Selected question recommendation method and system, electronic equipment and computer readable storage medium

Abstract

The invention belongs to the field of artificial intelligence and knowledge engineering, particularly relates to a topic selection recommendation method and system, electronic equipment and a computer readable storage medium, and aims to solve the problem that academic topic selection is low in accuracy and practicability. The knowledge graph is dynamically updated according to new academic data, so that the knowledge graph is dynamically evolved, candidate selected topics generated based on the knowledge graph can follow up the latest academic dynamics, the real-time performance and accuracy of the selected topics are improved, the generated candidate selected topics are verified and scored at least based on the knowledge graph, the reasonability of the selected topics is improved, and the accuracy of the selected topics is improved. And the accuracy and practicability of academic topic selection are further improved.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence and knowledge engineering, and specifically relates to a topic recommendation method, system, computer equipment, computer-readable storage medium and computer program product. Background Art

[0002] In academic research, the topic determines the direction of research and, to a large extent, its value. Good academic topic selection requires a sufficient understanding of existing achievements and discussions in the academic field. This often requires searching and refining from massive amounts of academic materials, which is time-consuming and labor-intensive, and lacks a clear value assessment. To improve the efficiency of academic topic selection, current academic topic selection is mainly achieved through knowledge graphs and large language models. However, existing knowledge graph-based solutions use static knowledge graphs, which cannot timely reflect the dynamic evolution of the academic field. They also rely too much on manual or basic text processing tools to extract entities and relationships, making it difficult to process multimodal academic data. Existing large language model-based solutions, while having advantages in processing multimodal academic data, cannot evaluate and verify topic selection from the perspective of the academic knowledge system as a whole. Moreover, both knowledge graph-based and large language model-based solutions lack user feedback mechanisms, making it difficult to continuously improve the accuracy and practicality of topic selection.

[0003] To sum up, existing technologies have deficiencies in the dynamic evolution mechanism of topic selection, evaluation verification, and feedback optimization system. How to overcome these deficiencies and improve the accuracy and practicality of academic topic selection is an urgent problem that needs to be solved. Summary of the Invention

[0004] In order to solve the above-mentioned problem in the prior art, namely, the problem of low accuracy and practicality of academic topic selection, the first aspect of the present application provides a topic recommendation method, the method comprising:

[0005] Extracting entities and semantic relationships related to the entities in academic data, wherein the entities include at least research questions and research methods;

[0006] Constructing a knowledge graph based on the entities and the semantic relationships;

[0007] If new academic data is detected, updating the knowledge graph according to the new academic data;

[0008] Generate multiple candidate topics based on the knowledge graph;

[0009] Calculate the score of each of the multiple candidate topics, where the score includes at least a rationality score based on the knowledge graph, and use topics with scores greater than a preset threshold as recommended topics.

[0010] Optionally, the method further includes: obtaining user feedback information on the recommended topics; adjusting the scores of the respective topics according to the feedback information; and selecting topics whose adjusted scores are greater than a preset threshold as recommended topics.

[0011] Optionally, the calculation of the rationality score of each of the multiple candidate topics includes: calculating the structural hole index of the entity corresponding to each topic based on the knowledge graph; and calculating the interdisciplinary connectivity of the entity corresponding to each topic based on the knowledge graph; and obtaining the rationality score according to the structural hole index and the interdisciplinary connectivity.

[0012] Optionally, the calculation of the rationality score of each of the multiple candidate topics includes: calculating the structural hole index of the entity corresponding to each topic based on the knowledge graph; and, calculating the interdisciplinary connectivity of the entity corresponding to each topic based on the knowledge graph; and, calculating the node influence of the entity corresponding to each topic based on the knowledge graph, and obtaining the rationality score according to the structural hole index, the interdisciplinary connectivity and the node influence.

[0013] Optionally, the calculating of the scores of each of the multiple candidate topics further includes: calculating the innovation score and / or feasibility score of the topic to be recommended, the topic to be recommended is a candidate topic whose rationality score is greater than a first preset threshold; obtaining a comprehensive score of the topic to be recommended based on the innovation score and / or the feasibility score; and recommending the candidate topic whose score is greater than the preset threshold as a recommended topic means recommending the topic to be recommended whose comprehensive score is greater than a second preset threshold as a recommended topic.

[0014] Optionally, the innovation score is an expert score obtained based on preset innovation standards, and the feasibility score is an expert score obtained based on preset feasibility standards.

