Vector search and large language model-based talent recommendation method and device and medium

Abstract

The invention discloses a talent recommendation method and device based on vector search and a large language model, and a medium, and the method comprises the following steps: building a talent scientific research achievement database based on a big data technology, vectorizing the database, and building a vector model database; obtaining a target talent demand template, vectorizing the target talent demand template to obtain a target talent demand vector model, and comparing the target talent demand vector model with the vector model database to obtain similarity; judging whether the similarity is greater than or equal to a set similarity threshold value or not; if the target talent demand template is greater than or equal to the target talent demand template, generating a clustering document, acquiring scientific research achievement vector model data meeting the target talent demand template based on the clustering document, and analyzing key information in the data; constructing a large language model, inputting the key information into the large language model, generating specific information of related talents, and performing screening to obtain recommendable talents; if yes, it is judged that no recommendable talents exist.

Description

Technical Field

[0001] The present invention relates to the technical field of talent recommendation, and in particular to a talent recommendation method, device and medium based on vector search and large language model. Background Art

[0002] With the development of deep learning technology, large language models have powerful application capabilities in various fields. In natural language processing, they can be used for text search, machine translation, text generation, sentiment analysis, named entity recognition and other functions. Before the emergence of large language models, vector retrieval technology had already developed. With the technology of deep learning, vector retrieval is widely used in artificial intelligence application fields such as image, audio, video search and recommendation, face recognition, speech recognition, etc.

[0003] The traditional search method is keyword search, which plays an extremely important role in Internet search. However, with the exponential increase in data volume, the speed and accuracy of keyword search have been challenged. The new vector search method provides a new and effective method for the application of big data technology. In addition, the combination of vector search and large language models provides a new method for data matching and recommendation, solving the problems of slow speed and poor accuracy of traditional search methods. Summary of the invention

[0004] The purpose of the present invention is to propose a talent recommendation method based on vector search and large language model to solve the problems of slow search speed and low accuracy of recommendation results in matching and recommendation methods in related technologies.

[0005] In order to achieve the above-mentioned purpose of the invention, the technical solution adopted by the present invention is: a talent recommendation method based on vector search and large language model, comprising the following steps:

[0006] Establish a talent research results database based on big data technology, vectorize the database, and establish a vector model database;

[0007] Obtain a target talent demand template, and vectorize the target talent demand template to obtain a target talent demand vector model, and compare the target talent demand vector model with a vector model database to obtain a similarity;

[0008] Determining whether the similarity is greater than or equal to a set similarity threshold;

[0009] If it is greater than or equal to, a cluster document is generated, and the scientific research achievement vector model data that meets the target talent demand template is obtained based on the cluster document, and the key information in the data is parsed;

[0010] Build a big language model, input key information into the big language model, generate specific information of relevant talents, screen them, and obtain talents that can be recommended;

[0011] If it is less than, it is determined that there is no talent to recommend.

[0012] Furthermore, a talent research results database is established based on big data technology, the database is vectorized, and a vector model database is established, including:

[0013] Apply big data technology to mine the world's public academic papers, patents and award-winning data to generate talent research results data and establish a dedicated database.

[0014] Vectorize the database to generate vectorized globally public academic papers, patents and award-winning data;

[0015] A vector model database is obtained based on vectorized global public academic papers, patents and award-winning data.

[0016] Furthermore, we will apply big data technology to mine the world's publicly available academic papers, patents and award-winning data to generate talent research results data and establish a dedicated database, including:

[0017] Input the instructions for target talent needs in the form of voice or text, and convert them into the language required by the talent research results database;

[0018] Based on the language required by the special database of talents' scientific research results, data on talents' scientific research results are obtained and a special database is established.

[0019] Furthermore, based on the clustered documents, the scientific research achievement vector model data that meets the target talent demand template is obtained, and the key information in the data is parsed, including:

[0020] Obtain cluster documents, analyze the data in the cluster documents, and obtain analysis data;

[0021] Compare the analysis data with the scientific research achievement vector model data of the target talent demand template to obtain a comparison result;

[0022] Determine whether the comparison result meets the requirements;

[0023] If it meets the requirements, analyze the key information in the data;

[0024] If it does not meet the requirements, the corresponding scientific research achievement vector model data will be eliminated according to the needs of the target talent demand template.

