent.

[0106] The server sorts the contact information of at least one related talent according to the order of their association weights, resulting in a sorted list of contact information. For example, the contact information of related talents with higher association weights is sorted first, and the contact information of related talents with lower association weights is sorted last.

[0107] Since association weights can be used to characterize the likelihood of obtaining the target talent's contact information through the contact information of associated talents, the higher the ranking of the contact information, the greater the likelihood of obtaining the target talent's contact information through that contact information.

[0108] In the aforementioned information acquisition method, in response to an information retrieval request, talent information of candidates matching the request is obtained from an open domain. Based on the talent information and social relationship information of each candidate, contact information of associated talents corresponding to each candidate is obtained from a pre-set information database. Since the talent information obtained from the open domain often does not include contact information, and associated talents are those who have a relationship with the candidate, when selecting a target talent from among the candidates, the contact information of the target talent can be determined by the contact information of at least one associated talent. This method of obtaining the contact information of the target talent through the contact information of associated talents is beneficial to improving the efficiency of information acquisition. Since the probability of obtaining the contact information of the target talent through associated talents with different relationships is different, different association weights are determined according to the association relationships. Thus, the contact information of at least one associated talent can be sorted according to the association weights, and the contact information of the target talent can be determined according to the sorted contact information, which is beneficial to further improving the efficiency of information acquisition.

[0109] In one exemplary embodiment, obtaining talent information of candidates matching an information retrieval request in an open domain includes: obtaining search keywords from the information retrieval request; obtaining a summary page containing at least one search keyword in the open domain; and determining the talent information of candidates based on the text containing at least one search keyword on the summary page.

[0110] The text contained in an information retrieval request often includes at least one search keyword. For example, an information retrieval request may be a combination of a work unit name and a job description.

[0111] Because open domains contain various web pages from search engines, a summary page can be a web page containing at least one search keyword. After retrieving the summary pages, the server further extracts the candidate's talent information from the text contained in each summary page. For example, if the search keywords in the information retrieval request include "Tsinghua University" and "scientist," the retrieved summary pages can contain only "Tsinghua University," only "scientist," or both. The server can extract "Tsinghua University" and the candidate's name from a summary page containing "Tsinghua University" and use it as the candidate's talent information. Alternatively, it can extract "scientist" and the candidate's name from a summary page containing "scientist" and use it as the candidate's talent information. Or, it can extract "Tsinghua University," "scientist," and the candidate's name from a summary page containing both "Tsinghua University" and "scientist" and use it as the candidate's talent information.

[0112] In some embodiments, for each candidate, the server may generate a tuple containing at least one search keyword.

[0113] In this embodiment, by using the search keywords in the information retrieval request, a summary page containing the search keywords is obtained, and then the talent information of the candidates is extracted from the summary page. This method of obtaining information by searching keywords is beneficial for quickly retrieving talent information of candidates that meet the user's search needs from the open domain.

[0114] In an exemplary embodiment, obtaining a summary page containing at least one search keyword in an open domain includes: for any search keyword, obtaining at least one candidate keyword that has a similar relationship with the current search keyword through a preset entity similarity model; obtaining a summary page containing at least one candidate keyword from the open domain; and determining the talent information of the candidate based on the text containing at least one search keyword in the summary page, including: determining the talent information of the candidate based on the text containing at least one candidate keyword in the summary page.

[0115] Candidate keywords refer to keywords that are similar to the current search keyword. For example, "South China University of Technology" and "SCUT" refer to the same entity, but when the search keyword is "South China University of Technology", it is impossible to recall summary pages and talent information containing "SCUT". Therefore, this application proposes to use a preset entity similarity model to obtain at least one candidate keyword that is similar to the current search keyword, thereby improving the retrieval recall rate.

[0116] The server will use word vector technology to convert the current keyword into a corresponding text vector. This text vector will then be input into a preset entity similarity model to obtain at least one output text vector. The output text vector is a text vector output by the preset entity similarity model that has a certain degree of similarity to the text vector corresponding to the current keyword. The server will convert at least one output text vector into corresponding text and use it as at least one candidate keyword.