[0015] Optionally, the method further includes: obtaining user feedback information on the recommended topic; and adjusting the structure and / or parameters of the knowledge graph based on the feedback information.

[0016] Optionally, adjusting the structure and parameters of the knowledge graph according to the feedback information includes: adjusting the attributes and relationships of the research problem entities according to the feedback information; and / or adding new connecting edges in the knowledge graph according to new semantic relationships in the feedback information.

[0017] Optionally, if new academic data is detected, updating the knowledge graph according to the new academic data includes: updating the time-varying edge weights between nodes in the knowledge graph:

[0018] ω ij (t) = f(Ii I j )·ɡ(t)·h(d i d j ) Where t is the time decay factor, I i and I j are the academic influence of nodes i and j in the knowledge graph, respectively. Node i is adjacent to node j. f is the academic influence coupling function, h is the node degree adjustment function, and d i and d j are the node degrees of nodes i and j respectively, λ is the preset subject area coefficient, and t0 represents the initial establishment time of the relationship edge between node i and node j.

[0019] In a second aspect of the present application, a topic recommendation system is provided, comprising:

[0020] An academic data processing module, configured to extract entities and semantic relationships related to the entities from the academic data, wherein the entities include at least research questions and research methods;

[0021] A knowledge graph construction module, configured to construct a knowledge graph based on the entities and the semantic relationships;

[0022] A knowledge graph updating module, configured to update the knowledge graph according to new academic data if new academic data is detected;

[0023] A candidate topic generation module, configured to generate multiple candidate topics based on the knowledge graph;

[0024] The recommendation module is used to calculate the score of each topic among the multiple candidate topics, where the score at least includes a rationality score based on the knowledge graph, and to recommend topics with scores greater than a preset threshold.

[0025] Optionally, the system further includes:

[0026] A feedback optimization module, used to obtain user feedback information on the recommended topic;

[0027] The recommendation module is further configured to adjust the scores of the respective topics according to the feedback information, and to use topics with adjusted scores greater than a preset threshold as recommended topics.

[0028] Optionally, the system further includes:

[0029] A feedback optimization module, used to obtain user feedback information on the recommended topic;

[0030] The knowledge graph updating module is also used to adjust the structure and / or parameters of the knowledge graph according to the feedback information.

[0031] According to a third aspect of the present application, an electronic device is provided, comprising:

[0032] At least one processor; and a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to implement the above-mentioned topic recommendation method.

[0033] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned topic recommendation method.

[0034] In a fifth aspect of the present application, a computer program product comprising instructions, when the instructions are executed by a computer device, causes the computer device to execute the above-mentioned topic recommendation method.

[0035] This application extracts entities and semantic relationships from academic data to construct a knowledge graph, and dynamically updates the knowledge graph based on new academic data, so that the knowledge graph can achieve dynamic evolution, so that the candidate topics generated based on the knowledge graph can keep up with the latest academic trends, improve the real-time and accuracy of the topics, and verify and score the generated candidate topics at least based on the knowledge graph to improve the rationality of the topics, thereby further improving the accuracy and practicality of academic topics. Moreover, further, this application can also evaluate the innovation and feasibility of the candidate topics, and improve the accuracy, practicality and value of the topics in more dimensions. In addition, further, this application can also adjust the topic evaluation and knowledge graph based on user feedback information, thereby improving the accuracy and quality of the topics to a greater extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Other features, objects and advantages of the present application will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0037] Figure 1 This is a flowchart of one implementation method of the application topic recommendation method;

[0038] Figure 2 This is a flowchart of another implementation method of the application topic recommendation method;

[0039] Figure 3 This is a flowchart of another implementation method of the application topic recommendation method;

[0040] Figure 4 This is a structural diagram of an implementation method of the topic recommendation system for this application;

[0041] Figure 5This is a structural diagram of another implementation of the application topic recommendation system;

[0042] Figure 6 It is a structural diagram of a computer system of a server for implementing the method, system, and device embodiments of the present application. DETAILED DESCRIPTION

[0043] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the relevant invention are shown in the accompanying drawings.

[0044] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this 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.