[0025] Furthermore, the key information includes the name, organization and partnership in the data.

[0026] Furthermore, a large language model is constructed, key information is input into the large language model, specific information of relevant talents is generated, and screening is performed to obtain talents that can be recommended, including:

[0027] Obtain historical recommendation data, generate a training set, iteratively train the initial model based on the training set, and obtain training results;

[0028] Determining whether the training result converges;

[0029] If converged, a large language model is generated, and specific information of relevant talents is generated based on the large language model, including nationality, age, and gender;

[0030] If it does not converge, adjust the number of iterations and train the model a second time until the model converges.

[0031] Furthermore, a target talent demand template is established based on the subject direction, research positioning, research field, research direction and research trend proposed by the employer.

[0032] Furthermore, the similarity calculation adopts cosine similarity calculation, and the calculation formula of the similarity is:

[0033] Among them, k i , l i They are the elements of the target talent demand vector and the single talent scientific research achievement data vector respectively.

[0034] The present invention also provides a talent recommendation device based on vector search and large language model, comprising a processor, a memory and at least one program, wherein the program is stored in the memory and is configured to be executed by the processor, and the program includes instructions for executing the talent recommendation method based on vector search and large language model as described in any one of the above items.

[0035] The present invention also provides a computer-readable storage medium, which stores a computer program, and the computer program enables a computer to execute to implement any of the above-mentioned talent recommendation methods based on vector search and large language models.

[0036] Due to the application of the above-mentioned technical scheme, the present invention has the following advantages compared with the prior art: the present application converts the instructions of target talent demand through input methods such as voice and text into the language required by the talent scientific research achievement dedicated database, realizes the vectorization of target talent demand templates and talent scientific research achievement data, establishes a talent scientific research achievement vector model database, and adopts distributed calculation of the similarity between the vector of the target talent demand template and the talent scientific research achievement vector model data, generates a target talent demand clustering document, and obtains the information of the target talent through a large language model, thereby improving the accuracy of talent recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1The overall algorithm block diagram of the talent recommendation method based on vector search and large language model provided by an embodiment of the present invention is shown.

[0038] Figure 2 A flow chart of a method for establishing a vector model database provided by an embodiment of the present invention is shown;

[0039] Figure 3 A flow chart of a dedicated database generation method provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0040] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0041] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, device, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0042] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0043] Please refer to Figure 1 , Figure 1 This is a flow chart of a talent recommendation method based on vector search and large language model in some embodiments of the present application. The talent recommendation method based on vector search and large language model is used in a terminal device, and the talent recommendation method based on vector search and large language model includes the following steps:

[0044] S101, establish a talent scientific research results database based on big data technology, vectorize the database, and establish a vector model database;

[0045] S102, obtaining a target talent demand template, and vectorizing the target talent demand template to obtain a target talent demand vector model, and comparing the target talent demand vector model with a vector model database to obtain a similarity;

[0046] S103, determining whether the similarity is greater than or equal to a set similarity threshold;

[0047] S104, if it is greater than or equal to, generating a cluster document, obtaining scientific research achievement vector model data that meets the target talent demand template based on the cluster document, and parsing key information in the data;

[0048] S105, constructing a large language model, inputting key information into the large language model, generating specific information of relevant talents, screening, and obtaining talents that can be recommended;

[0049] S106: If it is less than, it is determined that there is no talent to recommend.

[0050] Specifically, a talent science and technology achievement database is established. Through big data technology, data on talent science and technology achievements such as public papers, patents, and awards are mined. Due to the massive amount of data, servers, cloud storage, and other storage media are needed to store the data and form a distributed database.

[0051] The database of talents and scientific and technological achievements is vectorized, and a vector database is established. The vectorization of the database includes the vectorization of data storage and operation, which not only establishes a vector database, but also realizes vectorized calculation, and is carried out in specific devices and equipment.