[0117] In some embodiments, a preset entity similarity model stores the similarity relationships between multiple keyword samples. The training process of the preset entity similarity model is as follows: The server constructs a training dataset containing multiple keyword samples. The server uses word vector technology to convert each keyword sample into a corresponding text vector, and then compares the similarity of each text vector to obtain the similarity result between the text vectors. When the similarity result is greater than the preset similarity, it is determined that the corresponding keyword samples have a similar relationship; when the similarity result is not greater than the preset similarity, it is determined that the corresponding keyword samples do not have a similar relationship.

[0118] For example, "Tsinghua Tongfang" has a similar relationship with "Tongfang Co., Ltd.", "Tsinghua Tongfang Co., Ltd.", and "Tongfang Shares", while "Tsinghua Tongfang", "Tongfang Co., Ltd.", "Tsinghua Tongfang Co., Ltd.", and "Tongfang Shares" do not have a similar relationship with "Tsinghua University".

[0119] The preset entity similarity model continuously learns the similarity relationships between keyword samples during the training process, thus obtaining a preset entity similarity model that stores the similarity relationships between keyword samples.

[0120] The server can retrieve summary pages containing at least one candidate keyword from the open domain, and it can also retrieve summary pages containing at least one search keyword. In this way, the retrieved summary pages can not only contain web pages containing search keywords, but also web pages containing candidate keywords, thus improving the recall rate of summary pages.

[0121] The server extracts text containing at least one candidate keyword from the text content of each summary page, and can also extract text containing at least one search keyword. It then determines the talent information of the candidates from the extracted text. Because the recall rate of the summary pages is improved, the recall rate of the talent information is correspondingly improved.

[0122] In this embodiment, at least one candidate keyword with a similar relationship to the current search keyword is obtained through a preset entity similarity model. The recalled summary page can contain not only web pages containing the search keyword, but also pages containing candidate keywords, which improves the recall rate of the summary page. In addition, the talent information of the candidate is determined based on the candidate keywords in the summary page, which helps to improve the recall rate of talent information.

[0123] In an exemplary embodiment, based on the talent information and social relationship information of each candidate, the contact information of at least one associated talent corresponding to each candidate is obtained from a preset information database. This includes: for each candidate, obtaining target information that matches at least one piece of information in the social relationship information or talent information of the current candidate from the preset information database; designating the talent corresponding to the target information as the associated talent corresponding to the current candidate, and obtaining the contact information of the associated talent from the target information.

[0124] The pre-set information database stores the names, contact information, and social relationship information of potential candidates. The server searches the pre-set information database to retrieve target information that matches at least one piece of information from the candidate's social relationship information or talent information.

[0125] Matching target information not only with talent information, but also with social relationship information, and identifying the talent corresponding to the target information as related talent to the current candidate talent, helps to improve the talent recall rate and expand the candidate talent's relationship network.

[0126] Matching target information with talent information can be either identical or the similarity between the two meets a preset similarity condition, such as being greater than a preset similarity. Similarly, matching target information with social relationship information can also be either identical or the similarity between the two meets a preset similarity condition, such as being greater than a preset similarity. For example, if a talent's work unit is Tsinghua University and their social relationship information is attending Forum A, then talents with Tsinghua University affiliation and those who attended Forum A in the preset information database can both be considered as related talents.

[0127] The target information is related information about associated talents stored in a preset information database, including the contact information of the associated talents. Therefore, the server can obtain the contact information of associated talents from the target information.

[0128] In this embodiment, target information that matches at least one of the social relationship information or talent information of the current candidate is obtained from a preset information database. The talent corresponding to the target information is regarded as the associated talent corresponding to the current candidate. Since the target information can match not only talent information but also social relationship information, it is beneficial to improve the recall rate of information. In addition, obtaining the contact information of associated talents from the target information is beneficial to improving the accuracy and efficiency of talent information acquisition.

[0129] In an exemplary embodiment, determining the association weight between the target talent and each corresponding associated talent based on the association relationship between the target talent and at least one associated talent includes: obtaining the mapping relationship between the association relationship and the association weight; finding the mapping relationship and using the association weight corresponding to the association relationship as the association weight between the target talent and the corresponding associated talent.

[0130] The server pre-stores mapping relationships that represent the correspondence between relationships and their weights. These mapping relationships can be set according to actual needs. For example, the weight of a colleague relationship is 0.3, and the weight of a relationship based on shared activities is 0.5.