[0045] This application provides a topic recommendation method, such as Figure 1 As shown, the method includes:

[0046] Step S101, extracting entities and semantic relationships related to the entities in academic data, wherein the entities at least include research questions and research methods;

[0047] Step S102: constructing a knowledge graph based on the entities and the semantic relationships;

[0048] Step S103: if new academic data is detected, update the knowledge graph according to the new academic data;

[0049] Step S104: generating multiple candidate topics based on the knowledge graph;

[0050] Step S105: Calculate the score of each of the multiple candidate topics, where the score at least includes a rationality score based on the knowledge graph, and recommend topics with scores greater than a preset threshold.

[0051] Specifically, in one embodiment, step S101 can use a large language model to process academic data. The academic data can be multimodal data. If the academic field is the field of natural science, such as computer science related technology, the academic data can include the full text of relevant technology papers in top conferences in a certain period of time (for example, the past 5 years), open source experimental code of popular network protocol implementation projects of related technologies, patent documents of related technologies, etc.; if the academic field is the field of social science, such as education, the academic data can include national education policy documents, social survey data of compulsory education quality inspection in a certain region, well-known education reform case study reports, etc. The collected academic data can be classified and stored and pre-processed such as deduplication and word segmentation. Then, a large language model, such as the Huazhi large model, is used to extract the entities in the academic data and the semantic relationships related to the entities. Specifically, a prompt template containing entity type constraints, relationship extraction domain rules, and data quality verification requirements can be constructed to extract more accurate, clear and stable results. The extracted entity results include at least research questions and research methods, which are the two elements that need to be included in the academic topic of this application. For example, for language processing technology in computer science, the extracted entities include the research method "Transformer algorithm" and the research question "natural language processing task", and the academic topic can be "Research on natural language processing tasks based on Transformer algorithm". The above is only an example to illustrate that the extracted entities should at least include the results, and it is not a limitation of the method steps of this application. Specifically, in one embodiment, the extracted entities can also include scholars, etc. The extracted results can be output as structured data for the subsequent construction of a knowledge graph.

[0052] Specifically, step S102 constructs a graph based on the extracted entities and semantic relationships. For example, Neo4j can be used as a knowledge graph construction tool to define three core entity types: research questions, research methods, and scholars, and establish corresponding attribute fields. For example, the research question entity contains attributes such as the question name, the subject to which it belongs, and the research popularity; the research method entity contains attributes such as the tool name, technical principles, and applicable scenarios; and the scholar entity contains attributes such as name, affiliated institution, and research direction. At the same time, define the relationship types between entities, such as "use" (scholars use research methods to solve research problems) and "propose" (scholars propose research questions).

[0053] Furthermore, in order to enable the constructed knowledge graph to dynamically evolve with new academic data, step S103 updates the constructed knowledge graph according to the new academic data after detecting the new academic data. Specifically, the administrator can collect and store the relevant academic data based on new academic trends, such as the release of new academic achievements, and use this to update the relevant indicators of the corresponding entities in the knowledge graph, such as the academic influence of the entity (node); new entities and semantic relationships can also be extracted and added to the knowledge graph based on the collected new academic data through a large language model. In one embodiment, the administrator collects new academic data and inputs it into the storage area in a predetermined format. The administrator can trigger or input a completion action trigger (triggering data detection or directly triggering the knowledge graph update), or the system periodically scans the storage area. When the scan detects the presence of new academic data, the administrator calculates and updates the relevant indicators of the corresponding entities in the knowledge graph based on the new academic data, or extracts new entities and semantic relationships and adds them to the knowledge graph.

[0054] Specifically, in one embodiment, the system scans the aforementioned storage area, and when new academic data is detected, or the system scans the knowledge graph, and when data changes are detected in the knowledge graph, the updating of the knowledge graph includes: updating the time-varying edge weights W between the nodes (i.e., entities) in the knowledge graph. ij (t):

[0055] W ij (t) = f(I i I j )·ɡ(t)·h(d i d j ),

[0056] Where t is the time decay factor, I i and I j are the academic influence of nodes i and j in the knowledge graph, respectively. Node i is adjacent to node j. f is the academic influence coupling function, h is the node degree adjustment function, and d i and d j are the node degrees of nodes i and j respectively (i.e. the number of other directly connected nodes), λ is a preset subject area coefficient (for example, λ = 0.02 for basic disciplines and λ = 0.05 for applied disciplines), and t0 represents the initial establishment time of the relationship edge between node i and node j. Among them, the academic influence of the node (i.e., entity) can be obtained by calculating the weighted sum after normalization of the number of literature citations, H-index, alternative metrological indicators, etc., and the academic influence coupling function and the node degree adjustment function can adopt existing functions. This implementation method reflects the dynamic changes in the influence and relevance of knowledge in the time dimension when new academic data is generated, that is, when academic research changes, so that the knowledge graph can evolve naturally and dynamically, thereby making the generated topics more in line with the actual academic development and improving the timeliness and accuracy of topic recommendations.