[0052] Obtain the target talent demand template. The target talent demand template description can be input by voice, text, text pasting and file import through the device; according to the characteristics of the talent scientific and technological achievements database, the description content needs to be converted into a language in the device; in the device, the converted target talent demand template is vectorized.

[0053] Please refer to Figure 2 , Figure 2 This is a flow chart of a method for establishing a vector model database for a talent recommendation method based on vector search and large language model in some embodiments of the present application. According to an embodiment of the present invention, a talent scientific research achievement database is established based on big data technology, the database is vectorized, and a vector model database is established, specifically including:

[0054] S201, use big data technology to mine the world's public academic papers, patents and award data to generate talent research results data and establish a dedicated database.

[0055] S202, vectorize the database to generate vectorized globally public academic papers, patents and award-winning data;

[0056] S203, obtain a vector model database based on vectorized global public academic papers, patents and award-winning data.

[0057] Please refer to Figure 3 , Figure 3 This is a flowchart of a dedicated database generation method for a talent recommendation method based on vector search and large language model in some embodiments of the present application. According to an embodiment of the present invention, big data technology is used to mine globally published academic papers, patents and award-winning data to generate talent scientific research results data, and a dedicated database is established, specifically including:

[0058] S301, inputting instructions for target talent needs, in the form of voice or text, and converting them into the language required by the talent scientific research achievement database;

[0059] S302, acquiring the data on the scientific research achievements of talents based on the language required by the special database of scientific research achievements of talents, and establishing a special database.

[0060] According to an embodiment of the present invention, scientific research achievement vector model data that meets the target talent demand template is obtained based on cluster documents, and key information in the data is parsed, specifically including:

[0061] Obtain cluster documents, analyze the data in the cluster documents, and obtain analysis data;

[0062] Compare the analysis data with the scientific research achievement vector model data of the target talent demand template to obtain a comparison result;

[0063] Determine whether the comparison result meets the requirements;

[0064] If it meets the requirements, analyze the key information in the data;

[0065] If it does not meet the requirements, the corresponding scientific research achievement vector model data will be eliminated according to the needs of the target talent demand template.

[0066] According to the embodiment of the present invention, the key information includes the name, organization and cooperation relationship in the data.

[0067] According to an embodiment of the present invention, a large language model is constructed, key information is input into the large language model, specific information of relevant talents is generated, and screening is performed to obtain talents that can be recommended, specifically including:

[0068] Obtain historical recommendation data, generate a training set, iteratively train the initial model based on the training set, and obtain training results;

[0069] Determine whether the training results converge;

[0070] If converged, a large language model is generated, and specific information of relevant talents is generated based on the large language model, including nationality, age, and gender;

[0071] If it does not converge, adjust the number of iterations and train the model a second time until the model converges.

[0072] According to an embodiment of the present invention, a target talent demand template is established according to the discipline direction, research positioning, research field, research direction and research trend proposed by the employer.

[0073] Specifically, it includes: setting a similarity threshold, generating target talent clustering documents for talent scientific research results that are greater than the similarity threshold, obtaining talent scientific research results vector model data that meets the target talent demand template, and parsing out key information such as name, organization, and partnership in the data, inputting the parsed name, organization, and partnership information into the large language model, obtaining specific information of relevant talents, screening based on nationality, age, gender, and other requirements, and obtaining recommendable talents.

[0074] According to the embodiment of the present invention, the similarity calculation adopts cosine similarity calculation, and the calculation formula of the similarity is:

[0075] Among them, ki and li are the elements of the target talent demand vector and the single talent scientific research achievement data vector respectively.