[0131] In some embodiments, the association weight of colleague relationships is affected by whether they were employed at the same time and the similarity of their positions. For example, for colleague relationships with overlapping employment periods but a job similarity lower than a preset value, the association weight of such colleague relationships can be 0.4; for colleague relationships with overlapping employment periods and a job similarity not lower than a preset value, the association weight of such colleague relationships can be 0.9.

[0132] like Figure 3 The diagram shows the association weights between talents in one embodiment. The target talent selected by the user is talent (04), and the other talents are associated talents of the target talent. Talent (6) is associated with the target talent as an alumnus, with an association weight of 0.1. Talent (5) is associated with the target talent as a fellow participant, with an association weight of 0.5. Talent (7) is associated with the target talent as a colleague, with overlapping employment periods but a job similarity lower than a preset value, with an association weight of 0.4. Talent (3) is associated with the target talent as a colleague, with overlapping employment periods and a job similarity not lower than a preset value, with an association weight of 0.9. In some embodiments, the server can sort the associated talents according to their association weights. For example, if talent (3) has the highest association weight, the server can consider talent (3) as the best associated talent of the target talent.

[0133] In this embodiment, the association weights corresponding to the association relationships are obtained by querying the preset stored mapping relationships. The mapping relationships can be flexibly set according to the actual situation to ensure that the mapping relationships can determine appropriate association weights, thereby improving the efficiency of information acquisition.

[0134] In one exemplary embodiment, obtaining talent information of candidates matching the information retrieval request in an open domain includes: determining whether there is talent information matching the information retrieval request in a pre-constructed talent map; if not, obtaining talent information of candidates matching the information retrieval request in an open domain.

[0135] The talent graph refers to a pre-constructed knowledge graph storing information about various talents. Upon receiving an information retrieval request, the server first searches the talent graph to determine if any talent information matching the request exists. If matching talent information exists, it is returned and displayed as a retrieval result. If no matching talent information exists in the talent graph, further retrieval of candidate talent information is required from the open domain.

[0136] In this embodiment, by first searching the talent graph and then further searching the open domain if no talent information matching the information retrieval request is found in the talent graph, the efficiency of information acquisition can be improved.

[0137] In an exemplary embodiment, the information acquisition method further includes: for each candidate talent, acquiring the information type to which each talent information of the current candidate talent belongs; if the information type does not include all preset information types, supplementing the talent information of the current candidate talent according to the association relationship between the candidate talent and at least one corresponding associated talent; and updating the talent map based on the supplemented talent information of each candidate talent.

[0138] Information type refers to the type of talent information, such as occupation type, employer type, employment period type, and attendance period type. The server can return tuples of talent information, which contain talent information of various information types. Due to the diverse information structures in open domains, the retrieved summary page may not contain all the information types required by the tuple. Therefore, some information types of talent information may be missing, such as the employer type or the employment period type.

[0139] To ensure the completeness of talent information, this application proposes to supplement missing parts of the talent information. Specifically, the talent information of the current candidate can be supplemented based on the relationship between the candidate and at least one related talent. For example, if the talent information is missing due to a time period, the relationship indicates that the candidate attended a social event on behalf of the employer at time B. Therefore, the time period can be supplemented to time B. Furthermore, based on the supplemented time period, colleague relationships within the same time period can be identified more accurately.

[0140] The server can update the talent map based on the talent information supplemented by each candidate, which helps improve the completeness of the talent information in the talent map. When the information type includes all preset information types, the talent map can be updated directly based on the talent information of each candidate.

[0141] The updated talent map can be used for subsequent information retrieval. As the demand for information retrieval increases, the server can continuously obtain more talent information from the open domain, which helps to enrich the amount of information in the talent map and thus improve the efficiency of information acquisition.

[0142] In this embodiment, by supplementing the talent information according to the information type to which each talent information belongs, and using the supplemented talent information to update the talent map, it is beneficial to improve the completeness of talent information in the talent map, thereby improving the efficiency of information acquisition.

[0143] In an exemplary embodiment, the information acquisition method further includes updating the talent map based on the talent information of each candidate talent, the talent information of each associated talent, contact information, association relationships, and association weights.

