Information interaction method and device, information interaction system, electronic equipment and medium

Abstract

The present disclosure provides an information interaction method and device, and relates to the technical field of education. The specific implementation scheme is: receiving the to-be-processed corpus issued by a learning object; determining interaction corpus information based on the to-be-processed corpus and a pre-constructed vector database; sending the to-be-processed corpus and the interaction corpus information to a large language model to obtain courseware information and / or interaction information output by the large language model; sending the courseware information to an educational digital human with an educational role; and sending the interaction information to a partner digital human with a partner role, so that the educational digital human and the partner digital human interact with the learning object in a course system structure. The learning result information of the learning object is improved.

Description

TECHNICAL FIELD

[0001] The present disclosure belongs to the technical field of education, and relates to the technical fields of online education, artificial intelligence, etc., in particular to an information interaction method and device, an information interaction system, an electronic device, and a computer readable storage medium. BACKGROUND

[0002] In the current online education field, digital courseware has become an important means of educational informatization. However, although the existing digital courseware provides rich educational resources and learning methods, there are still some problems in the actual teaching process. On the one hand, digital courseware often lacks real interactive experience, making it difficult for students to truly immerse themselves in the learning environment; on the other hand, the existing digital courseware lacks a socialized learning mechanism, and cannot effectively simulate the learning and communication scene in real life, thereby affecting the learning experience of students. SUMMARY

[0003] The present disclosure provides an information interaction method and device, an electronic device, and a computer readable storage medium.

[0004] According to a first aspect, an information interaction method is provided, which includes: receiving a to-be-processed corpus issued by a learning object; determining interaction corpus information based on the to-be-processed corpus and a pre-constructed vector database; sending the to-be-processed corpus and the interaction corpus information to a large language model to obtain courseware information and / or interaction information output by the large language model; sending the courseware information to an educational digital person with an educational role to enable the educational digital person to interact with the learning object; and sending the interaction information to a partner digital person with a partner role to enable the partner digital person to interact with the learning object.

[0005] According to a second aspect, an information interaction device is provided, which includes: a receiving unit configured to receive a to-be-processed corpus issued by a learning object; a determining unit configured to determine interaction corpus information based on the to-be-processed corpus and a pre-constructed vector database; a obtaining unit configured to send the to-be-processed corpus and the interaction corpus information to a large language model to obtain courseware information and / or interaction information output by the large language model; a sending unit configured to send the courseware information to an educational digital person with an educational role to enable the educational digital person to interact with the learning object; and an interaction unit configured to send the interaction information to a partner digital person with a partner role to enable the partner digital person to interact with the learning object.

[0006] According to a third aspect, an information interaction system is provided, comprising a digital human interaction module, an educational digital human having an educational role, and a partner digital human having a partner role; the digital human interaction module is configured to receive a to-be-processed corpus issued by a learning object; determine interaction corpus information based on the to-be-processed corpus and a pre-constructed vector database; send the to-be-processed corpus and the interaction corpus information to a large language model to obtain courseware information and interaction information output by the large language model; send the courseware information to the educational digital human; send the interaction information to the partner digital human; the educational digital human is configured to generate teaching explanation information based on the courseware information, and provide the teaching explanation information to the learning object; receive a classroom question of the learning object, and generate a classroom answer to the classroom question of the learning object; the partner digital human is configured to generate learning corpus information based on the interaction information and the teaching explanation information, and provide the learning corpus information to the learning object.

[0007] According to a fourth aspect, an electronic device is provided, comprising at least one processor; and a memory connected with the at least one processor in communication, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in any implementation manner of the first aspect.

[0008] According to a fifth aspect, a non-transitory computer-readable storage medium storing computer instructions is provided, and the computer instructions are used to make a computer execute the method described in any implementation manner of the first aspect.

[0009] The information interaction method and device provided by the embodiments of the present disclosure first receive a to-be-processed corpus issued by a learning object; secondly, determine interaction corpus information based on the to-be-processed corpus and a pre-constructed vector database; thirdly, send the to-be-processed corpus and the interaction corpus information to a large language model to obtain courseware information and / or interaction information output by the large language model; then, send the courseware information to an educational digital human having an educational role, so that the educational digital human interacts with the learning object; finally, send the interaction information to a partner digital human having a partner role, so that the partner digital human interacts with the learning object. Thus, through the method, the learning object and the digital human having different roles can have a real interactive experience, the learning object can be immersed in the learning environment, and thus the learning communication scene in real life is effectively simulated, and the learning experience of the learning object is improved.

[0010] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present disclosure, nor to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0011] The accompanying drawings are used to better understand the present scheme, and do not constitute a limitation on the present disclosure. Among them:

[0012] Figure 1 is a flow chart of an embodiment of the information interaction method according to the present disclosure;

[0013] Figure 2 is a structural diagram of a course of the intelligent courseware self-learning platform in the present disclosure;

[0014] Figure 3 is a flow chart of another embodiment of the information interaction method according to the present disclosure;

[0015] Figure 4 is a flow chart of still another embodiment of the information interaction method according to the present disclosure;

[0016] Figure 5 is a structural schematic diagram of an embodiment of the information interaction device according to the present disclosure;

[0017] Figure 6 is a structural schematic diagram of another embodiment of the information interaction device according to the present disclosure;

[0018] Figure 7 is a structural schematic diagram of still another embodiment of the information interaction device according to the present disclosure;

[0019] Figure 8 is a structural schematic diagram of an embodiment of the information interaction system according to the present disclosure;

[0020] Figure 9 is a block diagram of an electronic device for implementing the information interaction method of the embodiments of the present disclosure. DETAILED DESCRIPTION

[0021] Unless otherwise explicitly stated, throughout the specification and claims, the term "comprise" or variations such as "comprises" or "comprising" will be understood to imply the inclusion of a stated element or group of elements but not the exclusion of any other element or group of elements.

[0022] The technical scheme of the present disclosure is described below through specific embodiments. It should be understood that the one or more steps mentioned in the present disclosure do not exclude other methods and steps before and after the combination steps, or other methods and steps can be inserted between these explicitly mentioned steps. It should also be understood that these examples are only used to illustrate the present disclosure and do not limit the scope of the present disclosure. Unless otherwise stated, the numbering of each method step is only for the purpose of identifying each method step, and is not limited to the arrangement order of each method or the scope of the implementation of the present disclosure. Changes or adjustments of the relative relationship, without substantial technical content changes, can also be considered as the implementation scope of the present disclosure.

[0023] The raw materials and instruments used in the examples are not specifically limited in origin, and can be purchased on the market or prepared according to conventional methods well known to those skilled in the art.

[0024] Defects of existing digital courseware: Although the existing digital courseware provides rich educational resources and learning methods, there are still some problems in the actual teaching process. On the one hand, the digital courseware often lacks real interactive experience, making it difficult for students to truly immerse themselves in the learning environment; on the other hand, the existing digital courseware lacks a socialized learning mechanism and cannot effectively simulate real-life learning communication scenarios, thereby affecting the students' learning result information experience.