[0057] Through step S103, the traditional static knowledge graph is made dynamic, and the structure and node relationships of the knowledge graph can be adjusted according to academic trends, making the knowledge graph more scientific, thereby improving the accuracy and foresight of topic recommendations based on the knowledge graph.

[0058] Step S104 generates multiple candidate topics based on the constructed knowledge graph. Specifically, in one embodiment, the entity and relationship information in the knowledge graph is integrated and converted into the format required for input to the large language model. The large language model is used to generate candidate topics. The model can be set to generate multiple candidate topics to increase the diversity of topics.

[0059] For the multiple candidate topics generated, step S105 evaluates the candidate topics, calculates their scores, and recommends topics based on the scores. The evaluation at least includes rationality verification based on the source knowledge graph, so the scoring at least includes rationality scoring based on the knowledge graph. Specifically, in one embodiment, the rationality score is calculated as follows:

[0060] Score i =a*SH i +b*CD i ,

[0061] Among them, SH i is the structural hole index, SH i =1-∑ j∈N(i) (w ij +w ji ) 2 , where w ij is the edge weight between node i and node j, taking the current value of the aforementioned time-varying edge weight, N(i) is the set of neighbors of node i, and node i is the entity i corresponding to the candidate topic;

[0062] CD i For interdisciplinary connectivity degrees, Among them, ||D i|| represents the number of disciplines associated with node i, and max(D) is the maximum discipline span in the system;

[0063] a and b are preset weight coefficients, a + b = 1. For each candidate topic, calculate the rationality score of each entity i corresponding to the candidate topic, i∈[1,m], m is the number of entities corresponding to the candidate topic, and get the score of each entity i i Then, calculate the Score i The sum or average value is used as the rationality score of the candidate topic.

[0064] Specifically, in one embodiment, the rationality score is calculated as follows: Score i =a*SH i +b*CD i +c*I i ,

[0065] Among them, SH i is the structural hole index, CD i For interdisciplinary connectivity, I i is the node influence, Among them, Citations i is the number of citations, max(Citations) is the maximum number of citations in the knowledge graph, SocialHeat i is the social heat, and α is the preset social heat coefficient;

[0066] a, b and c are preset weight coefficients, a+b+c=1. Similarly, the score of each entity i corresponding to the candidate topic is obtained. i Then, calculate the Score i The sum or average of the results is used as the rationality score of the candidate topic. This implementation method further uses influence as a factor in rationality verification, improving the rigor and rationality of topic verification.

[0067] After obtaining the rationality score of each candidate topic, the topics with rationality scores greater than a preset threshold are selected as recommended topics to complete the topic recommendation.

[0068] Specifically, in one embodiment, step S105 also includes: taking the topics whose rationality scores are greater than a preset first threshold as topics to be recommended, and further calculating the innovation scores of the topics to be recommended, or calculating the feasibility scores of the topics to be recommended, or calculating the innovation scores and feasibility scores of the topics to be recommended, and then taking the innovation scores, or the feasibility scores, or the weighted sum of the innovation scores and the feasibility scores as the comprehensive scores, and then taking the topics to be recommended whose comprehensive scores are greater than a second preset threshold as recommended topics to complete the topic recommendation.

[0069] Specifically, for each recommended question, the innovation score is calculated as follows:

[0070] Innovation k =a1·SH k +b1·CD k , where k is the serial number of the question to be recommended, Innovation k Score the innovation of the selected question k; SH k is the structural hole index of the topic k to be recommended. The structural hole index of topic k is the weighted sum or average of the structural hole indexes of its corresponding entities; CD k is the interdisciplinary connectivity of the topic k to be recommended. The interdisciplinary connectivity of topic k is the weighted sum or average of the structural hole indexes of its corresponding entities. a1 and b1 are preset weight coefficients, and a1+b1=1.

[0071] Specifically, for each recommended question, the feasibility score is calculated as follows: FinalScore k =a2·Resource k +b2·Technical k +c2 Temporal k ,in, Innovation k Score the feasibility of the recommended question k; Resource k is the resource matching degree of the recommended question k, Technical k is the technical maturity of the recommended question k, TRL is Technology Readiness Level; Temporal k Temporal feasibility of the selected question k. k =e -0.1·估计月数 , where the estimated number of months can be obtained from historical project statistics; a2, b2 and c2 are preset weight coefficients, a1+b1+C1=1.

