Information processing method, electronic device and computer storage medium

Abstract

An information processing method, an electronic device and a computer storage medium. The method comprises: acquiring subject question information input by a user (S202); parsing the subject question information to determine universal knowledge point information involved in a solving process for the subject question information; and determining from a knowledge point content library universal knowledge point content matching the universal knowledge point information (S204); performing question solving on the subject question information to obtain targeted answer information corresponding to the subject question information (S206); and for the subject question information, displaying the targeted answer information and the universal knowledge point content (S208).

Description

Information processing methods, electronic devices and computer storage media Cross-reference to related applications

[0001] This patent application claims priority to Chinese Patent Application No. 202510474621.1, filed on April 16, 2025, entitled "Information Processing Method, Electronic Device and Computer Storage Medium", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This disclosure relates to the field of computer technology, and in particular to an information processing method, electronic device, computer storage medium, and computer program product. Background Technology

[0003] Currently, many applications on the market offer question search functionality. For example, users can enter a question to search, and the application can output corresponding answers based on the question. However, the answers currently provided by these applications are usually specific to the question, making it difficult to help users comprehensively grasp the learning content related to the question, thus failing to meet users' learning needs. Summary of the Invention

[0004] In view of this, the present disclosure provides an information processing solution to at least partially solve the above problems.

[0005] According to a first aspect of the present disclosure, an information processing method is provided, comprising: acquiring subject question information for a preset subject input by a user; parsing the subject question information to determine universal test point information involved in the solution process of the subject question information; determining universal test point content matching the universal test point information from a pre-generated test point content library or a searched test point content library; solving the subject question information to obtain targeted answer information corresponding to the subject question information; and displaying the targeted answer information and the universal test point content for the subject question information.

[0006] According to a second aspect of the present disclosure, an information processing method is provided, applied to a user terminal, comprising: responding to an input operation by a user on the user terminal, acquiring subject question information for a preset subject input by the user; sending the subject question information to a server; the server being configured to parse the subject question information, determine universal test point information involved in the solution process of the subject question information; determine universal test point content matching the universal test point information from a pre-generated test point content library or a searched test point content library; solve the subject question information to obtain targeted answer information corresponding to the subject question information; and receiving and displaying the targeted answer information and the universal test point content returned by the server for the subject question information.

[0007] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus; the memory is used to store a computer program; and the processor is used to execute the information processing method described in the first or second aspect by running the computer program stored in the memory.

[0008] According to a fourth aspect of the present disclosure, a computer storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the information processing method as described in the first or second aspect.

[0009] According to a fifth aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the information processing method as described in the first or second aspect.

[0010] According to the information processing scheme provided in this embodiment, by obtaining subject question information for a preset subject input by the user, the subject question information is then parsed to determine the universal test point information involved in the solution process of the subject question information, and universal test point content matching the universal test point information is determined from the test point content database; furthermore, the subject question information is solved to obtain the targeted answer information corresponding to the subject question information, thereby allowing the targeted answer information and the aforementioned universal test point content to be displayed to the user.

[0011] The solution provided in this embodiment not only provides users with targeted answers to subject-specific questions, but also provides users with valuable, universally applicable test points involved in the solution process of those subject-specific questions. These universally applicable test points include explanations of the universally applicable test points, which can help users comprehensively and systematically grasp the common test points involved in subject-specific questions, thereby meeting users' learning needs.

[0012] The solution provided in this disclosure can offer targeted answers to users seeking specific solutions to specific questions. Simultaneously, it intuitively presents explanations of the universally applicable test points (common test points) involved in solving the specific subject matter question. This avoids the need for users to perform additional question-and-answer interactions to find these universal test points. For users, this disclosure allows them to more likely understand a single question and master a type of question.

[0013] In summary, the embodiments disclosed herein do not require users to perform any additional operations, reducing user operating costs and meeting users' learning needs in a more efficient manner. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments recorded in the embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings.

[0015] Figure 1 is a schematic diagram of an information processing system according to an embodiment of the present disclosure.

[0016] Figure 2 is a flowchart of the steps of an information processing method according to an embodiment of the present disclosure.

[0017] Figures 3A to 3C are schematic diagrams illustrating the process of an information processing method according to an embodiment of this disclosure.

[0018] Figure 4 is a flowchart of the steps of an information processing method according to another embodiment of this disclosure.

[0019] Figure 5 is a schematic diagram of an example of an information processing scheme according to an embodiment of this disclosure.

[0020] Figure 6 is a structural block diagram of an information processing device according to an embodiment of the present disclosure.

[0021] Figure 7 is a structural block diagram of an information processing apparatus according to another embodiment of the present disclosure.

[0022] Figure 8 is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions in the embodiments of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art should fall within the protection scope of this disclosure.

[0024] The specific implementation of the embodiments of this disclosure will be further described below with reference to the accompanying drawings.

[0025] Figure 1 illustrates an exemplary system to which the embodiments of this disclosure apply. As shown in Figure 1, the system 100 may include a server 102, a communication network 104, and / or one or more user devices 106. In the example in Figure 1, multiple user devices 106 are shown. The user devices 106 are equipped with applications for displaying subject-specific question information, targeted answer information, and general test point content.

[0026] The server 102 can be any suitable device for storing information, data, programs, and / or any other suitable type of content, including but not limited to distributed storage system devices, server clusters, computing server clusters, etc. In some embodiments, the server 102 can perform any suitable function. For example, when the server 102 implements the scheme of the embodiments of this disclosure, in some embodiments, the server 102 can be used to perform information processing methods. As an optional example, in some embodiments, the server 102 can first obtain subject question information for a preset subject input by the user, then parse the subject question information to obtain the universal test point information involved in solving the subject question information; and determine the universal test point content that matches the universal test point information from a pre-generated test point content library or a searched test point content library; and solve the subject question information to obtain the targeted answer information corresponding to the subject question information, that is, the universal test point content and targeted answer information for the subject question information can be obtained. In some embodiments, the server 102 may receive subject question information sent by the user device 106, and after generating targeted answer information and universal test point content in the manner described above, send the targeted answer information and universal test point content to the user device 106 so that the user device 106 can display the targeted answer information and universal test point content.

[0027] In some embodiments, the communication network 104 may be any suitable combination of one or more wired and / or wireless networks. For example, the communication network 104 may include any one or more of the following: the Internet, an intranet, a wide area network (WAN), a local area network (LAN), a wireless network, a digital subscriber line (DSL) network, a frame relay network, an asynchronous transfer mode (ATM) network, a virtual private network (VPN), and / or any other suitable communication network. The user equipment 106 may be connected to the communication network 104 via one or more communication links (e.g., communication link 112), and the communication network 104 may be linked to the server 102 via one or more communication links (e.g., communication link 114). The communication link may be any communication link suitable for transmitting data between the user equipment 106 and the server 102, such as a network link, a dial-up link, a wireless link, a hardwired link, any other suitable communication link, or any suitable combination of such links.

[0028] Optionally, user device 106 may be configured with applications that display subject-specific question information, targeted answer information, and general test point content. User device 106 may include any one or more user devices suitable for displaying information and interacting with users. In some embodiments, user device 106 may include any suitable type of device. For example, in some embodiments, user device 106 may include mobile devices, tablet computers, laptop computers, desktop computers, and / or any other suitable type of user device.

[0029] Based on the above system, this disclosure provides an information processing scheme, which will be described below through several embodiments.