[0025] In view of the defects of the existing digital courseware, the present disclosure proposes an information interaction method, which creates and integrates two digital people into the digital courseware platform. The design of these roles not only aims to simulate real human educational interaction scenarios, providing personalized educational strategies, learning support and social interaction, but also incorporates elements of confrontation, cooperation and competition to form a hybrid learning environment that simulates a real social environment. Figure 1 The flow 100 of one embodiment of the information interaction method according to the present disclosure is shown, which includes the following steps:

[0026] Step 101, receiving the to-be-processed corpus issued by the learning object.

[0027] In this embodiment, the execution subject on which the information interaction method runs can be an API (Application Programming Interface) of a digital courseware autonomous learning platform. The learning object issues the to-be-processed corpus to the digital courseware platform, wherein the learning object is an online learning object, which can be a student and a terminal possessed by the student. The to-be-processed corpus is various questions and interactive content proposed by the learning object to the AI digital person in the form of voice or text, such as course content, knowledge points, and problem descriptions of the learning object, etc. Specifically, the to-be-processed corpus can be multi-modal data, such as one or more of images, texts or voices.

[0028] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of personal information of the learning object comply with relevant laws and regulations and do not violate public order and good customs. In the technical solution of the present disclosure, all data related to the learning object (such as the to-be-processed corpus) can be obtained from a public data set or obtained from the learning object with the authorization of the learning object.

[0029] Step 102, determining the interactive corpus information based on the to-be-processed corpus and the pre-constructed vector database.

[0030] In this embodiment, the vector database is a knowledge base that stores vectors and corpora, where the vectors are stored one-to-one with the corpora. The interactive corpus information can include context information of the to-be-processed corpus and historical interaction information. The context information is the context corpus information in which the to-be-processed corpus is located, and the historical interaction information is the interaction content between the learning object and the educational digital person, the partner digital person in the historical period, such as user questions, feedback, selection, etc. The given question mainly comes from the interactive process between the learner and the AI digital person. Specifically, when the learner encounters doubts or needs to further understand a certain knowledge point in the learning process, they can ask questions to the AI digital person in the form of voice or text. These questions are the given questions.

[0031] For example, when the learning object is learning a mathematics course, it may not understand a certain formula or theorem and directly issues the to-be-processed corpus to the AI digital person as: “How is this formula derived?” or “What are some examples of the application of this theorem?” and so on. After the information interaction method receives the to-be-processed corpus, it converts the to-be-processed corpus into a vector representation, and then uses a similarity search algorithm to retrieve the most relevant interactive corpus information in the vector database, such as the implementation principle of the formula, etc.

[0032] In this embodiment, in the digital courseware self-learning platform, the user can upload course materials in any format, such as classroom notes, slides, textbooks, recorded voice recordings, etc. The digital courseware self-learning platform stores these course materials in the vector database for subsequent retrieval and generation. As shown in Figure 2 The structural diagram of the course of the digital courseware self-learning platform is shown. This structural diagram starts from a large unit, which is further divided into 13 teaching modules. Each teaching module contains 7 teaching steps, which are the main learning activities or stages within the module. In each teaching step, it is further divided into 7 micro-steps, which represent the specific actions or tasks required to achieve the teaching goals.

[0033] In Figure 2 The large unit is the top layer of the course system, which contains the theme or field of the entire course. The teaching modules: there are 13 teaching modules under the large unit, each module focuses on a specific knowledge or skill field within the large unit. Each teaching module contains 7 teaching steps, and each step is further divided into 7 micro-steps. These steps and micro-steps are the basic units of learning for the learning object, ensuring the systematicness and depth of learning. In each step, the AI digital person will appear in the 3rd, 4th, 5th, 6th, and 7th micro-step links to provide more specific and personalized learning support and interaction.

[0034] In the embodiment, the educational digital person is a digital person with an educational role and can provide courseware or educational knowledge for the learning object, and the partner digital person is a digital person with a partner role and can provide life or learning information for the learning object.

[0035] In the embodiment, the to-be-processed corpus is converted into a vector to obtain a to-be-processed vector; the to-be-processed vector is compared with vectors in a pre-constructed vector database in terms of similarity, vectors with similarity greater than a similarity threshold are determined, corpus of the vectors with similarity greater than the similarity threshold is extracted, and interactive corpus information is obtained.

[0036] In step 103, the to-be-processed corpus and the interactive corpus information are sent to the large language model to obtain courseware information and / or interactive information output by the large language model.

[0037] In the embodiment, the large language model analyzes the to-be-processed corpus to determine the learning needs of the learning object, determines courseware information that can be directly used for teaching or learning of the educational digital person according to the learning needs, and determines interactive information that can be directly used for the partner digital person according to the learning needs.

[0038] In the embodiment, the courseware information is generated courseware content and can be directly used for teaching or learning. When the courseware information is sent to the educational digital person, the educational digital person can directly convert the courseware information into corresponding teaching content, thereby facilitating learning of the learning object.

[0039] In the embodiment, the interactive information is interactive content generated for the partner digital person, which can enable the partner digital person to interact with the learning object in real time.

[0040] In step 104, the courseware information is sent to the educational digital person with an educational role to enable the educational digital person to interact with the learning object.

[0041] In the embodiment, the courseware information is sent to the educational digital person with an educational role, and the educational digital person can provide corresponding course explanation for the learning object in the digital intelligent courseware self-learning platform. The learning object learns according to the explanation of the educational digital person and issues the to-be-processed corpus again.

[0042] As the core interface of the system, the API is responsible for delivering these analysis results and prediction information to the AI digital person, enabling the AI digital person to adjust its response mode and teaching content in real time according to the user's behavior and learning state. For example, when the user shows confusion on a certain knowledge point, the API will call the machine learning model and user behavior analysis module to obtain relevant analysis results, and deliver these information to the AI digital person, enabling the AI digital person to provide targeted guidance and help in a timely manner.

[0043] In step 105, the interaction information is sent to the partner digital human with a partner role, so that the partner digital human interacts with the learning object.

[0044] In this embodiment, after the large language model outputs the interaction information, the interaction information is sent to the partner digital human with a partner role, which can realize the socialized learning mechanism in the virtual intelligent courseware autonomous learning platform, simulate the learning communication scene in real life, and improve the learning experience of the learning object.

[0045] The information interaction method provided by the embodiments of the present disclosure first receives the to-be-processed corpus issued by the learning object; secondly, determines the interaction corpus information based on the to-be-processed corpus and the pre-constructed vector database; thirdly, sends the to-be-processed corpus and the interaction corpus information to the large language model to obtain the courseware information and / or the interaction information output by the large language model; then, sends the courseware information to the educational digital human with an education role, so that the educational digital human interacts with the learning object; finally, sends the interaction information to the partner digital human with a partner role, so that the partner digital human interacts with the learning object. Therefore, through this method, the learning object and the digital human with different roles can have a real interactive experience, the learning object can be immersed in the learning environment, and thus the learning communication scene in real life is effectively simulated, and the learning experience of the learning object is improved.