[0076] The working principle of the present invention is as follows: the talent target demand template includes the discipline direction, research positioning, research field, research direction and research trend of the employer; target talent demand matching, vectorizing the target talent demand template, searching all vectors with the vector model database of the talent scientific research results, and performing similarity analysis, setting the similarity threshold, generating clustering documents based on the similarity, obtaining the scientific research results vector model data that meets the target talent demand template, and parsing out key information such as name, organization, and cooperative relationship in the data; talent recommendation based on the large language model, inputting the parsed name, organization, and cooperative relationship information into the large language model, obtaining the specific introduction of the relevant talents, screening, and obtaining recommendable talents. Its devices, equipment and storage media include: a talent discovery and recommendation device based on vectors and large language models, which is used for instructions on target talent needs; a talent discovery and recommendation device based on vectors and large language models, which is used for language vectorization of target talent demand templates and vector model databases of talent scientific research results when matching target talent needs, calculating similarities, generating clustering documents, and displaying talent recommendations based on large language models; a talent discovery and recommendation storage medium based on vectors and large language models, which is used for storage of massive global talent papers, patents and award data, as well as vectorized data.

[0077] To summarize, this application converts the target talent demand instructions into the language required by the talent scientific research results dedicated database through input methods such as voice and text, realizes the vectorization of target talent demand templates and talent scientific research results data, establishes a talent scientific research results vector model database, and uses distributed calculations to calculate the similarity between the vector of the target talent demand template and the talent scientific research results vector model data, generates target talent demand clustering documents, and obtains target talent information through a large language model to improve the accuracy of talent recommendations.

[0078] This embodiment also provides a talent recommendation device based on vector search and large language model, including a processor, a memory and at least one program, the program is stored in the memory and is configured to be executed by the processor, the program includes instructions for executing any of the above-mentioned talent recommendation methods based on vector search and large language model.

[0079] Those skilled in the art will appreciate that, for ease of description, the example in which both the memory and the processor are provided with one is used for description. In an actual terminal or server, there may be multiple processors and memories. The memory may also be referred to as a storage medium or a storage device, etc., which is not limited in the embodiments of the present application.

[0080] It should be understood that in the embodiments of the present application, the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The processor may also be a general-purpose microprocessor, a graphics processing unit (GPU), or one or more integrated circuits for executing related programs to implement the functions required to be executed in the embodiments of the present application.

[0081] The processor can also be an integrated circuit chip with signal processing capabilities. In the implementation process, the various steps of the present application can be completed by the integrated logic circuit of the hardware in the processor or the instructions in the form of software. The above-mentioned processor can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or a combination of hardware and software modules in the decoding processor to be executed. The software module can be located in a random access memory, a flash memory and a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register and other mature storage media in the art. The storage medium is located in the memory, and the processor reads the information in the memory, and combines its hardware to complete the functions required to be performed by the unit included in the method, device and storage medium of the embodiment of the present application.

[0082] It should also be understood that the memory mentioned in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache.

[0083] By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct RAM bus random access memory (DR RAM).

[0084] The memory may also be a read-only optical disc (Compact Disc Read-Only Memory, CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may be independent and connected to the processor via a bus. The memory may also be integrated with the processor, and the memory may store a program. When the program stored in the memory is executed by the processor, the processor is used to execute the various steps of the determination method in the above-mentioned embodiment of the present application.

[0085] It should be noted that when the processor is a general-purpose processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, the memory (storage module) is integrated into the processor. It should be noted that the memory described herein is intended to include but is not limited to these and any other suitable types of memory.

[0086] It should be understood that the term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0087] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software. The steps of the method disclosed in conjunction with the embodiment of the present application can be directly embodied as a hardware processor for execution, or a combination of hardware and software modules in a processor for execution. The software module can be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory, and completes the steps of the above method in conjunction with its hardware. To avoid repetition, it is not described in detail here.

[0088] Those of ordinary skill in the art will appreciate that the various illustrative logical blocks (ILBs) and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0089] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a program product of computer programming. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a processor, all or part of the processes or functions according to the embodiments of the present application are generated. The computer may be a general-purpose computer, a computer network, or other programmable device.

[0090] This embodiment also provides a computer-readable storage medium, which stores a computer program. The computer program enables a computer to execute to implement the above-mentioned talent recommendation method based on vector search and large language model.

