Method and system for perceiving and acquiring scientific and technological intelligence based on generative artificial intelligence

Abstract

The present invention belongs to the field of artificial intelligence technology, and discloses a method and system for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence. The method comprises the following steps: acquiring initial scientific and technological intelligence data based on intelligence perception; performing data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data; identifying all entity reference items in the reference scientific and technological intelligence data; acquiring candidate entities based on the entity reference items using a generative artificial intelligence model; determining a target entity based on the candidate entity; and establishing an association between the target entity and a scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data. Through the above-mentioned method, the method can cope with changes in the layout and structure of any website, effectively extract information from a variety of website structures, improve the efficiency of acquiring scientific and technological intelligence data, and further ensure the accuracy and completeness of the scientific and technological intelligence data using a generative artificial intelligence model.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and system for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence. Background Art

[0002] For a long time, scientific and technological intelligence has played an important supporting role in scientific research, technological innovation, industrial development, social progress, and management decision-making, fully embodying the core value of "eyes and ears, vanguard, and adviser". In order to overcome a series of key and bottleneck technologies in the increasingly fierce competition and seize the technological high ground, it is inseparable from the continuous and in-depth follow-up support of scientific and technological intelligence work.

[0003] Traditional intelligence acquisition suffers from information overload caused by massive open source intelligence data, and a large amount of "weak signal" intelligence that has predictive value but is vague and fragmented is difficult to detect. It is still necessary to rely on the business experience and research capabilities of research institutions or consulting experts to obtain effective intelligence information. This method is not only inefficient, but the information obtained also has problems with poor accuracy and completeness.

[0004] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of the present invention is to provide a method and system for perceiving and acquiring scientific and technological intelligence based on generative artificial intelligence, aiming to solve the technical problems that the existing technology has low efficiency in acquiring information and poor accuracy and completeness.

[0006] To achieve the above-mentioned purpose, the present invention provides a method for perceiving and acquiring scientific and technological intelligence based on generative artificial intelligence, and the method for perceiving and acquiring scientific and technological intelligence based on generative artificial intelligence comprises the following steps:

[0007] Acquire initial scientific and technological intelligence data based on intelligence perception;

[0008] Performing data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data;

[0009] Identify all entity references in the referenced scientific and technological intelligence data;

[0010] Obtaining a candidate entity based on the entity referent using a generative artificial intelligence model;

[0011] Determine a target entity according to the candidate entity;

[0012] An association is established between the target entity and a scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data.

[0013] Optionally, the obtaining of initial scientific and technological intelligence data based on intelligence perception includes:

[0014] Send an information extraction request to obtain the Uniform Resource Locator of the journal page associated with scientific and technological information;

[0015] Extracting corresponding fields from the journal page based on the uniform resource locator, and determining whether there are any remaining unobtained uniform resource locators;

[0016] If it does not exist, all the extracted fields are formatted and stored in the database to obtain the initial scientific and technological intelligence data.

[0017] Optionally, the performing data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data includes:

[0018] Normalizing the initial scientific and technological intelligence data to obtain processed scientific and technological intelligence data;

[0019] Detecting and repairing real words, homophones and short text misunderstandings on the processed scientific and technological intelligence data;

[0020] Calculate the comprehensive accuracy score of the detected and repaired S&T intelligence data;

[0021] If the comprehensive accuracy score does not reach the preset score, returning to execute the detection and repair of true words, homophones and short text misunderstandings on the processed scientific and technological intelligence data;

[0022] If the comprehensive accuracy score reaches the preset score, the scientific and technological intelligence data after detection and repair will be used as reference scientific and technological intelligence data.

[0023] Optionally, the obtaining a candidate entity based on the entity referent by using a generative artificial intelligence model includes:

[0024] Obtain the relationship between entities corresponding to entity referents through a generative artificial intelligence model;

[0025] Constructing an adjacency matrix according to the relationships between the entities;

[0026] Determining an embedding representation of a candidate entity based on the adjacency matrix, the coefficient matrix of the generative artificial intelligence model, the bias vector, and a preset activation function;

[0027] A candidate entity is obtained based on the embedded representation of the candidate entity.