[0030] Figure 2 is a flowchart of an information processing method according to an embodiment of the present disclosure. According to a first aspect of the present disclosure, an information processing method is provided. Referring to Figure 2, the method includes steps S202, S204, and S206.

[0031] S202. Obtain subject-specific question information input by the user for a preset subject.

[0032] For example, a discipline refers to a specific academic field or knowledge system. This disclosure does not limit the classification method or specific content of the preset discipline. For example, common discipline classifications include natural sciences, engineering technology, medicine, social sciences, humanities, and mathematics. The subject-specific question information for a preset discipline can be question information belonging to the scope of that preset discipline. When a user wants to obtain the answer to a subject-specific question within a certain subject category, they can input subject-specific question information on their user terminal, such as taking a picture of the subject-specific question information or inputting question text, etc., to obtain an image or text containing subject-specific question information. The user terminal then sends the image or text containing subject-specific question information to the server. Afterwards, for the image containing subject-specific question information, image information recognition processing can be used to obtain the subject-specific question information input by the user for the preset discipline; for the text input by the user, the text can be used as the subject-specific question information input by the user for the preset discipline. For example, for the subject of mathematics, the subject-specific question information input by the user could be: "Solve this triangle problem: For a triangle with side lengths of 3, 4, and 5, find the angle."

[0033] S204. Analyze the subject question information and determine the universal test point information involved in the solution process of the subject question information; determine the universal test point content that matches the universal test point information from the pre-generated test point content library or the test point content library obtained by search.

[0034] For example, for a specific subject, the subject-related questions submitted by users often involve some frequently asked knowledge points within that subject. In this embodiment, these knowledge points can be referred to as universal test points. Further, universal test point information can be identifiable information such as the name and identifier of the universal test points parsed from the subject-related question information. For example, in mathematics, the universal test point information involved in solving the subject-related question information may include: trigonometric functions, the Pythagorean theorem, trigonometric identities, quadratic equations, etc. For example, for the subject-related question information in mathematics: "Solve this triangle problem: For a triangle with side lengths of 3, 4, and 5, find the angles.", the universal test point information involved in solving this subject-related question information includes the Pythagorean theorem and trigonometric functions.

[0035] The universally applicable test point content may include the universally applicable test point itself. Furthermore, the universally applicable test point content may also include explanatory content for the universally applicable test point; that is, the universally applicable test point content may include explanatory content corresponding to the universally applicable test point information, and correspondingly, the universally applicable test point content may also include explanatory content corresponding to the universally applicable test point information. For example, the universally applicable test point content in this embodiment can be rich media content, that is, the universally applicable test point content can be embodied as multimedia content containing multiple media elements (such as at least two of the following: images, audio, and video).

[0036] In this embodiment, a test point content library containing multiple test point contents can be pre-generated or obtained through a search. After obtaining the subject question information, the universal test point information involved in the subject question information can be obtained by parsing the subject question information. Then, the universal test point information can be matched with the test point information corresponding to the test point contents contained in the test point content library, so that the successfully matched test point content can be used as the universal test point content involved in the process of solving the subject question information.

[0037] In some optional embodiments, parsing subject-specific question information and determining universally applicable test point information involved in the solution process for the subject-specific question information includes: identifying keywords in the subject-specific question information to obtain first test point information; performing semantic analysis on the subject-specific question information to obtain second test point information; and fusing the first and second test point information to determine the universally applicable test point information involved in the subject-specific question information. Correspondingly, determining universally applicable test point content that matches the universally applicable test point information from a pre-generated test point content library or a searched test point content library may include: determining universally applicable test point content that matches the universally applicable test point information from a pre-generated test point content library or a searched test point content library based on a pre-determined mapping relationship between test point information and test point content.

[0038] For example, when parsing question information, keyword recognition can be performed on the subject question information. For instance, by performing keyword recognition on the subject question information based on a pre-defined keyword database, the first test point information can be obtained, such as the matched test point keywords. Additionally, semantic analysis can be performed on the subject question information. Further semantic recognition can be performed based on the context of the subject question information to obtain the second test point information, such as the test point information implicit in the subject question information besides the test point keywords. Then, the first and second test point information are merged. For example, the first and second test point information can be simply merged, or duplicate information can be removed from the first and second test point information before merging the processed test point information, thereby obtaining the universal test point information involved in the subject question information.

[0039] In this embodiment, by identifying keywords in the subject question information, the explicit first test point information contained in the subject question information can be obtained. Then, by performing semantic analysis on the subject question information, the implicit second test point information contained in the subject question information can be obtained. After combining the first test point information and the second test point information, a more comprehensive and universal test point information involved in the subject question information can be obtained, thereby avoiding the omission of test points in the subject question information.

[0040] In some alternative embodiments, determining universal test point content that matches universal test point information from a pre-generated test point content library or a searched test point content library may also include: determining target historical question information that matches universal test point information; and, based on the pre-determined mapping relationship between historical question information and test point content, determining the test point content corresponding to the target historical question information as universal test point content that matches universal test point information from a pre-generated test point content library or a searched test point content library.

[0041] For example, the test point content in the test point content library can be generated based on a portion of the question information in a preset question bank, which may include historical question information input by the user. After constructing the test point content library, a mapping relationship can be established between each test point content in the test point content library and its corresponding historical question information. Subsequently, when matching universal test point content corresponding to universal test point information among multiple test point contents, the target historical question information corresponding to the universal test point information can be matched from the historical question information based on the universal test point, and then, based on the pre-set mapping relationship between historical question information and test point content, the test point content corresponding to the target historical question information can be determined as the aforementioned universal test point content. Optionally, after parsing and obtaining the universal test point information involved in the solution process of the subject question information, the universal test point information can be matched with the test point content contained in the test point content library, and the successfully matched test point content can be used as the aforementioned universal test point content.

[0042] In this embodiment, when multiple test point contents are generated in advance based on historical question information, a mapping relationship between historical question information and test point contents is constructed. This allows for the matching of universal test point contents with the subject question information input by the user in the future. The corresponding target historical question information can be matched, and the test point contents corresponding to the target historical question information can be quickly determined as universal test point contents based on the mapping relationship between historical question information and test point contents.

[0043] Here, the process of pre-generating a test point content library containing multiple test point contents in this embodiment is illustrated by way of example.

[0044] In some optional embodiments, before obtaining the subject question information input by the user, the method of this embodiment further includes: determining the test point title corresponding to the historical question information based on the historical question information input by the user for the preset subject; and generating test point content corresponding to the historical question information through a target large language model based on the historical question information and the corresponding test point title.

[0045] For example, the pre-generated multiple test point contents can be generated based on historical question information input by the user. During the generation of test point content, based on the historical question information, the corresponding historical answer information and corresponding historical follow-up questions can be obtained. Then, based on the historical answer information and corresponding historical follow-up questions, a test point title corresponding to the historical question information is generated, that is, the title of the test point content to be generated. Afterwards, the historical question information and the corresponding test point title are input into a pre-trained target large language model, which generates the test point content corresponding to the historical question information.

[0046] The target large language model in this embodiment is a large language model capable of generating knowledge point content based on question information and corresponding knowledge point titles. A large language model (LLM) refers to a deep learning model trained on a large amount of text data, enabling it to generate natural language text or understand the meaning of language text. These models can provide in-depth knowledge and language production on various topics by training on massive datasets. Large language models can learn the patterns and structures of natural language through large-scale unsupervised training, simulating the human language cognition and generation process to a certain extent.