[0144] After obtaining the talent information, contact information, relationships between candidates and related talents, and the corresponding relationship weights of each relationship, the server can write this data into the talent graph.

[0145] In this embodiment, the talent graph includes talent information of candidates recalled from the open domain, as well as talent information, contact information, and contact information of related talents and candidates. This enriches the information content of the talent graph and improves the efficiency of subsequent information retrieval. Furthermore, the addition of relationships between talents and their corresponding weights enhances the connections between them. Information on both candidates and related talents can be provided to users as retrieved talent information, increasing the recall rate. Users can understand the relationships between talents, thus helping them select suitable target talents.

[0146] In an exemplary embodiment, the information acquisition method further includes: correcting the talent information of each candidate to obtain corrected talent information; the correction includes at least one of entity alignment, information repair and data verification; and updating the talent map based on the corrected talent information of each candidate.

[0147] Talent information obtained from the open domain comes from a summary page. To ensure the real-time nature of information acquisition, the content of the summary page is often truncated, which can lead to errors in the acquired talent information. For example, taking the formation of talent information triplets as an example, a talent information triplet might include (Tsinghua, scientist, Li Si) or (Peking University, engineer, Zhang San). Due to the truncated text content in the summary page, an incorrect triplet (Tsinghua, scientist, Zhang San) is formed. To address this situation, this application proposes to correct the talent information, thereby updating the talent graph based on the corrected talent information, which helps ensure the accuracy of the information in the talent graph.

[0148] The correction methods can include at least one of entity alignment, information repair, and data verification. Entity alignment refers to merging talent information belonging to the same entity. For example, "SCUT" and "South China University of Technology" both belong to the same entity. Merging these two types of talent information into "South China University of Technology" will result in aligned talent information.

[0149] Information repair refers to repairing talent information using content stored offline on the server. The server extracts talent information from the offline storage. If the talent information extracted in real-time is inconsistent with the offline extracted information, the offline extracted information is used to overwrite the real-time extracted information. Because the offline stored content contains complete and comprehensive information, using offline stored content to repair the real-time extracted talent information helps improve the accuracy of the talent information.

[0150] Data validation encompasses three methods: rule-based, model-based, and manual. Rule-based methods involve deduplication, null value removal, and formatting of the tuples containing talent information. Model-based methods compare talent information with pre-stored reference information; for example, comparing pre-stored employer names with those in the talent information to determine accuracy and filter out low-quality information. Low-quality information may include incorrect employer names or irrelevant job titles. Manual methods employ manual sampling and review, establishing data review processes and quality assessment standards to ensure the accuracy of talent information. Furthermore, analyzing filtered and reviewed anomalies allows for continuous optimization and adjustment of information extraction and entity alignment algorithms, improving the accuracy and relevance of the information.

[0151] In this embodiment, the talent information extracted in real time is corrected, often through an asynchronous offline method, to ensure the accuracy of the talent information entering the talent graph. This combination of real-time and offline methods allows for more detailed analysis and processing, preventing the omission of important information from actual web pages. The asynchronous execution of offline tasks ensures that the talent information entering the talent graph remains up-to-date and comprehensive.

[0152] To illustrate the information acquisition method and its effects in this solution in detail, a specific embodiment is described below:

[0153] This paper uses a scenario where recruiters search for information on high-end talent as an example. The information acquisition method proposed in this application is mainly based on a talent graph, which includes real-time construction and offline construction. The real-time construction part is mainly used to respond to the user's information retrieval request. If the talent information matching the information retrieval request exists in the pre-built talent graph, the matched talent information is returned to the user. If it does not exist, the talent information of the candidate is retrieved through open domains. The offline construction part is executed asynchronously and on a timed basis, involving information collection and mining at the hourly to daily level. After retrieving the talent information of the candidate, the talent information can be updated in the talent graph.