[0028] Optionally, acquiring a candidate entity based on the embedded representation of the candidate entity comprises:

[0029] Obtaining contextual features of entity referents;

[0030] Constructing a feature vector of the context feature and a vector corresponding to the embedding representation of the candidate entity in the same dimension respectively;

[0031] Concatenate the feature vector of the context feature in the same dimension and the vector corresponding to the embedding representation of the candidate entity to obtain a candidate entity vector;

[0032] A candidate entity is obtained according to the candidate entity vector.

[0033] Optionally, determining the target entity according to the candidate entity includes:

[0034] Linking the candidate entity with a preset entry;

[0035] Compare the context features between the candidate entity and the corresponding entity of the linked entry;

[0036] disambiguating the candidate entity based on the comparison result;

[0037] Determine the target entity based on the disambiguated candidate entities.

[0038] Optionally, determining the target entity according to the disambiguated candidate entities includes:

[0039] Determine the similarity between the disambiguated candidate entity and the predicted entity using a preset classifier;

[0040] performing a similarity score according to the similarity;

[0041] The disambiguated candidate entity with the largest similarity score is taken as the target entity.

[0042] In addition, to achieve the above-mentioned purpose, the present invention also proposes a scientific and technological intelligence perception and acquisition system based on generative artificial intelligence, and the scientific and technological intelligence perception and acquisition system based on generative artificial intelligence includes:

[0043] The acquisition module is used to obtain initial scientific and technological intelligence data based on intelligence perception;

[0044] A cleaning module, used for cleaning the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data;

[0045] An identification module, used to identify all entity reference items in the reference scientific and technological intelligence data;

[0046] A construction module is used to obtain candidate entities based on the entity referent using a generative artificial intelligence model;

[0047] A screening module, used to determine a target entity according to the candidate entity;

[0048] The processing module is used to establish an association between the target entity and the scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data.

[0049] Optionally, the acquisition module is further used to send an information extraction request to obtain the uniform resource locator of the journal page associated with the scientific and technological information;

[0050] Extracting corresponding fields from the journal page based on the uniform resource locator, and determining whether there are any remaining unobtained uniform resource locators;

[0051] If it does not exist, all the extracted fields are formatted and stored in the database to obtain the initial scientific and technological intelligence data.

[0052] Optionally, the cleaning module is further used to normalize the initial scientific and technological intelligence data to obtain processed scientific and technological intelligence data;

[0053] Detecting and repairing real words, homophones and short text misunderstandings on the processed scientific and technological intelligence data;

[0054] Calculate the comprehensive accuracy score of the detected and repaired S&T intelligence data;

[0055] If the comprehensive accuracy score does not reach the preset score, returning to execute the detection and repair of true words, homophones and short text misunderstandings on the processed scientific and technological intelligence data;

[0056] If the comprehensive accuracy score reaches the preset score, the scientific and technological intelligence data after detection and repair will be used as reference scientific and technological intelligence data.

[0057] The present invention obtains initial scientific and technological intelligence data based on intelligence perception; performs data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data; identifies all entity reference items in the reference scientific and technological intelligence data; uses a generative artificial intelligence model to obtain candidate entities based on the entity reference items; determines a target entity based on the candidate entity; and constructs an association between the target entity and a scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data. Through the above-mentioned method, it is possible to cope with changes in the layout and structure of any website, effectively extract information from a variety of website structures, improve the efficiency of obtaining scientific and technological intelligence data, and further ensure the accuracy and completeness of the scientific and technological intelligence data by using a generative artificial intelligence model. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flow chart of the first embodiment of the method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence of the present invention;

[0059] Figure 2It is a flow chart of the second embodiment of the method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence of the present invention;

[0060] Figure 3 This is a structural block diagram of the first embodiment of the scientific and technological intelligence perception and acquisition system based on generative artificial intelligence of the present invention.

[0061] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0062] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0063] The embodiment of the present invention provides a method for perceiving and acquiring scientific and technological intelligence based on generative artificial intelligence, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of a method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence according to the present invention.

[0064] In this embodiment, the method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence includes the following steps:

[0065] Step S10: Acquire initial scientific and technological intelligence data based on intelligence perception.