[0047] In some optional embodiments, based on historical question information and corresponding test point titles, test point content corresponding to the historical question information is generated through a target large language model. This includes: based on the historical question information and corresponding test point titles, a search is performed from a preset knowledge base using a search enhancement generation method to obtain a first search result associated with the historical question information and test point titles; the historical question information, test point titles, and the first search result are parsed to generate test point content prompts, and the test point content prompts are input into a pre-trained target large language model to generate test point content corresponding to the historical question information.

[0048] In this embodiment, based on the historical question information input by the user, the corresponding test point titles are determined. Then, a pre-trained target large language model is used to generate test point content corresponding to the historical question information based on the historical question information and the corresponding test point titles, improving the efficiency of test point content generation. Furthermore, utilizing the content generation capabilities of the large language model ensures the quality of the test point content. Additionally, the test point content in this embodiment is generated based on the historical question information input by the user, meaning that test point content can be generated targeting users' common concerns, allowing the generated test point content to more efficiently meet users' clustered needs. That is, in the above embodiment of this disclosure, when generating test point content, on the one hand, core test point titles are extracted and refined from the user's historical question information, and then test point content is generated based on the extracted test point titles. The aforementioned method of extracting and refining core test point titles allows the final generated test point content to better cover the questions raised by users, thereby improving the universality and focus of the generated test point content. On the other hand, after extracting the test point titles, this embodiment of the disclosure uses the content generation capability of the large language model to generate test point content based on historical question information and test point title guidance model. Therefore, the final generated test point content can better match historical question information and test point titles, improving the relevance of the test point content.

[0049] In some optional embodiments, determining the test point title corresponding to the historical question information for a preset subject based on the historical question information input by the user includes: obtaining multiple historical follow-up questions for the historical question information for a preset subject input by the user; filtering the multiple historical follow-up questions through preset filtering rules to obtain filtered historical follow-up questions; the preset filtering rules are used to filter the historical follow-up questions based on their relevance to the historical question information and their title style; and filtering the filtered historical follow-up questions through a pre-trained target classification model to obtain the test point title corresponding to the historical question information.

[0050] For example, referring to Figure 3A, during the generation of exam point titles, multiple historical follow-up questions can be obtained based on a single historical question. Here, based on the click count of each historical follow-up question corresponding to the historical question, multiple historical follow-up questions with click counts exceeding a preset click threshold can be obtained. This allows for the identification of common questions among most users, and subsequent exam point content can be generated based on these common questions, enabling the generated exam point content to more efficiently meet the collective needs of users. Then, multiple historical follow-up questions are filtered using preset filtering rules to obtain filtered historical follow-up questions. The preset filtering rules are used to filter the historical follow-up questions based on their relevance to the historical question information and their title style. For example, regarding the relevance filtering, firstly, historical follow-up questions corresponding to other historical question information with a similarity to the historical question information reaching a preset threshold can be obtained; secondly, since multiple historical follow-up questions may contain follow-up questions from multiple rounds of questioning regarding the historical question information, the first round of questioning can be filtered out as the filtered historical follow-up questions. Regarding title style filtering, preset filtering rules can be used to set exam point title styles, including word count requirements, sentence structure requirements, etc., but this embodiment does not impose any restrictions on this. By filtering historical follow-up information that meets the exam point title styles using the preset filtering rules, the filtered historical follow-up information can be obtained.

[0051] After obtaining multiple filtered historical follow-up questions, the filtered historical follow-up questions can be input into a pre-trained target classification model. This target classification model further classifies the filtered historical follow-up questions, thereby obtaining the test point titles corresponding to the historical question information based on the classification results. Here, the target classification model is a classification model trained using pre-labeled positive and negative sample data of test point titles.

[0052] Classification models are machine learning tools used to divide data into different categories or labels, such as decision tree models, logistic regression models, and Naive Bayes classifiers. This embodiment does not limit the classification model used. In training the classification model, positive and negative sample data are typically used. Positive sample data refers to data samples corresponding to the target category. For example, in this embodiment, positive sample data could be pre-labeled qualified exam point titles. Negative samples, in contrast to positive samples, are data samples that do not belong to the target category. For example, in this embodiment, negative sample data could be pre-labeled unqualified exam point titles.

[0053] In this embodiment, when generating test point titles, core test point titles are further extracted and refined from users' historical questioning behavior (including the question information of the first question and the follow-up questions based on the question information of the first question). The above-mentioned extraction and refinement method of core test point titles can make the final generated test point content better cover the questions raised by users, thereby improving the universality and aggregation of the generated test point content.

[0054] Furthermore, in this embodiment, after obtaining historical follow-up information, multiple pieces of historical follow-up information are first filtered based on relevance and title style using preset filtering rules to obtain historical follow-up information with higher relevance and better style. Then, a pre-trained target classification model is used to perform a second filtering of the filtered historical follow-up information, thus obtaining the test point titles corresponding to the historical question information. In this embodiment, historical follow-up information is filtered multiple times based on relevance, title style, and classification model, thereby obtaining the test point titles corresponding to the historical question information based on the finally filtered historical follow-up information. Therefore, the final generated test point titles have a higher relevance to the historical question information, and are more accurate and have a better style.

[0055] In some optional embodiments, determining the test point title corresponding to the historical question information based on the historical question information input by the user includes: determining the historical question information input by the user and obtaining multiple test point information contained in the historical question information; inputting the multiple test point information into a pre-trained first large language model and outputting the test point title corresponding to the historical question information; the first large language model is a large language model trained with test point information samples as training data and test point title samples as labels.

[0056] For example, referring to Figure 3A, in addition to the aforementioned method of generating test point titles, test point titles can also be generated in the following way: First, historical question information can be parsed to obtain multiple test point information contained in the historical question information. Then, the multiple test point information is input into a pre-trained first large language model. The first large language model is a large language model that is pre-trained with test point information samples as training data and test point title samples as labels. After that, the test point titles corresponding to the historical question information can be output through the first large language model.

[0057] This process involves pre-constructing a training dataset. Question information samples can be obtained from a pre-defined question bank (including historical question information). Parsing these samples yields test point information samples, which can include both explicit and implicit test point samples. These test point information samples form the training data. Correspondingly, label data—test point title samples—is constructed for each test point information sample. Then, the test point information samples and their corresponding test point title samples form the training dataset. This training dataset is used to fine-tune the large language model, resulting in the first large language model.

[0058] In this embodiment, the first language model, trained using test point information samples as training data and test point title samples as labels, processes multiple test point information corresponding to historical question information to obtain test point titles corresponding to historical question information. This method can better generate corresponding test point titles for some historical question information with a small number or poor quality of follow-up information, and the generation efficiency and accuracy are high.

[0059] In some optional embodiments, based on historical question information and corresponding test point titles, test point content corresponding to the historical question information is generated through a target large language model, including: inputting the historical question information and corresponding test point titles into a pre-trained second large language model, and outputting multiple query questions for the historical question information; for the multiple query questions, retrieving query results for the multiple query questions from a pre-built subject knowledge base; inputting the query results and test point titles into a pre-trained third large language model, and outputting test point content corresponding to the historical question information.