[0072] In one embodiment, the innovation score may be obtained from experts according to a preset innovation standard, and the feasibility score may be obtained from experts according to a preset feasibility standard.

[0073] After obtaining the innovation score or feasibility score of the topic to be recommended, the innovation score or feasibility score can be used as the comprehensive score of the topic to be recommended. After simultaneously calculating the innovation score and feasibility score of the topic to be recommended, the weighted sum of the innovation score and the feasibility score can be further calculated as the comprehensive score of the topic to be recommended, wherein the innovation score and the feasibility score can be normalized first and then weighted summed to obtain the comprehensive score. The topics to be recommended whose comprehensive scores are greater than the preset second threshold are then used as recommended topics to complete the topic recommendation. This implementation method first screens the candidate topics according to the rationality score for the first time, and takes the candidate topics that meet the rationality verification (rationality score is greater than the preset first threshold) as the topics to be recommended, and then performs an innovation and / or feasibility evaluation on the topics to be recommended, and performs a second screening according to the innovation score and / or feasibility score. Compared with screening and recommendation based only on the rationality score, combining the innovation score and / or feasibility score improves the accuracy, practicality and value of the topic recommendation to a greater extent.

[0074] The topic recommendation method provided in this application overcomes the defects of the existing topic recommendation knowledge graph being static and lacking verification of topic rationality, and greatly improves the accuracy of topic recommendation.

[0075] In one embodiment, Figure 2 As shown, the topic recommendation method provided by this application includes:

[0076] Step S201, extracting entities and semantic relationships related to the entities in academic data, wherein the entities at least include research questions and research methods;

[0077] Step S202: constructing a knowledge graph based on the entities and the semantic relationships;

[0078] Step S203: if new academic data is detected, update the knowledge graph according to the new academic data;

[0079] Step S204: generating multiple candidate topics based on the knowledge graph;

[0080] Step S205: Calculate the score of each of the multiple candidate topics, where the score at least includes a rationality score based on the knowledge graph, and recommend topics with scores greater than a preset threshold.

[0081] Step S206 , obtaining user feedback information on the recommended topics; adjusting the scores of the respective topics according to the feedback information; and selecting topics whose adjusted scores are greater than a preset threshold as recommended topics.

[0082] Steps S201-205 may refer to the aforementioned steps S101-105. Step S206 further adjusts the score calculation of each of the aforementioned topics based on user feedback on the recommended topics, and selects topics with adjusted scores greater than a preset threshold as recommended topics, thereby optimizing the feedback of the recommended topics and further improving the accuracy of the topic recommendations.

[0083] Specifically, a user feedback collection system could be developed, including a web-based feedback portal and API interface integration. The web-based feedback portal provides a user interface where users can accept or reject recommended topics and provide information such as reasons for acceptance and suggestions for improvement. Through API integration with academic social platforms and paper management systems, the system automatically collects data such as changes in the collaborative network between scholars resulting from topic selection, as well as citations of related topics in papers. User feedback can be analyzed to determine user preferences and needs for recommended topics. For example, if a user prefers interdisciplinary topics, the weight of interdisciplinary connectivity in the candidate topic scoring calculation can be adjusted to increase the proportion of recommended interdisciplinary topics. If a user reports that some recommended topics are less feasible, the weight of the feasibility score can be increased. By adjusting the topic scoring calculation in this way, the relevant calculation parameter weights can be adjusted, and the topic scores can be recalculated. After obtaining the adjusted scores, the topics can be screened and recommended based on the new scores. This continuously optimizes the topic recommendation strategy, improving the accuracy and practicality of topic recommendations.

[0084] In one embodiment, Figure 3 As shown, the topic recommendation method provided by this application includes:

[0085] Step S301, extracting entities and semantic relationships related to the entities in academic data, wherein the entities at least include research questions and research methods;

[0086] Step S302: constructing a knowledge graph based on the entities and the semantic relationships;

[0087] Step S303: if new academic data is detected, update the knowledge graph according to the new academic data;

[0088] Step S304: generating multiple candidate topics based on the knowledge graph;

[0089] Step S305: Calculate the score of each of the multiple candidate topics, where the score at least includes a rationality score based on the knowledge graph, and recommend topics with scores greater than a preset threshold.