[0046] In some embodiments of the present disclosure, the above information interaction further includes: obtaining object behavior data of the learning object; based on the object behavior data, obtaining habit information and preference information of the learning object; based on the habit information and the preference information, determining teaching data and resource information, and sending the teaching data and the resource information to the educational digital human.

[0047] In this embodiment, the object behavior data is a series of actions or operations that the object has performed. These actions or operations are not isolated, but are closely related to the attributes of the object. The source of the behavior data can be operation records on the intelligent courseware autonomous learning platform, such as learning duration, learning frequency, accessed courses, knowledge points, completed exercises, test scores, clicks, searches, collections, and other interactive behaviors. The data form of the object behavior data can be structured data (such as log files, database records) or unstructured data (such as text, video records).

[0048] In this embodiment, the habit information can be the learning rules of the learning object, such as: daily learning time period, preferred learning content type (such as video, text, and practice questions), learning progress and rhythm, etc. The preference information can be the interests and inclinations of the learning object, such as: preference for certain topics (such as programming, mathematics), preference for certain teaching methods (such as interactive learning, self-directed learning), etc. The above habit information and preference information of the learning object based on the object behavior data include: extracting the learning actions of the learning object with an occurrence frequency greater than a preset number and the operation objects of the learning actions from the object behavior data, taking the learning actions and the operation objects as the habit information of the learning object; extracting the behavior preferences with an occurrence frequency greater than a preset number from the object behavior data as the preference information of the learning object.

[0049] In this embodiment, the teaching data is to generate personalized learning suggestions according to the behavior data of the learning object, such as: recommending suitable learning content, adjusting the learning plan (such as increasing the learning time of certain knowledge points), and providing learning skills and methods. The resource information can be related resources recommended according to the preferences of the learning object, such as: suitable courses, teaching materials, related practice questions, test questions, extended reading materials or videos, etc.

[0050] In this embodiment, the generated teaching data and resource information are sent to the educational digital person, and the educational digital person interacts with the learning object according to these information to provide personalized learning support. This method can significantly improve the learning experience and effect, and provide more accurate learning support for the learning object.

[0051] In some optional implementations of the present disclosure, the above habit information and preference information of the learning object based on the object behavior data include: inputting the object behavior data into a pre-trained behavior analysis model to obtain the habit information and preference information of the learning object output by the behavior analysis model.

[0052] In this optional implementation, the input data of the behavior analysis model can be the object behavior data, including learning duration, learning frequency, accessed content, test scores, etc. The output data is the habit information and preference information of the learning object, such as: learning time period preference (such as morning, evening), learning content preference (such as video, text, and practice questions), learning progress and rhythm, interest degree in certain topics (such as programming, mathematics), etc. When training the behavior analysis model, the training data can use historical behavior data and corresponding habit and preference labels.

[0053] In this optional implementation, the behavior analysis model can adopt a classifier or a regression model, mainly used for predicting user behavior and analyzing learning result information. These models can identify user behavior patterns and learning habits by learning and training on user historical data, thereby providing personalized teaching data and resources for AI digital people. For example, if the model predicts that the user may have difficulty in a certain knowledge point, the AI digital person can prepare relevant learning materials and exercises in advance to help the user better understand and master. User behavior analysis focuses on applying statistical analysis and pattern recognition techniques to in-depth analysis of user behavior patterns. It can help understand user learning habits and preferences, thereby providing more accurate teaching support for AI digital people. For example, by analyzing user behavior data on the platform, such as click rate, dwell time, course progress, etc., the user's interest and mastery of different knowledge points can be understood, and the teaching strategy and content of the AI digital person can be adjusted accordingly.

[0054] In this optional implementation, the behavior analysis model can input real-time or historical learning object behavior data, and output habit information and preference information of the learning object. The behavior analysis model can update the model regularly based on new behavior data to improve prediction accuracy, thereby improving the accuracy of habit information and preference information prediction of the learning object.

[0055] In some embodiments of the present disclosure, there are multiple partner digital people, and the partner roles of the respective partner digital people are different. The above information interaction method further includes: obtaining object behavior data of the learning object; obtaining habit information and preference information of the learning object based on the object behavior data; performing semantic recognition on the interaction information to determine a semantic recognition result; generating role identifiers of the respective partner digital people and output texts under the respective role identifiers based on the semantic recognition result, the habit information and the preference information; and sending the respective role identifiers and the output texts under the respective role identifiers to corresponding partner digital people in the multiple partner digital people, so that the partner digital people process the output texts from the perspective of the partner roles corresponding to the role identifiers.

[0056] In this embodiment, there are multiple partner digital people and the partner roles of the respective partner digital people are different. For example, the partner roles include: an antagonistic role, a competitive role, a cooperative role, and an interfering role. The role identifier can uniquely identify the partner role of the partner digital person. When there are multiple partner roles, there are multiple role identifiers. Each partner digital person can determine the partner role it needs to implement based on the role identifier sent to it by the execution subject on which the information interaction method runs.

[0057] In this embodiment, the habit information of the learning object includes the interactive object that the learning object is used to in the historical period, for example, the habit information represents that the learning object is used to sending information to the peer with stronger competitiveness in the historical period, and the role identifier generated by the execution subject on which the information interaction method runs corresponds to the competitive role.

[0058] In this embodiment, the preference information of the learning object includes the preferred interactive object of the learning object, for example, the preference information represents that the learning object prefers to communicate with the person with gentle character, and the role identifier generated by the execution subject on which the information interaction method runs corresponds to the cooperative role.

[0059] In this embodiment, the output text is the text that needs to be output to the learning object by the partner digital person, and the output text is the text determined through the semantic recognition result. After the partner digital person determines the partner role thereof through the role identifier, the output text can be processed into information corresponding to the partner role and output to the learning object, for example, the semantic recognition result is: guiding the learning object to learn knowledge point A, and the partner role outputs detailed introduction of the knowledge point A to the learning object under the role identifier of the cooperative role thereof, so that the learning object can more effectively absorb the knowledge point A.

[0060] The information interaction method provided in this embodiment includes: obtaining object behavior data of a learning object; obtaining habit information and preference information of the learning object based on the object behavior data; performing semantic recognition on interactive information to determine a semantic recognition result; generating a role identifier of each partner digital person and output text under each role identifier based on the semantic recognition result, the habit information and the preference information; and sending each role identifier and the output text under each role identifier to a corresponding partner digital person in a plurality of partner digital persons, so that the partner digital person processes the output text from the perspective of the partner role corresponding to the role identifier. The learning object can increase the diversity and freshness of learning through interaction with the partner digital person of different partner roles, constantly refresh the learning experience of the learning object, and avoid the learner from being bored with the learning materials. By providing the partner digital person of different AI roles, a learning environment that continuously changes and is full of vitality can be created, so as to improve the participation and learning efficiency of the learning object.