[0060] In some optional embodiments, historical question information and corresponding test point titles are input into a pre-trained second language model, which outputs multiple query questions for the historical question information. This includes: performing information retrieval based on the historical question information and corresponding test point titles to obtain second search results associated with the historical question information and test point titles; generating query question prompts based on the historical question information, test point titles, and second search results; and inputting the query question prompts into the pre-trained second language model to guide the second language model in generating multiple query questions for the historical question information.

[0061] In this embodiment, the query question represents the information collected from which test points related to historical question information to generate corresponding test point content. The query question is associated with the test point information contained in the historical question information. For example, if the historical question information is a question related to a law subject, the query question may include: Q1 Which legal provisions are included? Q2 What are some cases of theft? Q3 What are the differences between the crimes? etc. The query result is the test point content found for the query question. For example, the query result R1 for Q1 may be specific legal provisions, and the query result R2 for Q2 may be multiple cases of theft, etc.

[0062] The target large language model can include a second large language model and a third large language model. The second large language model is trained using question information samples and test point title samples as training data, and query question samples as labels. The third large language model is trained using test point paradigm samples as training data. The test point paradigm samples can refer to standardized test point content samples pre-defined according to different disciplines. For example, for the law discipline, the corresponding test point paradigm samples could be set so that each test point consists of four items: legal provisions, legal provision analysis, case studies, and case analysis, etc. This embodiment does not impose any limitations on this.

[0063] A training dataset for the second major language model can be pre-constructed. First, question information samples corresponding to different subjects can be obtained from a pre-set question bank (including historical question information). Correspondingly, test point title samples corresponding to the question information samples can be obtained using the test point title generation method described in the previous embodiment. For each subject, the training data for that subject consists of the question information samples and the corresponding test point information samples. Accordingly, label data for each subject—query question samples—is constructed based on the question information samples and corresponding test point information samples. Then, the training dataset for each subject is composed of the question information samples, the corresponding test point information samples, and the corresponding query question samples. Using this training dataset to train the major language model yields the second major language model.

[0064] In some optional embodiments, the process of training a large language model to obtain a second large language model using question information samples and test point title samples as training data and query question samples as labels may include: obtaining query question type information corresponding to multiple subjects; determining corresponding query question samples for each subject based on the query question type information corresponding to multiple subjects and for each subject's question information samples and test point title samples; and training a large language model using question information samples and test point title samples corresponding to each subject as training data and corresponding query question samples as labels to obtain a second large language model.

[0065] For example, query question type information corresponding to multiple different disciplines can be obtained. This query question type information can be characteristic information representing the discipline to which the query question corresponds. It can include the questioning angle of the query question and the proportion of questions corresponding to each angle. For example, for the discipline of law, Figure 3B illustrates the questioning angles and their corresponding proportions. Then, based on the query question type information corresponding to each discipline, query question samples corresponding to each discipline can be generated for each discipline's corresponding question information samples and exam point title samples. Finally, training data is composed of the question information samples and exam point title samples corresponding to each discipline. The large language model is then fine-tuned using the corresponding query question samples as labels to obtain the second large language model.

[0066] In this embodiment, by dividing different disciplines into corresponding query question type information, and based on the query question type information corresponding to each discipline, query question samples corresponding to each discipline are determined for the question information samples and test point title samples corresponding to each discipline, so as to train the second language model. This allows the second language model to output corresponding query questions for different disciplines more accurately and specifically.

[0067] Similarly, a training dataset for the third language model can be pre-built, and initial note paradigm samples can be constructed. Then, based on the style adjustment prompts corresponding to each subject, the initial test point paradigm samples are adjusted through a large language model (such as ChatGPT) to obtain test point paradigm samples corresponding to each subject. The test point paradigm samples corresponding to each subject constitute the training dataset. The large language model can be fine-tuned and trained using this training dataset to obtain the third language model.

[0068] In some optional embodiments, the process of constructing the test point paradigm sample includes: obtaining an initial test point paradigm sample; adjusting the initial test point paradigm sample through a large language model based on style adjustment prompts corresponding to multiple subjects to obtain the test point paradigm sample.

[0069] For example, when constructing test point paradigm samples, an initial test point paradigm sample can be built. Then, corresponding style adjustment prompts can be built for different subjects. These prompts guide the large language model to adjust the style of the initial test point paradigm sample. Finally, the initial test point paradigm sample and the style adjustment prompts for each subject are input into the large language model (e.g., ChatGPT). The large language model then adjusts the style of the initial test point paradigm sample based on the style adjustment prompts, outputting test point paradigm samples corresponding to each subject.

[0070] In this embodiment, by constructing style adjustment prompts based on multiple disciplines, the initial test point paradigm samples are adjusted using a large language model to obtain test point paradigm samples corresponding to each discipline, thereby training the third large language model. This allows the third large language model to generate test point content with corresponding styles for different disciplines more accurately and in a more targeted manner.

[0071] Referring to Figure 3C, after generating the test point titles, historical question information and corresponding test point titles can be input into the pre-trained second language model. Here, historical answer information corresponding to the historical question information can also be input into the second language model. Historical answer information includes the answer content and corresponding answer analysis; this embodiment does not impose limitations on this. The second language model can output multiple query questions related to the historical question information. These query questions are associated with the test point information contained in the historical question information. For example, if the historical question information is a question related to the legal discipline, the multiple query questions could include: Q1 Which legal provisions are included? Q2 What are some cases of theft? Q3 What are the differences between the crimes? etc. Then, for the generated multiple query questions, query results for each query question can be retrieved from the pre-built subject knowledge base, such as R1, R2, R3, etc. Finally, the query results and test point titles are input into the pre-trained third language model, which can then generate test point content corresponding to the historical question information.

[0072] In this embodiment, by inputting historical question information and corresponding test point titles into a pre-trained second language model, multiple query questions for the historical question information are output. For each query question, query results are retrieved from a pre-built subject knowledge base. The query results and test point titles are then input into a pre-trained third language model, which outputs test point content corresponding to the historical question information. This method can efficiently and in batches generate test point content corresponding to historical question information for different subjects. Furthermore, since the test point content is generated from query results retrieved from a pre-built subject knowledge base, the accuracy of the test point content is improved.

[0073] In some optional embodiments, historical question information and corresponding test point titles are input into a pre-trained second language model, which outputs multiple query questions for the historical question information. This includes: performing information retrieval based on the historical question information and corresponding test point titles to obtain second search results associated with the historical question information and test point titles; generating query question prompts based on the historical question information, test point titles, and second search results; and inputting the query question prompts into the pre-trained second language model to guide the second language model in generating multiple query questions for the historical question information.

[0074] The system analyzes historical question information and corresponding test point titles to generate query prompts. It then inputs the query prompts, historical question information, and corresponding test point titles into a pre-trained second language model, which outputs multiple query questions for the historical question information. The query prompts are used to guide the second language model to generate multiple query questions for the historical question information.

[0075] For example, after obtaining historical question information and corresponding test point titles, the target subject to which the historical question information and corresponding test point titles belong can be determined. For different target subjects, information retrieval can be performed in the target subject database to obtain second search results that are relevant to both the historical question information and test point titles. Then, according to the prompt word generation rules corresponding to the target subject, based on the historical question information, test point titles, and second search results, corresponding query prompt words, i.e., query prompts, are generated. Subsequently, the query prompt words are input into a pre-trained second language model, which is then guided to output multiple query questions related to the historical question information.