[0066] In this embodiment, the executor of this embodiment is a scientific and technological intelligence perception and acquisition device based on generative artificial intelligence, wherein the scientific and technological intelligence perception and acquisition device based on generative artificial intelligence has functions such as data processing, data communication and program running. The scientific and technological intelligence perception and acquisition device based on generative artificial intelligence can be a terminal data processing device such as a computer, and of course it can also be other devices with similar functions, and this embodiment does not limit this.

[0067] It should be noted that traditional intelligence acquisition suffers from information overload caused by massive open source intelligence data, and a large amount of "weak signal" intelligence that has predictive value but is vague and fragmented is difficult to detect. It is still necessary to rely on the business experience and research capabilities of research institutions or consulting experts to obtain effective intelligence information. This method is not only inefficient, but the information obtained also has problems with poor accuracy and completeness.

[0068] In order to solve the above technical problems, this embodiment obtains initial scientific and technological intelligence data based on intelligence perception; performs data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data; identifies all entity reference items in the reference scientific and technological intelligence data; uses a generative artificial intelligence model to obtain candidate entities based on the entity reference items; determines the target entity based on the candidate entity; and constructs an association between the target entity and a scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data. Through the above method, it is possible to cope with changes in the layout and structure of any website, effectively extract information from a variety of website structures, and improve the efficiency of obtaining scientific and technological intelligence data. At the same time, a generative artificial intelligence model is used to further ensure the accuracy and completeness of the scientific and technological intelligence data. Specifically, it can be achieved in the following way.

[0069] In the specific implementation, in this embodiment, it is necessary to first obtain the initial scientific and technological intelligence data based on intelligence perception. The process of obtaining the initial scientific and technological intelligence data is specifically to send an information extraction request to obtain the uniform resource locator of the journal page associated with the scientific and technological intelligence. The uniform resource locator can be obtained in the above manner for traditional HTML structure type websites, non-traditional HTML structure type websites, and websites based on AJAX technology. Then, based on the uniform resource locator, the corresponding fields are extracted from the journal page. In this embodiment, the corresponding field information can be extracted by xpath extraction technology combined with json formatting and regular expressions. All the extracted fields are formatted and stored in the database to obtain the initial scientific and technological intelligence data. It should be emphasized that in order to make the obtained data more comprehensive, in this embodiment, it is necessary to further determine whether all the uniform resource locators have been obtained. If there are remaining uniform resource locators that have not been obtained, continue to return to the step of obtaining the uniform resource locator.

[0070] Step S20: performing data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data.

[0071] In the specific implementation, in this embodiment, after the initial scientific and technological intelligence data is collected, in order to improve the accuracy of the data, it is necessary to perform data cleaning on the initial scientific and technological intelligence data. When performing data cleaning, the initial scientific and technological intelligence data needs to be normalized first, and then the processed initial scientific and technological intelligence data is detected and repaired for real words, homophones, and short passage misunderstandings. In this embodiment, real word and short passage misunderstanding detection can be performed through fixed collocations. For example, "wearing clothes" will not be expressed as "wearing a hat". In this embodiment, real word and short passage misunderstanding detection can also be performed in combination with the context. Then, the comprehensive accuracy score after detection and repair is calculated. For example, the sum of the fixed collocation score and the context score is the comprehensive accuracy score. For homophones, monitoring and repair can be performed by phonetic matching or pinyin matching. It can be understood that if the comprehensive accuracy score reaches the preset score, it means that the detection and repair are completed, and the scientific and technological intelligence data after detection and repair can be used as reference scientific and technological intelligence data.

[0072] Step S30: Identify all entity reference items in the reference scientific and technological intelligence data.

[0073] After obtaining the reference scientific and technological intelligence data, in this embodiment, it is necessary to obtain all entity reference items therefrom, such as names related to scientific and technological intelligence, to facilitate subsequent information extraction, text understanding and content indexing.

[0074] Step S40: Utilize a generative artificial intelligence model to obtain candidate entities based on the entity referent.

[0075] In this embodiment, a generative artificial intelligence model is further used to obtain candidate entities. Specifically, it is necessary to first obtain an embedded representation of the candidate entity, and then determine the candidate entity based on the embedded representation.