[0090] Step S306 , obtaining user feedback information on the recommended topics; adjusting the scores of the respective topics according to the feedback information; and selecting topics whose adjusted scores are greater than a preset threshold as recommended topics.

[0091] Step S307: Obtain user feedback information on the recommended topic; and adjust the structure and / or parameters of the knowledge graph according to the feedback information.

[0092] Among them, step S301306 can refer to the aforementioned steps S201-206. In this embodiment, not only is the score calculation adjusted according to user feedback in step S306, but the structure and / or parameters of the knowledge graph are also adjusted according to the feedback information in step S307. Specifically, the attributes and relationships of the research problem entity can be adjusted according to the user feedback information; and / or, new connection edges are added to the knowledge graph based on the new semantic relationships in the user feedback information. For example, if the user feedback shows that the research question involved in a recommended topic is inaccurate, the attributes and relationships of the research problem entity in the knowledge graph are updated; if a new scholar cooperation relationship is found (for example, obtained through the API interface of the user feedback collection system and the academic social platform), the corresponding connection edges are added to the knowledge graph. The structure and parameters of the knowledge graph are continuously adjusted and optimized through feedback to make it more in line with the actual situation of academic research, thereby further improving the accuracy and value of the recommended topics generated by the knowledge graph.

[0093] Although the various steps in the above embodiment are described in the above-mentioned order, those skilled in the art will understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of the present invention.

[0094] The second aspect of this application is Figure 4 As shown, a topic recommendation system is provided, the system comprising:

[0095] An academic data processing module, configured to extract entities and semantic relationships related to the entities from the academic data, wherein the entities include at least research questions and research methods;

[0096] A knowledge graph construction module, configured to construct a knowledge graph based on the entities and the semantic relationships;

[0097] A knowledge graph updating module, configured to update the knowledge graph according to new academic data if new academic data is detected;

[0098] A candidate topic generation module, configured to generate multiple candidate topics based on the knowledge graph;

[0099] The recommendation module is used to calculate the score of each topic among the multiple candidate topics, where the score at least includes a rationality score based on the knowledge graph, and to recommend topics with scores greater than a preset threshold.

[0100] In one embodiment, Figure 5 As shown, the topic recommendation system provided by the present application also includes a feedback optimization module, which is used to obtain user feedback information on the recommended topics; the recommendation module is also used to adjust the scores of each topic according to the feedback information, and use the topics with adjusted scores greater than a preset threshold as recommended topics.

[0101] In one embodiment, the feedback optimization module is used to obtain user feedback information on the recommended topic; the knowledge graph update module is also used to adjust the structure and / or parameters of the knowledge graph according to the feedback information.

[0102] The topic recommendation system provided by this application enables the knowledge graph to dynamically evolve with research development, and can fully verify and evaluate topics, greatly improving the accuracy of topic recommendations. Furthermore, the improved topic recommendation system of this application can also optimize the knowledge graph and the verification and evaluation of topics based on user feedback, further improving the accuracy and practicality of topic recommendations.

[0103] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the system described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0104] It should be noted that the topic recommendation system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present application can be decomposed or combined. For example, the modules in the above embodiments can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present application are only for distinguishing the modules or steps and are not considered to be improper limitations on the present application.

[0105] According to a third aspect of the present application, an electronic device is provided, comprising:

[0106] At least one processor; and a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to implement the above-mentioned topic recommendation method.

[0107] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned topic recommendation method.

[0108] In a fifth aspect of the present application, a computer program product comprising instructions, when the instructions are executed by a computer device, causes the computer device to execute the above-mentioned topic recommendation method.

[0109] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes and related instructions of the storage device and processing device described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] Those skilled in the art should be able to appreciate that, in conjunction with the modules and method steps of each example described in the embodiments disclosed herein, it is possible to implement them with electronic hardware, computer software, or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the art. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0111] Reference below Figure 6 , which shows a structural diagram of a computer system of a server for implementing the method, system, and device embodiments of the present application. Figure 6 The server shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0112] like Figure 6 As shown, the computer system includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 602 or the program loaded from the storage part 608 to the random access memory (RAM) 603. Various programs and data required for system operation are also stored in the RAM 603. The CPU 601, ROM 602 and RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0113] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, and the like; an output section 607 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 608 including a hard disk and the like; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. Removable media 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like, are installed in the drive 610 as needed so that computer programs read therefrom can be installed into the storage section 608 as needed.