[0076] In this embodiment, by analyzing historical question information and corresponding test point titles, query question prompts are generated; using query question prompts to guide the second language model to generate multiple query questions for historical question information can improve the accuracy of the generated query questions.

[0077] S206. Solve the subject-specific questions to obtain the corresponding answer information.

[0078] For example, after obtaining the subject-specific question information input by the user, targeted answer information for the subject-specific question information can be generated. In this embodiment, the specific implementation method for generating targeted answer information is not limited. For example, it can be implemented using a generative language model. A generative language model is a neural network model based on deep learning. Its basic architecture may include recurrent neural networks (RNNs), long short-term memory networks (LSTMs), and transformers, etc. Through pre-training and fine-tuning strategies, it can learn the general rules and structures of language on large-scale unlabeled text data, and then generate text that conforms to grammatical and semantic rules based on the learned language patterns and knowledge. In this embodiment, the type and structure of the generative language model used are not limited; any suitable generative language model capable of generating targeted answer information for subject-specific question information is applicable to the solution of this embodiment.

[0079] The targeted answer information in this embodiment may include the answer content for the subject-specific question, and may also include the corresponding answer explanation. This embodiment does not limit this. Taking the subject-specific question information listed in S204 above, "Solve this triangle problem: For a triangle with side lengths of 3, 4, and 5, find the angles," as an example, the targeted answer information may include the specific answer to the subject-specific question information: "A triangle with side lengths of 3, 4, and 5 is a right triangle, with angles of 36.87 degrees, 53.13 degrees, and 90 degrees respectively." Furthermore, the targeted answer information may also include the solution process for the above specific answer.

[0080] S208. For subject-specific questions, display targeted answer information and universally applicable test points.

[0081] The information processing method provided in this embodiment not only provides users with targeted answers to subject-specific questions, but also provides users with valuable, universally applicable test points related to those subject-specific questions. These universally applicable test point contents include explanations of the test points related to the subject-specific questions, which can help users comprehensively and systematically grasp the common test points related to the subject-specific questions to meet their learning needs.

[0082] Furthermore, the solution provided in this disclosure can, in scenarios where a user seeks a specific answer to a particular question, not only provide a targeted answer but also intuitively present the explanations of the universally applicable test points (common test points) involved in solving other questions within that specific subject. This avoids the need for users to perform additional question-and-answer interactions to find these universal test points. For users, this disclosure allows them to more likely understand a single question and master a type of question.

[0083] In summary, the embodiments disclosed herein do not require users to perform any additional operations, reducing user operating costs and meeting users' learning needs in a more efficient manner.

[0084] Optionally, in some embodiments, the process of displaying targeted answer information and general test point content for subject-specific question information may include: creating an answer display page and displaying targeted answer information corresponding to the subject-specific question information on the answer display page; and in response to interactive operations on the targeted answer information displayed on the answer display page, popping up a pop-up page containing general test point content.

[0085] For example, in the above embodiments of this disclosure, the targeted answer information corresponding to the subject-specific questions is first displayed intuitively on the answer display page. This allows users to quickly obtain the core information—the targeted answer to the subject-specific questions—without having to search for the answer on the page. Furthermore, while reading the targeted answer, users can instantly obtain relevant test points by performing preset interactive operations (for example, clicking keywords in the targeted answer, clicking preset icons, etc.), thereby effectively enhancing interactivity and the continuity of learning. In addition, displaying the targeted answer in the main content area of ​​the answer display page, and then displaying the test point explanations in a further pop-up window, creates a clear distinction between primary and secondary content, allowing users to clearly differentiate between core and extended information. In summary, the display method of targeted answer information and universal test point content in the above embodiments can further improve the user experience.

[0086] Optionally, in some embodiments, after displaying targeted answer information and general test point content for subject-specific question information, it is also possible to: generate tagging information in response to the tagging operation of the displayed targeted answer information and general test point content; and generate question-and-answer notes based on the subject-specific question information, targeted answer information, general test point content, and tagging information.

[0087] Specifically, in the above embodiments, by generating question-and-answer notes, users can more easily review and summarize the questions they asked, improving the overall user experience. More importantly, generating question-and-answer notes integrates questions, answers, key points, and tagging data, allowing users to quickly view relevant key points and their own tagging data while reading the answers, reducing search and organization time and improving learning efficiency.

[0088] Optionally, in some embodiments, after generating the question-and-answer notes, the following can be further performed: determining target question-and-answer notes involving the same universal test point content from a note library containing multiple question-and-answer notes; merging the target question-and-answer notes to generate test point category notes for the same universal test point content.

[0089] Specifically, in the above embodiment, after generating corresponding question-and-answer notes based on the subject-specific question information input by the user, the notes are further organized. This involves merging different question-and-answer notes covering the same universally applicable test point into the same universally applicable test point category, resulting in test point category notes for that specific test point. In this way, users can use the test point category notes corresponding to a particular universally applicable test point to conduct a detailed and systematic review and fill in any gaps in their knowledge of that test point, thereby helping them to quickly and efficiently master that universally applicable test point.

[0090] Figure 4 is a flowchart of another information processing method according to an embodiment of the present disclosure. According to a second aspect of the present disclosure, an information processing method is provided, applied to a user terminal. Referring to Figure 4, the method includes steps S402, S404, and S406.

[0091] S402. In response to the user's input operation on the user terminal, obtain the subject question information for the preset subject input by the user.

[0092] For example, the user terminal can provide an input interface through which the user can input subject-related questions. After the user terminal detects the user's input, it can obtain the subject-related questions entered by the user.

[0093] In some optional embodiments, in response to a user's input operation on a user terminal, obtaining subject question information for a preset subject input by the user includes: in response to a user's shooting operation on a user terminal, obtaining an image containing subject question information input by the user; and performing information recognition on the image containing subject question information to obtain the subject question information input by the user.

[0094] For example, a camera button can be set in the input interface of the user terminal. The user can take a picture of the question by taking a picture of the question. After the user terminal detects the user's shooting operation, it can obtain the image containing the subject question information input by the user. Then, the user terminal can perform information recognition on the image containing the subject question information, such as multimodal information recognition, to obtain the subject question information contained in the image.

[0095] This embodiment improves user input efficiency by providing users with a photo-based question search input method, making it easier for users to quickly input questions and enhancing the user experience.

[0096] In some optional embodiments, in response to a user's input operation on a user terminal, obtaining subject question information for a preset subject input by the user includes: in response to a user's text input operation on a user terminal, obtaining subject question information for a preset subject input by the user.

[0097] For example, the user terminal's input interface can also include an input box, through which the user can enter question information. The user terminal then detects the text input operation within the input box and obtains the subject question information entered by the user. This embodiment provides the user with a text input method, facilitating use in various question-asking scenarios and improving the user experience.

[0098] S404. Send the subject question information to the server; the server parses the subject question information and determines the universal test point information involved in the solution process of the subject question information; it determines the universal test point content that matches the universal test point information from the pre-generated test point content library or the searched test point content library; it solves the subject question information and obtains the targeted answer information corresponding to the subject question information.

[0099] The server in this embodiment is used to implement the corresponding information processing methods in the various method embodiments of the first aspect described above, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0100] S406. Receive and display the targeted answer information and general test points returned by the server for the subject-specific questions.

[0101] For example, after receiving the targeted answer information and general test point content for the subject-specific questions returned by the server, the user terminal displays the targeted answer information and general test point content. This embodiment does not limit the display method.