[0076] Step S50: determining a target entity according to the candidate entity.

[0077] After obtaining the candidate entities, in this embodiment, it is necessary to further filter out the target entity from the candidate entities. The filtering can be based on the similarity between the candidate entities and the predicted entity, and the candidate entity with the greatest similarity is used as the target entity.

[0078] Step S60: constructing an association between the target entity and the scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data.

[0079] After obtaining the target entity, in this embodiment, it is necessary to associate the target entity with the scientific and technological intelligence database to obtain scientific and technological intelligence data. It should be emphasized that in this embodiment, a scientific and technological intelligence knowledge base of a domain-specific knowledge graph is constructed, and through entity association, all the scientific and technological intelligence data finally obtained are stored in the scientific and technological intelligence knowledge base.

[0080] This embodiment obtains initial scientific and technological intelligence data based on intelligence perception; performs data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data; identifies all entity reference items in the reference scientific and technological intelligence data; uses a generative artificial intelligence model to obtain candidate entities based on the entity reference items; determines a target entity based on the candidate entity; and constructs an association between the target entity and a scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data. Through the above-mentioned method, it is possible to cope with changes in the layout and structure of any website, effectively extract information from a variety of website structures, and improve the efficiency of obtaining scientific and technological intelligence data. At the same time, a generative artificial intelligence model is used to further ensure the accuracy and completeness of the scientific and technological intelligence data.

[0081] refer to Figure 2 , Figure 2 This is a flow chart of the second embodiment of a method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence according to the present invention.

[0082] Based on the above first embodiment, in the method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence in this embodiment, step S40 further includes:

[0083] Step S401: Obtain the relationship between entities corresponding to entity referents through a generative artificial intelligence model.

[0084] The entity reference term in this embodiment is a word or phrase used to refer to a specific entity in the text. An entity reference term may correspond to one or more entities. In this embodiment, the relationship between the entities corresponding to the entity reference terms of the already constructed generative artificial intelligence model can be utilized, and the relationship between the entities can be, for example, a classification relationship, a part-whole relationship, an attribute relationship, and a causal relationship.

[0085] Step S402: construct an adjacency matrix according to the relationships between the entities.

[0086] With each entity as a node and based on the relationship between the above entities, an adjacency matrix can be further constructed in this embodiment. For example, n entities (nodes) can constitute an n×n adjacency matrix.

[0087] Step S403: Determine the embedded representation of the candidate entity based on the adjacency matrix, the coefficient matrix of the generative artificial intelligence model, the bias vector and the preset activation function.

[0088] Based on the adjacency matrix constructed above, this embodiment further combines the coefficient matrix, bias vector and preset activation function of the generative artificial intelligence model to determine the embedded representation of the candidate entity, for example, h=α(AW+b), wherein α represents the preset activation function, and nonlinear activation functions such as ReLU or Softmax can be used, A represents the adjacency matrix, W represents the coefficient matrix of the generative artificial intelligence model, and b represents the bias vector.

[0089] Step S404: Acquire a candidate entity based on the embedded representation of the candidate entity.

[0090] After the embedded representation of the candidate entity is obtained according to the above calculation, the candidate entity can be obtained according to the embedded representation of the candidate entity. When obtaining the candidate entity, the contextual features of the entity referent also need to be obtained in this embodiment. The contextual features include lexical features, syntactic features, semantic features, and paragraph features. Then, the feature vector of the contextual features and the vector corresponding to the embedded representation of the candidate entity are constructed separately in the same dimension, and finally the candidate entity vector is obtained by vector splicing, for example, V=S⊕G, ⊕ means splicing in the same dimension, S is the feature vector of the contextual features, and G is the vector corresponding to the embedded representation of the candidate entity.