[0154] like Figure 4 The diagram illustrates a flowchart of a talent graph construction method in one embodiment. In the real-time construction phase, the server responds to a user's information retrieval request by extracting information from an open domain to obtain talent information for candidates matching the retrieval request. This information is then grouped into triples, and entity alignment is performed on the subgraph to obtain the real-time graph layer. In the offline construction phase, webpages are categorized to obtain valid long text within the webpages. Information is extracted from this valid long text to obtain offline triples, which are then aligned with entities. The offline triples can be used to correct the real-time triples, thereby updating the talent graph with the corrected talent information. Correction includes at least one of entity alignment, information repair, and data verification. Additionally, the server can retrieve associated talents for each candidate from a pre-defined database, perform entity alignment on the talent information of each associated talent, and write both the candidate's and associated talent's information into the resulting graph layer of the talent graph. Users can perform mixed searches on candidates and associated talents to determine the selected target talent. Taking a user selecting a target talent from the candidate pool as an example, the server determines the association between the target talent and at least one related talent. It then searches for the mapping between the association and its weight, using the weight corresponding to the association as the association weight between the target talent and its related talent. Based on the association weight, the contact information of at least one related talent is sorted, and the contact information of the target talent is determined based on this sorted information.

[0155] like Figure 5 The diagram illustrates a flowchart of an information acquisition method in one embodiment. The information extraction methods in the open domain include keyword retrieval and text vector retrieval. The server can directly return the retrieved candidate information. Furthermore, the server can supplement contact information of related talents from a pre-set database, thereby returning an information extraction result containing candidate information and contact information of related talents.

[0156] Keyword retrieval refers to a method of quickly locating relevant talent information using search keywords in an information retrieval request. This method first extracts a summary page containing the search keywords, then extracts the text containing the search keywords from the summary page, thus obtaining the candidate's talent information. This keyword retrieval method achieves precise matching and high retrieval accuracy.

[0157] Because keyword retrieval cannot identify the same entity expressed in different ways, it suffers from insufficient recall. For example, "South China University of Technology" and "SCUT" belong to the same entity; when the keyword is "South China University of Technology," the keyword retrieval cannot recall "SCUT." Based on this, this application proposes a text vector retrieval method. For any search keyword, it is converted into a corresponding text vector using word vector technology. The text vector is then input into a preset entity similarity model to obtain at least one candidate keyword with a similar relationship to the current search keyword. A summary page containing at least one candidate keyword is obtained from the open domain. Based on the text on the summary page containing at least one candidate keyword, the talent information of the candidate is determined. In this way, the text vector retrieval method can deeply explore candidate talent and effectively improve the recall rate of talent information.

[0158] The talent information, contact information, relationships between each candidate and related talent, and the corresponding relationship weights of each relationship can be written into the talent graph to enrich the information content of the talent graph and help improve the efficiency of information retrieval in subsequent information searches.

[0159] In addition, for the talent information of candidates recalled by the server, if the information type does not include all the preset information types, the server can supplement the talent information of the current candidate according to the relationship between the candidate and at least one corresponding related talent, and update the talent map based on the supplemented talent information of each candidate.

[0160] The aforementioned information acquisition method, in response to an information retrieval request, obtains talent information of candidates matching the request from an open domain. Based on the talent information and social relationship information of each candidate, it retrieves the contact information of associated talents corresponding to each candidate from a pre-set information database. Since the talent information obtained from the open domain often does not include contact information, and associated talents are those who have a relationship with the candidate, when selecting a target talent from among the candidates, the contact information of the target talent can be determined through the contact information of at least one associated talent. This method of obtaining the target talent's contact information through the contact information of associated talents is beneficial to improving information acquisition efficiency. Since the probability of obtaining the target talent's contact information through associated talents with different relationships varies, different association weights are determined based on the association relationships. This allows the contact information of at least one associated talent to be sorted according to the association weights, and the contact information of the target talent is determined according to the sorted contact information, which is beneficial to further improving information acquisition efficiency. Furthermore, this application embodiment, by combining real-time and offline construction components, achieves dynamic fusion, updating, and continuous expansion of the talent graph. This not only improves the timeliness and response speed of the talent graph but also ensures the comprehensiveness and integrity of the information, providing users with an efficient, real-time, and comprehensive talent information service. By combining keyword retrieval and text vector retrieval, it can process multi-source, complex structured data and further mine closely related and similarly capable talents through social relationship information, supplementing the recall results. In scenarios with complex talent data structures, it effectively improves the recall and accuracy of retrieval. By combining open domains and pre-set information databases, and through reasoning about association relationships, it can construct a multi-dimensional talent relationship network including colleagues, alumni, industry peers, and participants in the same activities, and obtain association weights by analyzing data such as tenure, activity participation, and social interaction. Based on the constructed talent graph and designed association weights, it can efficiently find the contact information of target talents, greatly improving the efficiency and rate of information acquisition.