[0102] The information processing method provided in this embodiment involves a user terminal detecting user input and then acquiring the subject-specific question information. This information is sent to a server. The server, based on the general test point information contained in the subject-specific question information, determines the matching general test point content from a test point content database and solves the question to obtain a targeted answer. This answer is then returned to the user terminal. Upon receiving the targeted answer and general test point content from the server, the user terminal can display the answer to the subject-specific question and provide corresponding general test point content. This general test point content includes explanations of the test points corresponding to the general test point information, helping the user comprehensively and systematically grasp the test point content related to the subject-specific question information to meet their learning needs. Furthermore, no additional user intervention is required, reducing user operating costs and meeting their learning needs more efficiently.

[0103] In some optional embodiments, when receiving and displaying the targeted answer information and general test point content for the subject-specific questions returned by the server, the method of this embodiment further includes: receiving at least one historical follow-up question information associated with the general test point information sent by the server, and displaying the historical follow-up question information.

[0104] Specifically, while receiving targeted answer information and general test point content for subject-specific questions from the server, the system can also simultaneously receive one or more historical follow-up questions associated with the general test point information. Afterwards, while displaying the targeted answer information and general test point content, the historical follow-up questions can also be displayed, allowing users to further select and view historical follow-up questions that meet their needs, eliminating the need for users to repeatedly enter the same follow-up questions, thus improving the user experience.

[0105] In some optional embodiments, after displaying the historical follow-up information, the method of this embodiment further includes: in response to the user's click operation on the historical follow-up information, obtaining the historical answer information corresponding to the historical follow-up information received in advance from the server, and displaying the historical answer information.

[0106] For example, when receiving one or more historical follow-up questions associated with general test point information from the server, the system can also receive historical answer information corresponding to those historical follow-up questions. After displaying one or more historical follow-up questions, the system can detect user clicks on those questions and, based on the clicks, further display the corresponding historical answer information. This allows users to quickly access the historical answer information for each follow-up question, eliminating the need for the server to repeatedly generate answer information for the historical follow-up questions and saving device resources.

[0107] The following describes, with reference to the scenario diagram shown in FIG5, an exemplary description of the overall implementation process of the information processing scheme of this disclosure embodiment. As shown in FIG5, the user takes an image containing subject question information at the input interface of the user terminal, obtains the subject question information, and sends the subject question information to the server. The server obtains the subject question information, and then parses the subject question information to obtain the universal test point information contained in the subject question information; and determines the universal test point content that matches the universal test point information from the test point content library; the universal test point content includes test point explanation content corresponding to the universal test point information; in addition, it generates targeted answer information for the subject question information through a generative language model, that is, it can obtain the universal test point content and targeted answer information for the subject question information, and returns the universal test point content and targeted answer information for the subject question information to the user terminal. The user terminal receives the targeted answer information and universal test point content for the subject question information returned by the server and displays them on the display interface.

[0108] It should be understood that more details of the overall implementation process shown in Figure 5 can be understood in conjunction with the preceding embodiments, and the overall implementation process shown in Figure 5 is only some examples for easy understanding of the embodiments of this disclosure, and is not intended to limit the embodiments of this disclosure in any way.

[0109] It is understood that the foregoing description of the information processing method is merely an exemplary description of some embodiments of this disclosure and is not intended to limit the embodiments of this disclosure in any way.

[0110] Referring to FIG6, a structural block diagram of an information processing apparatus according to an exemplary embodiment of the present disclosure is shown.

[0111] According to a third aspect of the present disclosure, an information processing device is provided, comprising: a first acquisition module 602, a test point matching module 604, a solution module 606, and a first display module 608.

[0112] The system comprises the following modules: a first acquisition module 602, used to acquire subject-specific question information input by the user for a preset subject; a test point matching module 604, used to parse the subject-specific question information and determine the universal test point information involved in the solution process of the subject-specific question information; and to determine the universal test point content that matches the universal test point information from a pre-generated test point content library or a searched test point content library; a solution module 606, used to solve the subject-specific question information and obtain the corresponding targeted answer information; and a first display module 608, used to display the targeted answer information and the universal test point content for the subject-specific question information.

[0113] In some optional embodiments, the test point matching module 604 is specifically used for: identifying keywords in the subject question information to obtain first test point information; performing semantic analysis on the subject question information to obtain second test point information; integrating the first and second test point information to determine the universal test point information involved in the subject question information; and determining the universal test point content that matches the universal test point information from a pre-generated test point content library or a searched test point content library based on a pre-determined mapping relationship between test point information and test point content.

[0114] In some optional embodiments, the first display module 608 is specifically used to create an answer display page and display targeted answer information corresponding to the subject question information on the answer display page; in response to interactive operations on the targeted answer information displayed on the answer display page, a pop-up window page containing universal test point content is displayed.

[0115] In some optional embodiments, the information processing apparatus further includes a note generation module for generating tagging information in response to tagging operations on the displayed targeted answer information and general test point content; and generating question-and-answer notes based on subject question information, targeted answer information, general test point content, and tagging information.

[0116] In some optional embodiments, the note generation module is further configured to: after generating the question-and-answer notes, determine the target question-and-answer notes involving the same universal test point content from the note library containing multiple question-and-answer notes; merge the target question-and-answer notes to generate test point category notes for the same universal test point content.

[0117] In some optional embodiments, the information processing device further includes: a test point content generation module, used to determine the test point title corresponding to the historical question information based on the user input of historical question information for a preset subject; and to generate test point content corresponding to the historical question information through a target large language model based on the historical question information and the corresponding test point title.

[0118] In some optional embodiments, the test point content generation module, when performing the step of generating test point content corresponding to historical question information through a target large language model based on historical question information and corresponding test point titles, specifically performs the following: based on historical question information and corresponding test point titles, it performs a search from a preset knowledge base using a search enhancement generation method to obtain a first search result associated with historical question information and test point titles; it parses historical question information, test point titles, and the first search result to generate test point content prompts, and inputs the test point content prompts into a pre-trained target large language model to generate test point content corresponding to historical question information through the target large language model.

[0119] In some optional embodiments, when the test point content generation module performs the step of determining the test point title corresponding to the historical question information based on the historical question information for a preset subject input by the user, it is specifically used to: obtain multiple historical follow-up questions based on the historical question information for a preset subject input by the user; filter the multiple historical follow-up questions through preset filtering rules to obtain filtered historical follow-up questions; the preset filtering rules are used to filter the historical follow-up questions based on their relevance to the historical question information and their title style; and filter the filtered historical follow-up questions through a pre-trained target classification model to obtain the test point title corresponding to the historical question information.

[0120] In some optional embodiments, when the test point content generation module performs the step of determining the test point title corresponding to the historical question information based on the historical question information for a preset subject input by the user, it is specifically used to: determine the historical question information for a preset subject input by the user, and obtain multiple test point information contained in the historical question information; input the multiple test point information into a pre-trained first large language model, and output the test point title corresponding to the historical question information; the first large language model is a large language model trained with test point information samples as training data and test point title samples as labels.

[0121] In some optional embodiments, the test point content generation module, when performing the step of generating test point content corresponding to the historical question information based on historical question information and corresponding test point titles through a target large language model, specifically performs the following: inputting the historical question information and corresponding test point titles into a pre-trained second large language model, and outputting multiple query questions for the historical question information; the second large language model is a large language model trained with question information samples and test point title samples as training data and query question samples as labels; for the multiple query questions, retrieving query results for the multiple query questions from a pre-constructed subject knowledge base; inputting the query results and test point titles into a pre-trained third large language model, and outputting test point content corresponding to the historical question information; the third large language model is a large language model trained with test point paradigm samples as training data.