[0091] Furthermore, after obtaining the candidate entity, in this embodiment, the candidate entity is linked to a preset entry, which is an entry in a preset database, and each entry corresponds to an entity. After the link is completed, the contextual features of the candidate entity and the entity corresponding to each entry are compared, so that the candidate entity can match the entity corresponding to each entry. If they do not match, disambiguation can be performed. After completing entity disambiguation, the target entity in this embodiment can be determined from the disambiguated candidate entities. Specifically, a preset classifier can be used to determine the similarity between the disambiguated candidate entity and the predicted entity; a similarity score is performed based on the similarity; and the disambiguated candidate entity with the largest similarity score is used as the target entity. In this embodiment, a softmax classifier can be used, for example, sim(x,y)=softmax(x,y), where x and y represent the predicted entity and the disambiguated candidate entity, respectively.

[0092] This embodiment obtains the relationship between each entity corresponding to the entity referent through a generative artificial intelligence model; constructs an adjacency matrix according to the relationship between the entities; determines the embedded representation of the candidate entity according to the adjacency matrix, the coefficient matrix of the generative artificial intelligence model, the bias vector and the preset activation function, and obtains the contextual features of the entity referent; constructs the feature vector of the contextual feature and the vector corresponding to the embedded representation of the candidate entity in the same dimension respectively; performs vector concatenation on the feature vector of the contextual feature and the vector corresponding to the embedded representation of the candidate entity in the same dimension to obtain a candidate entity vector; obtains a candidate entity according to the candidate entity vector, and links the candidate entity with a preset entry; compares the contextual features between the candidate entity and the entity corresponding to the linked entry; disambiguates the candidate entity based on the comparison result; determines the similarity between the disambiguated candidate entity and the predicted entity using a preset classifier; performs a similarity score according to the similarity; and takes the disambiguated candidate entity with the largest similarity score as the target entity. By the above, accurate target entities can be obtained, and the accuracy and completeness of scientific and technological intelligence data can be further guaranteed.

[0093] Reference Figure 3 , Figure 3 This is a structural block diagram of the first embodiment of the scientific and technological intelligence perception and acquisition system based on generative artificial intelligence of the present invention.

[0094] like Figure 3 As shown, the scientific and technological intelligence perception and acquisition system based on generative artificial intelligence proposed in the embodiment of the present invention includes:

[0095] The acquisition module 10 is used to obtain initial scientific and technological intelligence data based on intelligence perception.

[0096] The cleaning module 20 is used to clean the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data.

[0097] The identification module 30 is used to identify all entity reference items in the reference scientific and technological intelligence data.

[0098] A construction module 40 is used to obtain candidate entities based on the entity referent using a generative artificial intelligence model.

[0099] The screening module 50 is used to determine the target entity according to the candidate entities.

[0100] The processing module 60 is used to establish an association between the target entity and the scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data.

[0101] This embodiment obtains initial scientific and technological intelligence data based on intelligence perception; performs data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data; identifies all entity reference items in the reference scientific and technological intelligence data; uses a generative artificial intelligence model to obtain candidate entities based on the entity reference items; determines a target entity based on the candidate entity; and constructs an association between the target entity and a scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data. Through the above-mentioned method, it is possible to cope with changes in the layout and structure of any website, effectively extract information from a variety of website structures, and improve the efficiency of obtaining scientific and technological intelligence data. At the same time, a generative artificial intelligence model is used to further ensure the accuracy and completeness of the scientific and technological intelligence data.

[0102] In one embodiment, the acquisition module 10 is also used to send an information extraction request to obtain the uniform resource locator of a journal page associated with scientific and technological intelligence; based on the uniform resource locator, the corresponding fields are extracted from the journal page, and it is determined whether there are any remaining uniform resource locators that have not been obtained; if not, all the extracted fields are formatted and stored in a database to obtain initial scientific and technological intelligence data.

[0103] In one embodiment, the cleaning module 20 is also used to normalize the initial scientific and technological intelligence data to obtain processed scientific and technological intelligence data; perform real word, homonym and short article misunderstanding detection and repair on the processed scientific and technological intelligence data; calculate the comprehensive accuracy score of the scientific and technological intelligence data after detection and repair; if the comprehensive accuracy score does not reach the preset score, return to execute the real word, homonym and short article misunderstanding detection and repair on the processed scientific and technological intelligence data; if the comprehensive accuracy score reaches the preset score, use the detected and repaired scientific and technological intelligence data as reference scientific and technological intelligence data.