[0161] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0162] Based on the same inventive concept, this application also provides an information acquisition device for implementing the information acquisition method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more information acquisition device embodiments provided below can be found in the limitations of the information acquisition method described above, and will not be repeated here.

[0163] In one exemplary embodiment, such as Figure 6 As shown, an information acquisition device 600 is provided, including: a first acquisition module 620, a second acquisition module 640, a first determination module 660, and a second determination module 680, wherein:

[0164] The first acquisition module 620 is used to respond to an information retrieval request and acquire talent information of candidates that match the information retrieval request in an open domain;

[0165] The second acquisition module 640 is used to acquire the contact information of at least one related talent corresponding to each candidate talent from a preset information database based on the talent information and social relationship information of each candidate talent.

[0166] The first determining module 660 is used to determine the association weight between the target talent and each corresponding associated talent in response to the selection of the target talent among the candidates.

[0167] The second determining module 680 is used to sort the contact information of at least one related talent according to the association weight, and determine the contact information of the target talent based on the sorted contact information.

[0168] The aforementioned information acquisition device, in response to an information retrieval request, acquires talent information of candidates matching the request in an open domain. Based on the talent information and social relationship information of each candidate, it retrieves contact information of associated talents corresponding to each candidate from a preset information database. Since the talent information acquired in the open domain often does not include contact information, and associated talents are those who have a relationship with the candidate, when selecting a target talent from among the candidates, the contact information of the target talent can be determined by using the contact information of at least one associated talent. This method of obtaining the target talent's contact information through the contact information of associated talents is beneficial to improving information acquisition efficiency. Since the probability of obtaining the target talent's contact information through associated talents with different relationships varies, different association weights are determined based on the association relationships. This allows the contact information of at least one associated talent to be sorted according to the association weights, and the contact information of the target talent is determined according to the sorted contact information, which is beneficial to further improving information acquisition efficiency.

[0169] In one embodiment, the first acquisition module 620 is further configured to: acquire search keywords from the information retrieval request; acquire a summary page containing at least one search keyword in the open domain; and determine the talent information of the candidate based on the text containing at least one search keyword in the summary page.

[0170] In one embodiment, the first acquisition module 620 is further configured to: for any search keyword, acquire at least one candidate keyword that has a similar relationship with the current search keyword through a preset entity similarity model; acquire the summary page containing at least one candidate keyword from the open domain; and determine the talent information of the candidate based on the text containing at least one search keyword in the summary page. The first acquisition module 620 is further configured to: determine the talent information of the candidate based on the text containing at least one candidate keyword in the summary page.

[0171] In one embodiment, based on the talent information and social relationship information of each candidate, the contact information of at least one associated talent corresponding to each candidate is obtained from a preset information database. The second acquisition module 640 is further configured to: for each candidate, obtain target information that matches at least one piece of information in the social relationship information or talent information of the current candidate from the preset information database; regard the talent corresponding to the target information as the associated talent corresponding to the current candidate, and obtain the contact information of the associated talent from the target information.

[0172] In one embodiment, based on the association relationship between the target talent and at least one corresponding associated talent, the association weight between the target talent and each corresponding associated talent is determined. The first determining module 660 is further configured to: obtain the mapping relationship between the association relationship and the association weight; find the mapping relationship and use the association weight corresponding to the association relationship as the association weight between the target talent and the corresponding associated talent.

[0173] In one embodiment, the first acquisition module 620 is further configured to: determine whether there is talent information matching the information retrieval request in the pre-constructed talent map; if not, acquire talent information matching the information retrieval request in the open domain.

[0174] In one embodiment, the information acquisition device 600 further includes a first update module, which is used to: for each candidate, acquire the information type to which each talent information of the current candidate belongs; if the information type does not include all preset information types, supplement the talent information of the current candidate according to the association relationship between the candidate and at least one corresponding related talent; and update the talent map based on the supplemented talent information of each candidate.