[0061] In some embodiments of the present disclosure, the above method further includes: determining learning result information of the learning object based on the habit information and the preference information; determining a teaching strategy and content of an educational digital person based on the learning result information, and sending the teaching strategy and content to the educational digital person.

[0062] In this embodiment, the habit information and the preference information can be input into a pre-trained result classification model to obtain learning result information of the learning object output by the result classification model, wherein the result classification model is used for learning result classification of the habit information and the preference information of the learning object, and the habit information and the preference information beneficial to the learning effect are divided into one category, and the habit information and the preference information not beneficial to the learning effect are divided into another category.

[0063] In this embodiment, the learning result information is the current result of the learning object, and the learning result information includes learning duration, learning frequency, knowledge point mastery, error rate and the like of the learning object. Through the learning result information, the learning progress and effect of the learning object can be judged.

[0064] In this embodiment, the teaching strategy and content can be information obtained after adjustment of original teaching data and resource information. The teaching strategy and content can be information changing the data order in the teaching data and resource information, or information obtained after new teaching suggestions and / or resource information are added to the educational data and resource information. For example, through the teaching data and resource information, it can be determined that the learning object performs weakly on some knowledge points, and the educational digital person can provide more exercises or adjust the teaching method (such as increasing interactivity, using more intuitive teaching materials, etc.). If the learning object has mastered some content, the educational digital person can skip or simplify the relevant part to improve the learning efficiency. The educational digital person carries out personalized teaching on the learning object according to the received teaching strategy and content.

[0065] Optionally, the above determining the teaching strategy and content of the educational digital person based on the learning result information comprises: performing multi-index effect evaluation on the learning result information to obtain an evaluation result; and based on the evaluation result, matching the teaching strategy and content of the teaching digital person from a pre-set teaching strategy set. Wherein, the multi-indexes include: quantitative indexes such as test scores, knowledge point answering accuracy, error type distribution (such as concept confusion, calculation error), etc.; practical indexes such as learning duration, interaction frequency, micro-step completion efficiency, resource click preference (such as video, text, experiment), etc.; processing indexes such as voice emotion analysis (such as frustration, excitement), text interaction emotion recognition (such as NLP emotion classification), physiological signals (such as eye movement tracking attention monitoring), etc.

[0066] The evaluation model is used for multi-index evaluation of the learning result, wherein the evaluation model can use lightweight machine learning (such as random forest, XGBoost) to grade the learning effect, such as excellent, up to standard, and needing intervention; or associate error knowledge points based on a knowledge graph (Knowledge Graph) to identify weak links, such as "Newton's law application" associated error.

[0067] The information interaction method provided by the embodiment determines the teaching strategy and content of the educational digital person based on the learning result information, sends the teaching strategy and content to the educational digital person, and the educational digital person can dynamically adjust the teaching strategy to provide more personalized and efficient learning experience and improve the learning experience of the learning object.

[0068] Optionally, the information interaction method further includes determining the performance information of the digital persons in different roles in the partner digital person based on the learning result information, and sending the corresponding performance information to each digital person in the partner digital person. The digital persons in different roles include an excellent student AI role, a bad student AI role, a neutral student AI role, and a digital person for increasing the interest of the class. The partner digital person of the excellent student AI role is used to play the role of a model student, guide and encourage other students to learn actively, and the performance information of the partner digital person of the excellent student AI role includes sharing learning skills and knowledge points, encouraging peers to learn and discuss and share knowledge. The partner digital person of the bad student AI role is used to simulate the behavior of a student in real life that may reduce learning efficiency to improve the attractiveness of the teaching content, and the performance information of the partner digital person of the bad student AI role includes showing some bad learning habits for educating learners to avoid similar behaviors to stimulate learners to reflect and discuss through negative teaching materials. The partner digital person of the neutral student AI role is used to play an ordinary student to reflect the ordinary performance and attitude of most students. The performance information of the partner digital person of the bad student AI role includes emotional reaction information reflecting common learning emotions such as doubt, curiosity, etc., and basic interaction information participating in basic learning activities and discussions. The digital person for increasing the interest of the class is used to enhance the attractiveness and emotional depth of the class by providing emotional support and interesting interaction. The performance information of the digital person for increasing the interest of the class includes emotional support information for showing empathy and emotional support to help reduce learning pressure, and game-based learning information for introducing game elements and challenges to make the learning process more lively and interesting.

[0069] In some optional implementations of the present disclosure, the learning result information includes different result levels and knowledge points associated with each result level, and the determination of the teaching strategy and content of the educational digital person based on the learning result information and the sending of the teaching strategy and content to the educational digital person include determining the stratified education strategy of the educational digital person corresponding to each result level based on the result level, introducing the corresponding knowledge points in the stratified education strategy of each result level, and extracting associated resources from the courseware library based on the knowledge points. The stratified education strategies of all result levels, the knowledge points, and the associated resources are used as the teaching strategy and content, and the teaching strategy and content are sent to the educational digital person.

[0070] In this optional implementation, the learning result information is divided into different result levels (e.g., excellent, good, medium, and needs improvement). Each result level is associated with a specific knowledge point or knowledge field. For example, the excellent level may be associated with higher-order or expanded knowledge points, and the needs improvement level may be associated with basic or consolidated knowledge points. Determine the hierarchical education strategy: based on each result level, design a corresponding hierarchical education strategy for each level. For example, for learning objects with an "excellent" level, the educational digital person can use an expansion teaching strategy to provide more challenging content or in-depth exploration of related knowledge points; for learning objects with a "needs improvement" level, the educational digital person can use a consolidation teaching strategy to provide more basic exercises or repeated explanations of key content.

[0071] In this optional implementation, the hierarchical education strategy includes: introducing high-order challenge tasks (such as interdisciplinary project design), enabling "adversarial learning" roles; strengthening knowledge expansion (such as case deepening), triggering "collaborative learning" scenarios; starting basic knowledge reprocessing (such as micro-step relearning), calling "expert agent" targeted guidance; if the same knowledge point has consecutive errors ≥ 3 times, automatically switch to "step-by-step guidance mode" (such as 3D animation disassembly steps); when eye movement data indicates that attention is scattered, switch to "gamified learning" or "bad student interference scenarios" to refocus.