[0091] It should be noted that computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., coaxial cable, optical fiber) or wireless (e.g., infrared, wireless, microwave, etc.) means, or can be transmitted from one website, computer, server or data center to a mobile phone processor by wired means. Computer-readable storage media can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media. Available media can be magnetic media (e.g., floppy disk, hard disk), optical media (e.g., DVD), or semiconductor media (e.g., solid-state hard disk), etc.

[0092] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A talent recommendation method based on vector search and large language model, characterized in that: The following steps are involved: Establish a talent research results database based on big data technology, vectorize the database, and establish a vector model database; Obtain a target talent demand template, and vectorize the target talent demand template to obtain a target talent demand vector model, and compare the target talent demand vector model with a vector model database to obtain a similarity; Determining whether the similarity is greater than or equal to a set similarity threshold; If it is greater than or equal to, a cluster document is generated, and the scientific research achievement vector model data that meets the target talent demand template is obtained based on the cluster document, and the key information in the data is parsed; Build a big language model, input key information into the big language model, generate specific information of relevant talents, screen them, and obtain talents that can be recommended; If it is less than, it is determined that there is no talent to recommend.

2. The talent recommendation method based on vector search and large language model as claimed in claim 1, characterized in that: Establish a talent scientific research results database based on big data technology, vectorize the database, and establish a vector model database, including: Apply big data technology to mine the world's public academic papers, patents and award-winning data to generate talent research results data and establish a dedicated database. Vectorize the database to generate vectorized globally public academic papers, patents and award-winning data; A vector model database is obtained based on vectorized global public academic papers, patents and award-winning data.

3. The talent recommendation method based on vector search and large language model as claimed in claim 2, characterized in that: Apply big data technology to mine the world's public academic papers, patents and award data to generate talent research results data and establish a dedicated database, including: Input the instructions for target talent needs in the form of voice or text, and convert them into the language required by the talent research results database; Based on the language required by the special database of talents' scientific research results, data on talents' scientific research results are obtained and a special database is established.

4. The talent recommendation method based on vector search and large language model as claimed in claim 3, characterized in that: Based on the clustering documents, we obtain the scientific research achievement vector model data that meets the target talent demand template, and parse out the key information in the data, including: Obtain cluster documents, analyze the data in the cluster documents, and obtain analysis data; Compare the analysis data with the scientific research achievement vector model data of the target talent demand template to obtain a comparison result; Determine whether the comparison result meets the requirements; If it meets the requirements, analyze the key information in the data; If it does not meet the requirements, the corresponding scientific research achievement vector model data will be eliminated according to the needs of the target talent demand template.

5. The talent recommendation method based on vector search and large language model as claimed in claim 4, characterized in that: The key information includes the name, organization and partnership in the data.

6. The talent recommendation method based on vector search and large language model as claimed in claim 5, characterized in that: Build a big language model, input key information into the big language model, generate specific information of relevant talents, screen them, and obtain talents that can be recommended, including: Obtain historical recommendation data, generate a training set, iteratively train the initial model based on the training set, and obtain training results; Determining whether the training result converges; If converged, a large language model is generated, and specific information of relevant talents is generated based on the large language model, including nationality, age, and gender; If it does not converge, adjust the number of iterations and train the model a second time until the model converges.

7. The talent recommendation method based on vector search and large language model as claimed in claim 5, characterized in that: Establish a target talent demand template based on the subject direction, research positioning, research field, research direction and research trend proposed by the employer.

8. The talent recommendation method based on vector search and large language model as claimed in claim 5, characterized in that: The similarity calculation adopts cosine similarity calculation, and the calculation formula of the similarity is: Among them, k i , l i They are the elements of the target talent demand vector and the single talent scientific research achievement data vector respectively.

9. A talent recommendation device based on vector search and large language model, characterized in that: It includes a processor, a memory and at least one program, wherein the program is stored in the memory and is configured to be executed by the processor, and the program includes instructions for executing the talent recommendation method based on vector search and large language model as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which enables a computer to execute to implement the talent recommendation method based on vector search and a large language model according to any one of claims 1 to 8.

Images