[0104] In one embodiment, the construction module 40 is also used to obtain the relationship between each entity corresponding to the entity referent through a generative artificial intelligence model; construct an adjacency matrix based on the relationship between the entities; determine the embedded representation of the candidate entity based on the adjacency matrix, the coefficient matrix of the generative artificial intelligence model, the bias vector and the preset activation function; and obtain the candidate entity based on the embedded representation of the candidate entity.

[0105] In one embodiment, the construction module 40 is also used to obtain contextual features of the entity referent; construct the feature vector of the contextual features and the vector corresponding to the embedded representation of the candidate entity in the same dimension; perform vector concatenation on the feature vector of the contextual features and the vector corresponding to the embedded representation of the candidate entity in the same dimension to obtain a candidate entity vector; and obtain a candidate entity based on the candidate entity vector.

[0106] In one embodiment, the screening module 50 is further used to link the candidate entity with a preset entry; compare contextual features between the candidate entity and the entity corresponding to the linked entry; disambiguate the candidate entity based on the comparison result; and determine the target entity based on the disambiguated candidate entity.

[0107] In one embodiment, the screening module 50 is further used to determine the similarity between the disambiguated candidate entity and the predicted entity using a preset classifier; perform a similarity score based on the similarity; and use the disambiguated candidate entity with the largest similarity score as the target entity.

[0108] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive Solid State Disk (SSD)), etc.

[0109] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0110] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0111] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0112] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0113] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0114] In addition, for technical details that are not described in detail in this embodiment, please refer to the method for perceiving and acquiring scientific and technological intelligence based on generative artificial intelligence provided in any embodiment of the present invention, which will not be repeated here.

[0115] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0116] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0117] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

[0118] It is understandable that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and the explanation, examples and beneficial effects of the relevant contents can refer to the corresponding parts in the above method.

Claims

1. A method for perceiving and acquiring scientific and technological intelligence based on generative artificial intelligence, characterized in that: The method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence includes: Acquire initial scientific and technological intelligence data based on intelligence perception; Performing data cleaning on the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data; Identify all entity references in the referenced scientific and technological intelligence data; Obtaining a candidate entity based on the entity referent using a generative artificial intelligence model; Wherein, the obtaining of candidate entities based on the entity referent using a generative artificial intelligence model further comprises: Obtain the relationship between entities corresponding to entity referents through a generative artificial intelligence model; Constructing an adjacency matrix according to the relationship between the entities, wherein the adjacency matrix A is n×n, and n represents an entity; Determine the embedding representation of the candidate entity according to the adjacency matrix, the coefficient matrix of the generative artificial intelligence model, the bias vector and the preset activation function, wherein the embedding representation of the candidate entity is h=α(AW+b), α represents the preset activation function, adopts ReLU or Softmax nonlinear activation function, A represents the adjacency matrix, W represents the coefficient matrix of the generative artificial intelligence model, and b represents the bias vector; Obtaining contextual features of entity referents; Constructing a feature vector of the context feature and a vector corresponding to the embedding representation of the candidate entity in the same dimension respectively; Concatenate the feature vector of the context feature in the same dimension with the vector corresponding to the embedding representation of the candidate entity to obtain a candidate entity vector, wherein the candidate entity vector V=S⊕G, where ⊕ means concatenation in the same dimension, S is the feature vector of the context feature, and G is the vector corresponding to the embedding representation of the candidate entity; Acquire a candidate entity according to the candidate entity vector; Determine a target entity according to the candidate entity; An association is established between the target entity and a scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data.

2. The method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence according to claim 1, characterized in that: The acquisition of initial scientific and technological intelligence data based on intelligence perception includes: Send an information extraction request to obtain the Uniform Resource Locator of the journal page associated with scientific and technological information; Extracting corresponding fields from the journal page based on the uniform resource locator, and determining whether there are any remaining unobtained uniform resource locators; If it does not exist, all the extracted fields are formatted and stored in the database to obtain the initial scientific and technological intelligence data.