[0072] In this optional implementation, in each hierarchical education strategy, introduce the knowledge points associated with the result level. Based on these knowledge points, extract relevant associated resources (such as videos, exercises, case analysis, interactive modules, etc.) from the courseware library to support the teaching of the educational digital person. Send the teaching strategy and content to the educational digital person: send the determined hierarchical education strategy, knowledge points, and associated resources to the educational digital person. The educational digital person dynamically adjusts the teaching content and method based on the received information, providing personalized learning experiences for learning objects. For example, assume that the learning result information of the learning object in the "mathematics geometry" field is evaluated as "needs improvement", and the execution subject on which the information interaction method runs will perform the following steps: determine the hierarchical education strategy corresponding to the "needs improvement" level as a consolidation teaching strategy, associated with "basic geometry knowledge points" (such as triangle properties, angle calculation, etc.), extract relevant associated resources (such as basic explanation videos, exercises, interactive geometry tools, etc.) from the courseware library. Send these strategies and resources to the educational digital person, which will teach based on the specific circumstances of the learning object.

[0073] The method for determining and sending teaching strategies and content provided by the optional implementation determines the hierarchical education strategies of the educational digital person corresponding to each result classification based on the result classification; introduces the corresponding knowledge points in the hierarchical education strategies of each result classification, and extracts the associated resources from the courseware library based on the knowledge points, takes all the hierarchical education strategies of the result classification, the knowledge points and the associated resources as the teaching strategies and content, and sends the teaching strategies and content to the educational digital person. Based on different result classifications of the learning object, the hierarchical education strategies are determined, which can enable the educational digital person to realize precise and hierarchical teaching, and improve the learning experience of the learning object.

[0074] In some optional implementations of the present disclosure, the above determining the interactive corpus information based on the to-be-processed corpus and the pre-constructed vector database comprises: converting the to-be-processed corpus into a corpus vector; matching the course materials and the emotional elements from the pre-constructed vector database based on the corpus vector; and taking the course materials and the emotional elements as the interactive corpus information.

[0075] In the optional implementation, the corpus vector refers to a feature vector of the to-be-processed corpus. The way of converting the to-be-processed corpus into a corpus vector is different according to different data types of the to-be-processed corpus. For example, when the to-be-processed corpus is text data, the to-be-processed corpus can be converted into a corpus vector by word embedding. Video and audio data can extract the corpus vector of the to-be-processed corpus by convolutional neural network and recurrent neural network.

[0076] In the optional implementation, the course materials are the course information content involved in the learning of the learning object, such as courseware of different grades and different subjects; the emotional elements are various characteristics and factors related to emotions in the behavior of the learning object, including emotions, emotional expression and emotional experience, and are a kind of expression and embodiment of the inner world of human beings. Emotional elements play an important role in individual psychology and social behavior, and affect the mental health and social behavior of individuals.

[0077] In the optional implementation, by converting the to-be-processed corpus into a vector representation, the relevant course materials and emotional elements are detected, which can facilitate the execution subject to better understand and respond to the needs of the learning object.

[0078] In the optional implementation, the vector database includes a plurality of library vectors, each library vector has corresponding course materials and emotional elements, and the above matching the course materials and the emotional elements from the pre-constructed vector database based on the corpus vector comprises: comparing the similarity of the corpus vector and the library vector in the vector database, and in response to the similarity of the corpus vector and the library vector being greater than a similarity threshold, taking the course materials and the emotional elements corresponding to the library vector as the matched course materials and emotional elements.

[0079] The method for obtaining interactive corpus information provided by the optional implementation manner converts the to-be-processed corpus into a corpus vector, matches course materials and emotional elements from a pre-constructed vector database based on the corpus vector, and takes the course materials and emotional elements as the interactive corpus information, thereby improving the comprehensiveness of the obtained interactive corpus information.

[0080] Optionally, the information guidance method further includes: receiving feedback data of the learning object; sending the feedback data, the to-be-processed corpus and the interactive corpus information to the large language model to obtain new courseware information and new interactive information output by the large language model; sending the new courseware information to the educational digital human; and sending the new interactive information to the partner digital human. The feedback data is data for feedback on the educational digital human and / or the partner digital human, such as appreciation or opposition to the educational strategy output by the educational digital human, or positive or negative responses to the performance of the partner digital human.

[0081] Figure 3 Flow 300 of another embodiment of the information interaction method according to the present disclosure is shown, and the information interaction method of the present embodiment is applied to an educational digital human. The information interaction method includes the following steps:

[0082] Step 301: receiving courseware information.

[0083] In the present embodiment, the courseware information is the generated courseware content, which can be directly used for teaching or learning. When the courseware information is sent to the educational digital human, the educational digital human can directly convert the courseware information into corresponding teaching content, thereby facilitating the learning of the learning object. The courseware information in step 301 can be the courseware information sent by the execution subject (such as the API of the digital courseware self-learning platform) of the information interaction method shown in the present disclosure to the educational digital human of the present disclosure, and the courseware information is information generated due to the to-be-processed corpus sent by the learning object. Figure 1

[0084] Step 302: generating teaching explanation information based on the courseware information and providing the teaching explanation information to the learning object.

[0085] In the present embodiment, the teaching explanation information is the information explanation content sent by the educational digital human to the learning object, and the teaching explanation information can effectively answer or explain the corresponding to-be-processed corpus for the learning object.

[0086] In the present embodiment, the execution subject on which the information interaction method runs can perform information decomposition on the courseware information according to the pre-set teaching strategy and content to generate the teaching explanation information. For example, if the teaching strategy and content are progressive teaching, the execution subject decomposes the different levels of content in the courseware information in a progressive manner to generate the teaching explanation information.​

[0087] Optionally, the execution subject on which the information interaction method runs can also obtain teaching data and resource information from the API of the intelligent courseware self-learning platform, determine a teaching strategy based on the teaching data, and generate teaching explanation information applicable to the teaching strategy based on the resource information and the courseware information according to the teaching strategy.

[0088] Optionally, the execution subject on which the information interaction method runs can also obtain teaching data and resource information from the API of the intelligent courseware self-learning platform, determine a teaching strategy based on the teaching data, and generate teaching explanation information applicable to the teaching strategy based on the resource information and the courseware information according to the teaching strategy.

[0089] Step 303: Receive a classroom question of the learning object.

[0090] In this embodiment, the classroom question is a question issued by the learning object after receiving the teaching explanation information.

[0091] Step 304: Generate a classroom answer based on the classroom question, and send the classroom answer to the learning object.

[0092] In this optional implementation manner, the classroom answer is reply information to the classroom question, and the digital educational figure selects relevant knowledge points from the courseware information after obtaining the classroom question, and organizes the knowledge points according to the internal relationship of the knowledge points to obtain the classroom answer.

[0093] In this optional implementation manner, the digital educational figure can directly send the classroom answer to the learning object through the intelligent courseware self-learning platform, so as to enable the learning object to effectively learn the courseware information.

[0094] The information interaction method provided in this embodiment receives courseware information, generates teaching explanation information based on the courseware information, and provides the teaching explanation information to a learning object, receives a classroom question of the learning object, generates a classroom answer based on the classroom question, and sends the classroom answer to the learning object, thereby providing an effective learning method for the learning object and improving the learning experience of the learning object.