[0122] In some optional embodiments, the test point content generation module, when performing the step of inputting historical question information and corresponding test point titles into a pre-trained second language model and outputting multiple query questions for the historical question information, specifically performs the following: performs information retrieval based on the historical question information and corresponding test point titles to obtain a second search result associated with the historical question information and test point titles; generates query question prompts based on the historical question information, test point titles, and the second search result; and inputs the query question prompts into the pre-trained second language model to guide the second language model to generate multiple query questions for the historical question information.

[0123] In some optional embodiments, the information processing device further includes: a training module, used to acquire query question type information corresponding to multiple disciplines; based on the query question type information corresponding to multiple disciplines, determine corresponding query question samples for question information samples and test point title samples corresponding to each discipline; use question information samples and test point title samples corresponding to each discipline as training data, and use the corresponding query question samples as labels to train the large language model to obtain a second large language model.

[0124] In some optional embodiments, the information processing device further includes: a construction module for obtaining an initial test point paradigm sample; and adjusting the initial test point paradigm sample using a large language model based on style adjustment prompts corresponding to multiple disciplines to obtain a test point paradigm sample.

[0125] The information processing apparatus of this embodiment is used to implement the corresponding information processing methods in the various method embodiments of the first aspect described above, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here. Furthermore, the functional implementation of each module in the information processing apparatus of this embodiment can be referred to the description of the corresponding part in the aforementioned method embodiments, which will also not be repeated here.

[0126] Referring to FIG7, a structural block diagram of an information processing apparatus according to an exemplary embodiment of the present disclosure is shown.

[0127] According to a fourth aspect of the present disclosure, an information processing apparatus is provided, applied to a user terminal, including a second acquisition module 702, a sending module 704, and a second display module 706.

[0128] The second acquisition module 702 is used to respond to the user's input operation on the user terminal and acquire the subject question information for the preset subject input by the user; the sending module 704 is used to send the question information to the server; the server is used to parse the subject question information, determine the universal test point information involved in the solution process of the subject question information; determine the universal test point content that matches the universal test point information from the pre-generated test point content library or the searched test point content library; solve the subject question information to obtain the targeted answer information corresponding to the subject question information; the second display module 706 is used to receive the targeted answer information and universal test point content returned by the server for the subject question information and display them.

[0129] In some optional embodiments, the second acquisition module 702 is further configured to: in response to the user's shooting operation on the user terminal, acquire an image containing subject question information input by the user; perform information recognition on the image containing subject question information to obtain the subject question information input by the user.

[0130] In some optional embodiments, the second acquisition module 702 is further configured to: in response to the user's text input operation on the user terminal, acquire the subject question information for the preset subject input by the user.

[0131] In some optional embodiments, the second display module 706 is further configured to: receive at least one historical follow-up question message associated with the universal test point information sent by the server, and display the historical follow-up question message.

[0132] In some optional embodiments, the second display module 706 is further configured to: in response to a user's click operation on historical follow-up information, obtain historical answer information corresponding to the historical follow-up information received in advance from the server, and display the historical answer information.

[0133] The information processing apparatus of this embodiment is used to implement the corresponding information processing methods in the various method embodiments of the second aspect described above, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here. Furthermore, the functional implementation of each module in the information processing apparatus of this embodiment can be referred to the description of the corresponding parts in the aforementioned method embodiments, which will also not be repeated here.

[0134] According to a fifth aspect of the present disclosure, an electronic device is provided, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus; the memory is used to store a computer program; and the processor is used to execute the information processing method of the first or second aspect described above by running the computer program stored in the memory.

[0135] Figure 8 shows a structural block diagram of an optional electronic device according to an embodiment of the present disclosure. The present disclosure does not limit the specific implementation of the electronic device 800. As an example, referring to Figure 8, the electronic device 800 provided in this embodiment includes: a processor 802, a communications interface 804, a memory 806, and a communication bus 808.

[0136] The processor 802, communication interface 804, and memory 806 communicate with each other through the communication bus 808.

[0137] Communication interface 804 is used to communicate with other electronic devices or servers.

[0138] The processor 802 is used to execute the computer program 810, specifically the relevant steps in any of the aforementioned information processing method embodiments.

[0139] Specifically, computer program 810 may include program code that includes computer operation instructions.

[0140] The processor 802 may be a CPU, a GPU (Graphics Processing Unit), an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present disclosure. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.

[0141] Memory 806 is used to store computer program 810. Memory 806 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0142] Specifically, computer program 810 can be used to cause processor 802 to execute the information processing method in any of the foregoing embodiments.

[0143] The specific implementation of each step in computer program 810 can be found in the corresponding descriptions of the steps and units in any of the foregoing information processing method embodiments, and will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices and modules described above can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0144] The electronic device 800 in this disclosure has been described in detail in the foregoing information processing method embodiments. Therefore, its related content and beneficial effects can be understood by referring to the above method embodiments, and will not be repeated here.

[0145] According to a sixth aspect of the present disclosure, the present disclosure also provides a computer storage medium storing a computer program thereon, which, when executed by a processor, implements the information processing method described in any of the foregoing method embodiments. The computer storage medium includes, but is not limited to, compact disc read-only memory (CD-ROM), random access memory (RAM), floppy disk, hard disk, or magneto-optical disk.

[0146] According to a seventh aspect of the present disclosure, the present disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the information processing method as described in any of the above-described method embodiments.

[0147] The electronic device 800 / computer storage medium / computer program product embodiments in this disclosure have been described in detail in the foregoing information processing method embodiments. Therefore, their related content and beneficial effects can be understood by referring to the above method embodiments, and will not be repeated here.

[0148] Furthermore, it should be noted that the user-related information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to sample data used for training the model, data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this disclosure are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0149] It should be noted that, depending on the implementation needs, the various components / steps described in the embodiments of this disclosure can be broken down into more components / steps, or two or more components / steps or parts of the operation of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of this disclosure.

[0150] The methods described above according to embodiments of this disclosure can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code originally stored on a remote recording medium or a non-transitory machine-readable medium and subsequently stored on a local recording medium, downloaded over a network. Thus, the methods described herein can be stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an Application Specific Integrated Circuit (ASIC) or a Field Programmable Gate Array (FPGA)). It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components (e.g., Random Access Memory (RAM), Read-Only Memory (ROM), Flash Memory, etc.) capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods described herein. Furthermore, when a general-purpose computer accesses code used to implement the methods shown herein, the execution of the code transforms the general-purpose computer into a dedicated computer for performing the methods shown herein.

[0151] Those skilled in the art will recognize that the units and method steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in 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 specific applications, but such implementations should not be considered beyond the scope of the embodiments disclosed herein.

[0152] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". It should be noted that the concepts of "first", "second", etc., mentioned in the embodiments of this disclosure are only used to distinguish different devices, modules, or units, and are not used to limit the order of functions performed by these devices, modules, or units or their interdependencies. It should be noted that the modifications of "a" and "a plurality" mentioned in the embodiments of this disclosure are illustrative and not restrictive, and those skilled in the art should understand that unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0153] The above embodiments are only used to illustrate the embodiments of this disclosure, and are not intended to limit the embodiments of this disclosure. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the embodiments of this disclosure. Therefore, all equivalent technical solutions also fall within the scope of the embodiments of this disclosure, and the patent protection scope of the embodiments of this disclosure should be defined by the claims.