3. The method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence according to claim 1, characterized in that: The data cleaning of the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data includes: Normalizing the initial scientific and technological intelligence data to obtain processed scientific and technological intelligence data; Detecting and repairing real words, homophones and short text misunderstandings on the processed scientific and technological intelligence data; Calculate the comprehensive accuracy score of the detected and repaired S&T intelligence data; If the comprehensive accuracy score does not reach the preset score, returning to execute the detection and repair of true words, homophones and short text misunderstandings on the processed scientific and technological intelligence data; If the comprehensive accuracy score reaches the preset score, the scientific and technological intelligence data after detection and repair will be used as reference scientific and technological intelligence data.

4. The method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence according to claim 1, characterized in that: The determining the target entity according to the candidate entity comprises: Linking the candidate entity with a preset entry; Compare the context features between the candidate entity and the corresponding entity of the linked entry; disambiguating the candidate entity based on the comparison result; Determine the target entity based on the disambiguated candidate entities.

5. The method for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence according to claim 4, characterized in that: The determining the target entity according to the disambiguated candidate entities includes: Determine the similarity between the disambiguated candidate entity and the predicted entity using a preset classifier; performing a similarity score according to the similarity; The disambiguated candidate entity with the largest similarity score is taken as the target entity.

6. A scientific and technological intelligence perception and acquisition system based on generative artificial intelligence, characterized in that: The scientific and technological intelligence perception and acquisition system based on generative artificial intelligence: The acquisition module is used to obtain initial scientific and technological intelligence data based on intelligence perception; A cleaning module, used for cleaning the initial scientific and technological intelligence data to obtain reference scientific and technological intelligence data; An identification module, used to identify all entity reference items in the reference scientific and technological intelligence data; A construction module is used to obtain candidate entities based on the entity referent using a generative artificial intelligence model, wherein the obtaining candidate entities based on the entity referent using a generative artificial intelligence model further includes: Obtain the relationship between entities corresponding to entity referents through a generative artificial intelligence model; Constructing an adjacency matrix according to the relationship between the entities, wherein the adjacency matrix A is n×n, and n represents an entity; Determine the embedding representation of the candidate entity according to the adjacency matrix, the coefficient matrix of the generative artificial intelligence model, the bias vector and the preset activation function, wherein the embedding representation of the candidate entity is h=α(AW+b), α represents the preset activation function, adopts ReLU or Softmax nonlinear activation function, A represents the adjacency matrix, W represents the coefficient matrix of the generative artificial intelligence model, and b represents the bias vector; Obtaining contextual features of entity referents; Constructing a feature vector of the context feature and a vector corresponding to the embedding representation of the candidate entity in the same dimension respectively; Concatenate the feature vector of the context feature in the same dimension with the vector corresponding to the embedding representation of the candidate entity to obtain a candidate entity vector, wherein the candidate entity vector V=S⊕G, where ⊕ means concatenation in the same dimension, S is the feature vector of the context feature, and G is the vector corresponding to the embedding representation of the candidate entity; Acquire a candidate entity according to the candidate entity vector; A screening module, used to determine a target entity according to the candidate entity; The processing module is used to establish an association between the target entity and the scientific and technological intelligence knowledge base to obtain target scientific and technological intelligence data.

7. The system for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence according to claim 6, characterized in that: The acquisition module is also used to send an information extraction request to obtain the uniform resource locator of the journal page associated with the scientific and technological information; Extracting corresponding fields from the journal page based on the uniform resource locator, and determining whether there are any remaining unobtained uniform resource locators; If it does not exist, all the extracted fields are formatted and stored in the database to obtain the initial scientific and technological intelligence data.

8. The system for sensing and acquiring scientific and technological intelligence based on generative artificial intelligence as claimed in claim 6, characterized in that: The cleaning module is also used to normalize the initial scientific and technological intelligence data to obtain processed scientific and technological intelligence data; Detecting and repairing real words, homophones and short text misunderstandings on the processed scientific and technological intelligence data; Calculate the comprehensive accuracy score of the detected and repaired S&T intelligence data; If the comprehensive accuracy score does not reach the preset score, returning to execute the detection and repair of true words, homophones and short text misunderstandings on the processed scientific and technological intelligence data; If the comprehensive accuracy score reaches the preset score, the scientific and technological intelligence data after detection and repair will be used as reference scientific and technological intelligence data.

Images