[0095] Optionally, the above information interaction method further includes: extracting a vector of the teaching explanation information and the classroom answer, and storing the vector and the teaching explanation information and the classroom answer in a vector database.

[0096] In some optional implementations of the present disclosure, the above information interaction method further includes: extracting semantic vectors and sentiment vectors in the teaching explanation information and the classroom answer to obtain a first vector; and storing the first vector, the teaching explanation information, and the classroom answer in a vector database.

[0097] In the optional implementation, the first vector is a vector obtained after vector conversion of semantics and emotion of the teaching explanation information and the classroom answer, and the semantics vector and the emotion vector in the teaching explanation information and the classroom answer can be effectively represented by the first vector by extracting the semantics vector and the emotion vector in the teaching explanation information and the classroom answer. The first vector is stored in the vector database, and the teaching explanation information and the classroom answer can be determined by vector retrieval.

[0098] Figure 4 A flow 400 of still another embodiment of the information interaction method according to the present disclosure is shown, and the information interaction method of the embodiment is applied to a partner digital person. The information interaction method includes the following steps:

[0099] In step 401, the interactive information and the teaching explanation information sent by the educational digital person are received.

[0100] In the embodiment, the interactive information is interactive content generated for the partner digital person, and the interactive information can enable the partner digital person to interact with the learning object in real time. The interactive information can be information generated due to the learning object issuing the to-be-processed corpus.

[0101] In step 402, learning corpus information is generated based on the interactive information and the teaching explanation information.

[0102] In the embodiment, the learning corpus information is content of understanding of the teaching explanation information by the partner digital person to the learning object when the learning object learns the teaching explanation information. The learning corpus information is information issued by the partner digital person to help the learning object to deeply understand the teaching explanation information.

[0103] In the embodiment, the step 402 includes: determining an interactive content template with the learning object based on the interactive information; extracting interactive knowledge points based on the teaching explanation information; and filling the interactive knowledge points into the interactive content template to obtain the learning corpus information. For example, the interactive content template is: xx, do you know how to understand yy in this place? After obtaining the knowledge points, the knowledge points are filled into the “yy” position of the interactive content template to obtain the learning corpus information.

[0104] In this embodiment, the partner digital person also has a corresponding role, and the content of the learning corpus information is different when the role of the partner digital person is different. For example, the partner digital person includes an AI role of a good student, an AI role of a bad student, an AI role of a neutral student, and a digital person for increasing the interest of the class. The partner digital person in the AI role of a good student is used to play the role of a model student, guide and encourage other students to learn actively, and the learning corpus information of the partner digital person in the AI role of a good student includes sharing learning skills knowledge points, encouraging peers to learn and discuss and share knowledge. The partner digital person in the AI role of a bad student is used to simulate the behavior of a student in real life that may reduce learning efficiency to improve the attractiveness of the teaching content, and the learning corpus information of the partner digital person in the AI role of a bad student includes: showing some bad learning habits for educating learners to avoid similar behaviors to stimulate learners to reflect and discuss through negative teaching materials. The partner digital person in the AI role of a neutral student is used to play an ordinary student, reflecting the ordinary performance and attitude of most students. The learning corpus information of the partner digital person in the AI role of a bad student includes: emotional response information reflecting common learning emotions such as doubt, curiosity, etc.; basic interaction information participating in basic learning activities and discussions. The digital person for increasing the interest of the class is used to enhance the attractiveness and emotional depth of the class by providing emotional support and interesting interaction. The learning corpus information of the digital person for increasing the interest of the class includes: emotional support information for showing empathy and emotional support to help reduce learning pressure; game-based learning information for introducing game elements and challenges to make the learning process more lively and interesting.

[0105] In step 403, the learning corpus information is provided to the learning object.

[0106] The information interaction method provided in this embodiment receives interaction information and teaching explanation information sent by the educational digital person, generates learning corpus information based on the interaction information and the teaching explanation information, and provides the learning corpus information to the learning object, thereby effectively providing learning support for the learning object.

[0107] Optionally, the above information interaction method further includes extracting a vector of the learning corpus information, and storing the vector and the learning corpus information in a vector database. In this embodiment, the learning corpus information can include learning resources, collaboration tasks, emotional support, and game-based content.

[0108] In some optional implementations of the present disclosure, the above information interaction method further includes extracting a semantic vector and an emotional vector of the learning corpus information, obtaining a second vector, and storing the second vector and the learning corpus information in a vector database.

[0109] In the optional implementation, the second vector is a vector obtained after vector conversion of semantics and emotion of the learning corpus information, and the semantics vector and the emotion vector in the learning corpus information can be effectively represented by extracting the semantics vector and the emotion vector in the learning corpus information and the second vector. The second vector is stored in the vector database, and the learning corpus information can be determined by vector retrieval.

[0110] Further referring to Figure 5 , as an implementation of the method shown in the above figures, the present disclosure provides an embodiment of an information interaction device, which corresponds to the method embodiment shown in Figure 1 .

[0111] As shown in Figure 5 , the information interaction device 500 provided in the embodiment includes a corpus receiving unit 501, a determining unit 502, an obtaining unit 503, a sending unit 504, and an interaction unit 505. The corpus receiving unit 501 can be configured to receive the to-be-processed corpus issued by the learning object. The determining unit 502 can be configured to determine the interactive corpus information based on the to-be-processed corpus and the pre-constructed vector database. The obtaining unit 503 can be configured to send the to-be-processed corpus and the interactive corpus information to the large language model to obtain the course information and / or the interaction information output by the large language model. The sending unit 504 can be configured to send the course information to the educational digital person with an educational role to enable the educational digital person to interact with the learning object. The interaction unit 505 can be configured to send the interaction information to the partner digital person with a partner role to enable the partner digital person to interact with the learning object.

[0112] In the embodiment, the specific processing of the corpus receiving unit 501, the determining unit 502, the obtaining unit 503, the sending unit 504, and the interaction unit 505 in the information interaction device 500 and the technical effects brought by the specific processing can be respectively referred to the related descriptions of the steps 101, 102, 103, 104, and 105 in the corresponding embodiment, which will not be described herein again. Figure 1 The corresponding embodiment in the corresponding embodiment, the steps 101, 102, 103, 104, 105, the related description of the corresponding embodiment in the corresponding embodiment, which will not be described herein again.

[0113] In some embodiments of the present disclosure, the device 500 further includes an analysis unit (not shown in the figure), which is configured to: obtain object behavior data of the learning object; obtain habit information and preference information of the learning object based on the object behavior data; determine the teaching data and the resource information based on the habit information and the preference information, and send the teaching data and the resource information to the educational digital person.