Claims

1. An information processing method comprising: obtaining subject question information input by a user for a preset subject; analyzing the subject question information to determine universal examination point information involved in a solving process for the subject question information; determining universal examination point content matching the universal examination point information from a pre-generated examination point content library or a searched examination point content library; solving the subject question information to obtain specific answer information corresponding to the subject question information; displaying the specific answer information and the universal examination point content for the subject question information.

2. The method of claim 1, wherein, The analyzing the subject question information to determine universal examination point information involved in a solving process for the subject question information comprises: performing keyword recognition on the subject question information to obtain first examination point information; performing semantic analysis on the subject question information to obtain second examination point information; fusing the first examination point information and the second examination point information to determine universal examination point information involved in the subject question information; The determining universal examination point content matching the universal examination point information from a pre-generated examination point content library or a searched examination point content library comprises: According to a pre-determined mapping relationship between examination point information and examination point content, determine universal examination point content matching the universal examination point information from a pre-generated examination point content library or a searched examination point content library.

3. The method of claim 1, wherein, The displaying the specific answer information and the universal examination point content for the subject question information comprises: creating an answer display page and displaying specific answer information corresponding to the subject question information in the answer display page; in response to an interactive operation on the specific answer information displayed in the answer display page, popping up a pop-up window page containing the universal examination point content.

4. The method according to any one of claims 1 to 3, wherein, The method further comprises: in response to a marking operation on the displayed specific answer information and the universal examination point content, generating marking information; based on the subject question information, the specific answer information, the universal examination point content, and the marking information, generating a question and answer note.

5. The method of claim 4, wherein, After generating the question and answer note, the method further comprises: from a note library containing a plurality of question and answer notes, determining target question and answer notes involving the same universal examination point content; merging each target question and answer note to generate an examination point category note for the same universal examination point content.

6. The method of claim 1, wherein, Before obtaining subject question information input by a user for a preset subject, the method further comprises: determining an examination point title corresponding to historical question information input by a user for the preset subject; generating examination point content corresponding to the historical question information through a target large language model according to the historical question information and the corresponding examination point title.

7. The method of claim 6, wherein, The generating examination point content corresponding to the historical question information through a target large language model according to the historical examination point information and the corresponding examination point title comprises: According to the historical question information and the corresponding examination point title, a first retrieval result associated with the historical question information and the examination point title is obtained by searching a preset knowledge base in a retrieval enhancement generation mode; The historical question information, the examination point title, and the first retrieval result are analyzed to generate an examination point content prompt word, and the examination point content prompt word is input into a target large language model trained in advance to generate an examination point content corresponding to the historical question information through the target large language model.

8. The method of claim 6, wherein, The determination of the examination point title corresponding to the historical question information input by the user for the preset subject includes: Obtaining a plurality of historical follow-up question information for the historical question information input by the user for the preset subject; The plurality of historical follow-up question information is screened through a preset screening rule to obtain screened historical follow-up question information; the preset screening rule is used for association degree screening and title style screening of historical follow-up question information with the historical question information; The screened historical follow-up question information is screened through a target classification model trained in advance to obtain the examination point title corresponding to the historical question information.

9. The method of claim 6, wherein, The determination of the examination point title corresponding to the historical question information input by the user for a preset subject includes: Determine the historical question information input by the user for the preset subject, and obtain a plurality of examination point information contained in the historical question information; The plurality of examination point information is input into a first large language model trained in advance to output an examination point title corresponding to the historical question information; the first large language model is a large language model trained with examination point information samples as training data and examination point title samples as labels.

10. The method of any one of claims 6-9, wherein, The generation of the examination point content corresponding to the historical question information through the target large language model according to the historical question information and the corresponding examination point title includes: The historical question information and the corresponding examination point title are input into a second large language model trained in advance to output a plurality of query questions for the historical question information; the second large language model is a large language model trained with question information samples and examination point title samples as training data and query question samples as labels; For the plurality of query questions, query results for the plurality of query questions are retrieved from a pre-constructed subject knowledge base; The query results and the examination point title are input into a third large language model trained in advance to output an examination point content corresponding to the historical question information; the third large language model is a large language model trained with examination point paradigm samples as training data.

11. The method of claim 10, wherein, The input of the historical question information and the corresponding examination point title into the second large language model trained in advance to output a plurality of query questions for the historical question information includes: Based on the historical question information and the corresponding examination point title, a second retrieval result associated with the historical question information and the examination point title is obtained through information retrieval, and a query question prompt word is generated according to the historical question information, the examination point title, and the second retrieval result; inputting the query question prompt word into a pre-trained second large language model to guide the second large language model to generate a plurality of query questions for the historical question information through the query question prompt word.

12. The method of claim 11, wherein, The method further comprises: obtaining query question type information corresponding to a plurality of subjects; based on the query question type information corresponding to the plurality of subjects, determining corresponding query question samples for question information samples and examination point title samples corresponding to each subject; using the question information samples and the examination point title samples corresponding to each subject as training data, and using the corresponding query question samples as labels to train a large language model, to obtain the second large language model.

13. The method of claim 10, wherein, The construction process of the examination point paradigm sample comprises: obtaining an initial examination point paradigm sample; based on style adjustment prompt information corresponding to a plurality of subjects, adjusting the initial examination point paradigm sample through a large language model to obtain the examination point paradigm sample.

14. An information processing method applied to a user terminal, comprising: in response to a user input operation on the user terminal, obtaining subject question information input by the user for a preset subject; sending the subject question information to a server; the server is used for analyzing the subject question information, determining the universal examination point information involved in the solving process of the subject question information, determining the universal examination point content matched with the universal examination point information from the pre-generated examination point content library or the searched examination point content library, and solving the subject question information to obtain the corresponding specific answer information of the subject question information; receiving and displaying the specific answer information and the universal examination point content returned by the server for the subject question information.

15. The method of claim 14, wherein, The response to the user input operation on the user terminal to obtain the subject question information input by the user for the preset subject comprises: in response to a user shooting operation on the user terminal, obtaining an image containing subject question information input by the user; performing information recognition on the image containing subject question information to obtain the subject question information input by the user; and / or in response to a user text input operation on the user terminal, obtaining subject question information input by the user for a preset subject.

16. The method of any one of claims 14-15, wherein, When receiving and displaying the specific answer information and the universal examination point content returned by the server for the subject information, the method further comprises: receiving at least one piece of historical follow-up information associated with the universal examination point information sent by the server, and displaying the historical follow-up information.

17. The method of claim 16, wherein, After displaying the historical follow-up information, the method further comprises: in response to a user click operation on the historical follow-up information, obtaining historical answer information corresponding to the historical follow-up information pre-received from the server, and displaying the historical answer information.

18. An electronic device, comprising: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete communication with each other through the communication bus; the memory is used for storing a computer program; The processor is configured to execute the computer program stored on the memory to perform the method of any one of claims 1-17.

19. A computer storage medium having stored thereon a computer program which, when executed by a processor, performs the method of any one of claims 1-17.

20. A computer program product comprising computer instructions that direct a computing device to perform the method of any one of claims 1-17.

Images