[0114] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611. When the computer program is executed by the central processing unit (CPU) 601, the above-mentioned functions defined in the method of the present application are executed. It should be noted that the computer-readable medium mentioned above in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code embodied on a computer-readable medium may be transmitted using any suitable medium, including, but not limited to, wireless, wire, optical cable, RF, etc., or any suitable combination thereof.

[0115] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0116] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0117] The terms "first", "second", etc. are used to distinguish similar objects, rather than to describe or indicate a particular order or sequence.

[0118] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0119] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. A topic recommendation method, characterized in that: include: Extracting entities and semantic relationships related to the entities in academic data, wherein the entities include at least research questions and research methods; Constructing a knowledge graph based on the entities and the semantic relationships; If new academic data is detected, updating the knowledge graph according to the new academic data; Generate multiple candidate topics based on the knowledge graph; Calculate the score of each of the multiple candidate topics, where the score includes at least a rationality score based on the knowledge graph, and use topics with scores greater than a preset threshold as recommended topics.

2. The topic recommendation method according to claim 1, characterized in that: Also includes: Obtaining user feedback on the recommended topic; Adjusting the score of each of the selected topics according to the feedback information; Topics whose adjusted scores are greater than the preset threshold will be recommended.

3. The topic recommendation method according to claim 1 or 2, characterized in that: Calculating the rationality score of each of the multiple candidate topics includes: Calculating the structural hole index of the entity corresponding to each topic based on the knowledge graph; and Calculate the interdisciplinary connectivity of entities corresponding to each topic based on the knowledge graph; The rationality score is obtained according to the structural hole index and the interdisciplinary connectivity.

4. The topic recommendation method according to claim 3, characterized in that: Calculating the score of each of the plurality of candidate topics further includes: Calculating the innovation score and / or feasibility score of the candidate topic to be recommended, wherein the candidate topic to be recommended is the candidate topic whose rationality score is greater than the first preset threshold; Obtaining a comprehensive score for the recommended topic based on the innovation score and / or the feasibility score; The method of taking candidate topics with scores greater than a preset threshold as recommended topics is to take the candidate topics with comprehensive scores greater than a second preset threshold as recommended topics.

5. The topic recommendation method according to claim 1 or 2, characterized in that: Also includes: Obtaining user feedback on the recommended topic; Adjust the structure and / or parameters of the knowledge graph according to the feedback information; The adjusting the structure and parameters of the knowledge graph according to the feedback information includes: Adjusting the attributes and relationships of the research problem entity according to the feedback information; and / or, According to the new semantic relationship in the feedback information, new connection edges are added to the knowledge graph.

6. The topic recommendation method according to claim 1, characterized in that: If new academic data is detected, updating the knowledge graph according to the new academic data includes: Update the time-varying edge weights between nodes in the knowledge graph: ω ij (t)=f(I i I j )·ɡ(t)·h(d i d j ) Where t is the time decay factor, I i and I j are the academic influence of nodes i and j in the knowledge graph, respectively. Node i is adjacent to node j. f is the academic influence coupling function, h is the node degree adjustment function, and d i and d j are the node degrees of nodes i and j respectively, λ is the preset subject area coefficient, and t0 represents the initial establishment time of the relationship edge between node i and node j.

7. A topic recommendation system, characterized in that: include: An academic data processing module, configured to extract entities and semantic relationships related to the entities from the academic data, wherein the entities include at least research questions and research methods; A knowledge graph construction module, configured to construct a knowledge graph based on the entities and the semantic relationships; A knowledge graph updating module, configured to update the knowledge graph according to new academic data if new academic data is detected; A candidate topic generation module, configured to generate multiple candidate topics based on the knowledge graph; The recommendation module is used to calculate the score of each topic among the multiple candidate topics, where the score at least includes a rationality score based on the knowledge graph, and to recommend topics with scores greater than a preset threshold.

8. The topic recommendation system according to claim 7, characterized in that: Also includes: A feedback optimization module, used to obtain user feedback information on the recommended topic; The recommendation module is further configured to adjust the scores of the respective topics according to the feedback information, and to recommend topics whose adjusted scores are greater than a preset threshold. The knowledge graph updating module is also used to adjust the structure and / or parameters of the knowledge graph according to the feedback information.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the topic recommendation method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the topic recommendation method according to any one of claims 1 to 6.

Images