[0114] In some embodiments of the present disclosure, the above-mentioned multiple partner digital humans have different partner roles, and the above-mentioned device 500 further comprises a role assigning unit (not shown in the figure), which is configured to: obtain object behavior data of the learning object; obtain habit information and preference information of the learning object based on the object behavior data; perform semantic recognition on the interaction information to determine a semantic recognition result; generate a role identifier of each partner digital human and output text under each role identifier based on the semantic recognition result, the habit information and the preference information; and send each role identifier and the output text under each role identifier to a corresponding partner digital human in the multiple partner digital humans, so that the partner digital human processes the output text from the perspective of the partner role corresponding to the role identifier.

[0115] In some embodiments of the present disclosure, the above-mentioned device 500 further comprises a planning unit (not shown in the figure), which is configured to: determine learning result information of the learning object based on the habit information and the preference information, the learning result information including different result levels and knowledge points associated with each result level; determine a stratified education strategy of an educational digital human corresponding to each result level based on the result level; introduce the corresponding knowledge points in the stratified education strategy of each result level and extract associated resources from a courseware library based on the knowledge points; and send the stratified education strategies of all result levels, the knowledge points and the associated resources as teaching strategies and contents to the educational digital human.

[0116] In some embodiments of the present disclosure, the determining unit 502 of the above-mentioned device is further configured to: convert the to-be-processed corpus into a corpus vector; match course materials and emotional elements from a pre-constructed vector database based on the corpus vector; and take the course materials and the emotional elements as the interaction corpus information.

[0117] The information interaction device provided by the embodiments of the present disclosure first receives the to-be-processed corpus issued by the learning object through the corpus receiving unit 501; secondly, determines the interaction corpus information based on the to-be-processed corpus and the pre-constructed vector database through the determining unit 502; thirdly, sends the to-be-processed corpus and the interaction corpus information to the large language model through the obtaining unit 503 to obtain courseware information and / or interaction information output by the large language model; then, sends the courseware information to the educational digital human with an educational role through the sending unit 504 to enable the educational digital human to interact with the learning object; and finally, sends the interaction information to the partner digital human with a partner role through the interaction unit 505 to enable the partner digital human to interact with the learning object. Thus, through the method, the learning object and the digital human with different roles can have a real interactive experience, the learning object can be immersed in the learning environment, and thus the learning communication scene in real life is effectively simulated, and the learning experience of the learning object is improved.

[0118] Further referring to Figure 6 , as an implementation of the method shown in the above figures, the present disclosure provides another embodiment of an information interaction device, which is applied to an educational digital person, and the device embodiment corresponds to the method embodiment shown in Figure 3 .

[0119] As shown in Figure 6 , the information interaction device 600 provided in the embodiment includes a courseware receiving unit 601, an explanation unit 602, a question receiving unit 603, and an answer generating unit 604. The courseware receiving unit 601 can be configured to receive courseware information. The explanation unit 602 can be configured to generate teaching explanation information based on the courseware information and provide the teaching explanation information to a learning object. The question receiving unit 603 can be configured to receive a classroom question of the learning object. The answer generating unit 604 can be configured to generate a classroom answer based on the classroom question and send the classroom answer to the learning object.

[0120] In the embodiment, the specific processing of the courseware receiving unit 601, the explanation unit 602, the question receiving unit 603, and the answer generating unit 604 in the information interaction device 600 and the technical effects brought by the specific processing can be respectively referred to the related descriptions of the steps 301, 302, 303, and 304 in the corresponding embodiment, which will not be repeated here. Figure 3

[0121] In some embodiments of the present disclosure, the information interaction device 600 further includes a first storage unit (not shown in the figure), which is configured to extract semantic vectors and sentiment vectors in the teaching explanation information and the classroom answer to obtain first vectors, and store the first vectors, the teaching explanation information, and the classroom answer in a vector database.

[0122] Further referring to Figure 7 , as an implementation of the method shown in the above figures, the present disclosure provides another embodiment of an information interaction device, which is applied to an educational digital person, and the device embodiment corresponds to the method embodiment shown in Figure 4 .

[0123] As shown in Figure 7 ​As shown, the information interaction device 700 provided in this embodiment includes: an information receiving unit 701, a corpus generation unit 702, and a providing unit 703. The information receiving unit 701 can be configured to receive interactive information and teaching explanation information sent by the educational digital human. The corpus generation unit 702 can be configured to generate learning corpus information based on the interactive information and teaching explanation information. The providing unit 703 can be configured to provide learning corpus information to the learner. In this embodiment, the specific processing of the information receiving unit 701, the corpus generation unit 702, and the providing unit 703 in the information interaction device 700, and the resulting technical effects, can be found by referring to [reference needed]. Figure 4 The relevant descriptions of steps 401, 402, and 403 in the corresponding embodiments will not be repeated here.

[0124] In some embodiments of this disclosure, the information interaction device 700 further includes a second storage unit (not shown in the figure), which is configured to: extract semantic vectors and sentiment vectors from the learning corpus information to obtain a second vector; and store the second vector and the learning corpus information in a vector database accordingly.

[0125] like Figure 8 As shown, the information interaction system 800 provided in this embodiment includes: a digital human interaction module 801, an educational digital human 802, and a partner digital human 803. The digital human interaction module 801 receives the corpus to be processed from the learner; determines the interaction corpus information based on the corpus to be processed and a pre-built vector database; sends the corpus to be processed and the interaction corpus information to a large language model to obtain courseware information and interaction information output by the large language model; sends the courseware information to the educational digital human 802; and sends the interaction information to the partner digital human 803. The educational digital human 802 provides teaching explanation information to the learner based on the courseware information; receives classroom questions from the learner and generates classroom answers to send to the learner. The partner digital human 803 provides learning corpus information to the learner based on the interaction information and the teaching explanation information.

[0126] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0127] Figure 9A schematic block diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their patterns are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0128] like Figure 9 As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded into random access memory (RAM) 903 from storage unit 908. RAM 903 may also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0129] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0130] The computing unit 901 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 901 performs various methods and processes described above, such as the information interaction method. For example, in some embodiments, the information interaction method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the information interaction method described above can be performed. Alternatively, in other embodiments, the computing unit 901 can be configured to perform the information interaction method by any other appropriate means, such as by means of firmware.

[0131] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0132] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable information processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, implements the methods / operations specified in the flowcharts and / or block diagrams. The program code can execute entirely on a machine, partly on a machine, as a stand-alone software package, partly on a machine and partly on a remote machine or entirely on a remote machine or server.

[0133] In the context of this disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0134] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0135] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0136] It should be understood that various forms of flow shown above can be re-ordered, added to, or deleted from without departing from the spirit of the present disclosure. For example, the steps recited in the present disclosure can be performed in parallel, in series, or in a different order, without departing from the desired results of the technology disclosed in the present disclosure, which are not limited herein.

[0137] The foregoing description of specific exemplary embodiments of the disclosure has been presented for the purposes of illustration and explanation. It is not intended to be exhaustive or to limit the disclosure to the precise forms disclosed, and obviously many modifications and variations are possible in light of the above teaching. It is intended that the embodiments be chosen and / or described such that the disclosure is practical and sufficient for one skilled in the art to practice and utilize the disclosure, as well as to customize the disclosure for various uses and conditions. The scope of the disclosure is intended to be limited only by the claims and their equivalents.