[0175] In one embodiment, the information acquisition device 600 further includes a second update module, which is used to update the talent map based on the talent information of each candidate talent, the talent information of each associated talent, contact information, association relationship and association weight.

[0176] In one embodiment, the information acquisition device 600 further includes a third update module, which is used to: correct the talent information of each candidate to obtain corrected talent information; the correction includes at least one of entity alignment, information repair and data verification; and update the talent map based on the corrected talent information of each candidate.

[0177] Each module in the aforementioned information acquisition device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0178] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores a talent map. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When executed by the processor, the computer program implements an information acquisition method.

[0179] Those skilled in the art will understand that Figure 7The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0180] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0181] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0182] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0183] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0184] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0185] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0186] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An information acquisition method, characterized in that, The method includes: In response to an information retrieval request, talent information of candidates matching the information retrieval request is obtained in an open domain; Based on the talent information and social relationship information of each candidate, retrieve the contact information of at least one related talent corresponding to each candidate from the preset information database; In response to the selection of target talents among the candidates, the association weight between the target talent and each corresponding associated talent is determined based on the association relationship between the target talent and at least one associated talent. The contact information of the at least one related talent is sorted according to the association weight, and the contact information of the target talent is determined based on the sorted contact information.

2. The method according to claim 1, characterized in that, The step of obtaining talent information of candidates matching the information retrieval request in the open domain includes: Obtain the search keywords from the information retrieval request; Retrieve a summary page containing at least one search keyword in an open domain; Talent information of candidates is determined based on text containing at least one search keyword on the summary page.

3. The method according to claim 2, characterized in that, The step of obtaining a summary page containing at least one search keyword in an open domain includes: For any search keyword, at least one candidate keyword with a similar relationship to the current search keyword is obtained through a preset entity similarity model; Retrieve a summary page containing at least one candidate keyword from the open domain; The step of determining candidate talent information based on text containing at least one search keyword on the summary page includes: Talent information of candidates is determined based on the text containing at least one candidate keyword on the summary page.

4. The method according to claim 1, characterized in that, The step of retrieving contact information of at least one associated talent for each candidate from a pre-set database based on the candidate's talent information and social relationship information includes: For each candidate, target information matching at least one piece of information from the preset information database, either social relationship information or talent information, is retrieved. The talent corresponding to the target information is regarded as the associated talent corresponding to the current candidate talent, and the contact information of the associated talent is obtained from the target information.

5. The method according to claim 1, characterized in that, The step of determining the association weight between the target talent and each corresponding associated talent based on the association relationship between the target talent and at least one associated talent includes: Obtain the mapping relationship between association relationships and association weights; Find the mapping relationship and use the association weight corresponding to the association relationship as the association weight between the target talent and the corresponding associated talent.

6. The method according to claim 1, characterized in that, The step of obtaining talent information of candidates matching the information retrieval request in the open domain includes: Determine whether there is talent information in the pre-constructed talent map that matches the information retrieval request; If not found, then retrieve the talent information of the candidate matching the information retrieval request from the open domain.

7. The method according to claim 6, characterized in that, The method further includes: For each candidate, obtain the individual information types of each candidate's talent information; If the information type does not include all preset information types, the talent information of the current candidate is supplemented according to the association relationship between the candidate and at least one corresponding related talent. The talent map is updated based on the talent information supplemented by each candidate.

8. The method according to claim 6, characterized in that, The method further includes: The talent map is updated based on the talent information of each candidate, the talent information of each associated talent, contact information, relationship, and association weight.

9. The method according to claim 6, characterized in that, The method further includes: The talent information of each candidate is corrected to obtain the corrected talent information; the correction includes at least one of entity alignment, information repair and data verification. The talent map is updated based on the revised talent information of each candidate.

10. An information acquisition device, characterized in that, The device includes: The first acquisition module is used to respond to an information retrieval request and acquire talent information of candidates that match the information retrieval request in an open domain; The second acquisition module is used to acquire the contact information of at least one related talent corresponding to each candidate from a preset information database based on the talent information and social relationship information of each candidate. The first determining module is used to determine the association weight between the target talent and each corresponding associated talent in response to the selection of the target talent among the candidates; The second determining module is used to sort the contact information of the at least one related talent according to the association weight, and determine the contact information of the target talent based on the sorted contact information.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

Images