Claims

1. An information interaction method, the method comprising: receiving a to-be-processed corpus issued by a learning object; determining interaction corpus information based on the to-be-processed corpus and a pre-constructed vector database; sending the to-be-processed corpus and the interaction corpus information to a large language model to obtain course information and / or interaction information output by the large language model; sending the course information to an educational digital human with an educational role to enable the educational digital human to interact with the learning object; sending the interaction information to a partner digital human with a partner role to enable the partner digital human to interact with the learning object; obtaining object behavior data of the learning object; obtaining habit information and preference information of the learning object based on the object behavior data; determining learning result information of the learning object based on the habit information and the preference information; determining performance information of respective digital humans of different roles in the partner digital human based on the learning result information, and sending corresponding performance information to respective digital humans in the partner digital human; wherein the digital humans of different roles include: a good student AI role, a bad student AI role, a neutral student AI role, and a digital human for increasing interest in class; the performance information of the partner digital human of the bad student AI role includes: displaying bad learning habits and emotional response information, which is used to enable an educational learner to avoid similar behaviors, and the performance information of the digital human for increasing interest in class includes: emotional support information, which is used to exhibit empathy and emotional support; and game-based learning information, which is used to introduce game elements and challenges to make the learning process more lively and interesting.

2. The method of claim 1, further comprising: determining teaching data and resource information based on the habit information and the preference information, and sending the teaching data and resource information to the educational digital human.

3. The method of claim 1, wherein the partner digital human has a plurality of different partner roles, and the method further comprises: performing semantic recognition on the interaction information to determine a semantic recognition result; generating a role identifier of each partner digital human and output text under each role identifier based on the semantic recognition result, the habit information, and the preference information; and sending each role identifier and the output text under each role identifier to a corresponding partner digital human in the plurality of partner digital humans to enable the partner digital human to process the output text from the perspective of the partner role corresponding to the role identifier.

4. The method of claim 2 or 3, further comprising: the learning result information includes different result levels and knowledge points associated with each result level; and based on each result level, a stratified education strategy of the educational digital human corresponding to the result level is determined; introducing corresponding knowledge points in the stratified education strategy of each result level, and extracting associated resources from a course library based on the knowledge points; all stratified education strategies, knowledge points, and associated resources of the result levels are used as teaching strategies and content, and the teaching strategies and content are sent to the educational digital human.

5. The method of claim 1, wherein the determining the interactive corpus information based on the to-be-processed corpus and the pre-constructed vector database comprises: converting the to-be-processed corpus into a corpus vector; matching, based on the corpus vector, a course material and an emotional element from the pre-constructed vector database; and taking the course material and the emotional element as the interactive corpus information.

6. An information interaction method applied to an educational digital person of the information interaction method of any one of claims 1-5, the method comprising: receiving courseware information; generating teaching explanation information based on the courseware information and providing the teaching explanation information to a learning object; receiving a classroom question of the learning object; and generating a classroom answer based on the classroom question and sending the classroom answer to the learning object.

7. The method of claim 6, further comprising: extracting semantic vectors and emotional vectors in the teaching explanation information and the classroom answer to obtain a first vector; and storing the first vector, the teaching explanation information, and the classroom answer in a vector database.

8. An information interaction method applied to a partner digital person of the information interaction method of any one of claims 1-5, the method comprising: receiving interactive information and teaching explanation information sent by an educational digital person; generating learning corpus information based on the interactive information and the teaching explanation information; and providing the learning corpus information to a learning object.

9. The method of claim 8, further comprising: extracting semantic vectors and emotional vectors in the learning corpus information to obtain a second vector; and storing the second vector and the learning corpus information in a vector database.

10. An information interaction apparatus, comprising: a corpus receiving unit configured to receive to-be-processed corpus issued by a learning object; a determining unit configured to determine interactive corpus information based on the to-be-processed corpus and a pre-constructed vector database; an obtaining unit configured to send the to-be-processed corpus and the interactive corpus information to a large language model to obtain courseware information and / or interactive information output by the large language model; a sending unit configured to send the courseware information to an educational digital person having an educational role to enable the educational digital person to interact with the learning object; an interaction unit configured to send the interactive information to a partner digital person having a partner role to enable the partner digital person to interact with the learning object; and an obtaining unit configured to obtain object behavior data of the learning object, to obtain habit information and preference information of the learning object based on the object behavior data, to determine learning result information of the learning object based on the habit information and the preference information, and to provide the learning result information to the learning object. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ Determine performance information of each of the digital humans in the partner digital humans based on the learning result information, and send the corresponding performance information to each of the digital humans in the partner digital humans; wherein the digital humans of different roles include: a good student AI role, a bad student AI role, a neutral student AI role, and a digital human for increasing interest in class; the performance information of the partner digital human of the bad student AI role includes: showing bad learning habits and emotional response information, for making the education learner avoid similar behaviors, and the performance information of the digital human for increasing interest in class includes: emotional support information for showing empathy and emotional support, and game-based learning information for introducing game elements and challenges to make the learning process more lively and interesting.

11. An information interaction system, the system comprising: The digital human interaction module, the educational digital human with an educational role, and the partner digital human with a partner role; The digital human interaction module is configured to receive the to-be-processed corpus issued by the learning object. Determine interaction corpus information based on the to-be-processed corpus and a pre-constructed vector database, send the to-be-processed corpus and the interaction corpus information to a large language model to obtain courseware information and interaction information output by the large language model, send the courseware information to the educational digital human, and send the interaction information to the partner digital human. Obtain object behavior data of the learning object. Determine habit information and preference information of the learning object based on the object behavior data. Determine learning result information of the learning object based on the habit information and the preference information. Determine performance information of each of the digital humans in the partner digital humans based on the learning result information, and send the corresponding performance information to each of the digital humans in the partner digital humans; wherein the digital humans of different roles include: a good student AI role, a bad student AI role, a neutral student AI role, and a digital human for increasing interest in class; the performance information of the partner digital human of the bad student AI role includes: showing bad learning habits and emotional response information, for making the education learner avoid similar behaviors, and the performance information of the digital human for increasing interest in class includes: emotional support information for showing empathy and emotional support, and game-based learning information for introducing game elements and challenges to make the learning process more lively and interesting. The educational digital human generates teaching explanation information based on the courseware information, provides the teaching explanation information to the learning object, and receives a class question of the learning object to generate a class answer to the class question. The partner digital human generates learning corpus information based on the interaction information and the teaching explanation information, and provides the learning corpus information to the learning object.

12. An electronic device, comprising: Comprise: At least one processor; and a memory connected in communication with the at least one processor; Wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method of any one of claims 1-9.

13. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are for causing the computer to perform the method of any one of claims 1-9.

Images