A method and related device for presenting a topic, a learning machine and a storage medium

Abstract

The application discloses a question-answering interaction method and related device, a learning machine and a storage medium, wherein the question-answering interaction method comprises the following steps: driving a virtual teacher to explain a topic title; determining whether a question instruction is detected; in response to detecting the question instruction, driving the virtual teacher to pause the current explanation, waiting for receiving a question sentence until determining that the question sentence is inputted completely, and obtaining an answer sentence generated by a large language model based on at least the question sentence; driving the virtual teacher to broadcast the answer sentence to the question sentence; in response to detecting that the interaction of question and answer between the virtual teacher is completed, obtaining new explanation content generated by the large language model based on the interaction of question and answer; wherein the new explanation content starts to re-explain the topic title from the interruption of the explanation before the interaction of question and answer; and driving the virtual teacher to broadcast the new explanation content. The above scheme can improve the interactivity and feedback timeliness as much as possible, especially in the primary school mathematics teaching scene.

Description

Technical Field

[0001] This application relates to the field of speech synthesis technology, and in particular to a lecture-and-interaction method and related devices, learning machines and storage media. Background Technology

[0002] In the current teaching context, especially in primary school mathematics, there is a widespread demand for efficient and interactive learning tools.

[0003] Currently, especially when students struggle with problems, traditional teaching methods such as video explanations and pure text analyses, while effective to some extent, often lack interactivity and timely feedback, severely limiting students' learning outcomes and the development of their interest. Therefore, how to maximize interactivity and timely feedback, particularly in elementary school mathematics teaching, has become an urgent issue to address. Summary of the Invention

[0004] The main technical problem addressed by this application is to provide a question-and-answer interactive method and related devices, learning machines, and storage media that can maximize interactivity and timely feedback, especially in elementary school mathematics teaching scenarios.

[0005] To address the aforementioned technical problems, the first aspect of this application provides a method for interactive question-and-answer sessions, comprising: driving a virtual teacher to explain a target question; determining whether a question instruction has been detected; in response to detecting a question instruction, driving the virtual teacher to pause the current explanation and wait to receive a question statement until it is determined that the question statement has been input; and obtaining a large language model generating at least an answer statement based on the question statement; driving the virtual teacher to broadcast the answer statement to the question statement; in response to detecting that the interactive question-and-answer session with the virtual teacher has ended, obtaining new explanation content generated by the large language model based on the interactive question-and-answer session, wherein the new explanation content re-explains the target question from the point where the explanation was interrupted before the interactive question-and-answer session; and driving the virtual teacher to broadcast the new explanation content.

[0006] To address the aforementioned technical problems, a second aspect of this application provides a question-and-answer interaction device, comprising: a first driving module, a first judgment module, a first waiting module, a second driving module, an explanation acquisition module, and a third driving module. The first driving module drives a virtual teacher to explain a target question. The first judgment module determines whether a question instruction is detected. The first waiting module, in response to the detection of a question instruction, drives the virtual teacher to pause the current explanation and wait to receive a question statement until the question statement is judged to be input completely, and acquires an answer statement generated by a large language model based on at least the question statement. The second driving module drives the virtual teacher to broadcast the answer statement to the question statement. The explanation acquisition module, in response to the detection that the interaction between the virtual teacher and the teacher has ended, acquires new explanation content generated by the large language model based on the interaction, wherein the new explanation content re-explains the target question from the point where the explanation was interrupted before the interaction. The third driving module drives the virtual teacher to broadcast the new explanation content.

[0007] To address the aforementioned technical problems, a third aspect of this application provides an electronic device comprising at least a memory and a processor coupled to each other, wherein the memory stores program instructions and the processor executes the program instructions to implement the interactive lecture method described in the first aspect.

[0008] To solve the above-mentioned technical problems, the fourth aspect of this application provides a learning machine, including a display device, a speaker device, a pickup device, and a control device, wherein the display device, the speaker device, and the pickup device are electrically connected to the control device, and the control device is the electronic device described in the third aspect above.

[0009] To address the aforementioned technical problems, the fifth aspect of this application provides a computer-readable storage medium storing program instructions executable by a processor, the program instructions being used to implement the interactive lecture method of the first aspect described above.

[0010] The above scheme drives a virtual teacher to explain the target problem and determines whether a question command is detected. Upon detecting a question command, the virtual teacher pauses its current explanation and waits to receive the question until the question is input. It also obtains at least one answer statement generated based on the question, and then broadcasts the answer. Afterwards, upon detecting the completion of the Q&A session with the virtual teacher, a large language model generates new explanation content based on the Q&A. This new explanation resumes the explanation of the target problem from the point of interruption before the Q&A session, driving the virtual teacher to broadcast the new content. This allows for interactive Q&A with the virtual teacher regarding points of confusion, significantly improving interactivity compared to traditional teaching. Furthermore, if a question command is detected during the virtual teacher's explanation, the explanation is paused, and the virtual teacher answers the question, ensuring timely answers to any questions that arise during the explanation, maximizing feedback timeliness. The generation of new explanation content based on the Q&A session allows for timely adjustments to the explanation. Therefore, it maximizes interactivity and timely feedback, especially in elementary school mathematics teaching scenarios. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating an embodiment of the interactive question-and-answer method of this application;

[0012] Figure 2 This is a schematic diagram of a process of an embodiment of the interactive question-and-answer device of this application;

[0013] Figure 3 This is a schematic diagram of the framework of an embodiment of the electronic device of this application;

[0014] Figure 4 This is a schematic diagram of the framework of an embodiment of the learning machine of this application;

[0015] Figure 5 This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium of this application. Detailed Implementation

[0016] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0017] In the following description, specific details such as particular system architectures, interfaces, and technologies are presented for illustrative purposes rather than for limiting purposes, in order to provide a thorough understanding of this application.

[0018] In this paper, the terms "system" and "network" are often used interchangeably. The term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the slash " / " generally indicates that the preceding and following related objects have an "or" relationship. Furthermore, "many" in this paper indicates two or more objects.

[0019] Please see Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the interactive question-and-answer method of this application.

[0020] Specifically, this may include the following steps:

[0021] Step S11: Drive the virtual teacher to explain the target question.

[0022] In one implementation scenario, the target question can come from a question bank. For example, the target question could come from a built-in question bank, or it could come from an online question bank; this is not limited to any particular source. In this case, the user can freely select a question from the question bank to obtain the target question. Alternatively, the target question can also be entered manually, by scanning, or other methods. Taking obtaining the target question by scanning as an example, one can scan questions on paper materials (such as textbooks or exam papers) and perform recognition to obtain the target question. Other cases can be deduced similarly; the methods for obtaining the target question will not be listed here.

[0023] In one implementation scenario, after obtaining the target question, the answer to the target question can be further obtained. Based on at least one of the target question and the answer, the required target knowledge points for solving the problem can be analyzed. Then, based on these target knowledge points, explanation content for the target question can be generated. On this basis, a virtual teacher can be driven to broadcast the generated explanation content, enabling interaction between the virtual teacher and the user. It should be noted that the driving process of the virtual teacher can be found in technical details such as 3D modeling and voice-driven processes, which will not be elaborated upon here.

[0024] In one implementation scenario, a virtual teacher may include, but is not limited to, cartoon characters, etc. The specific content of the virtual teacher is not limited here.

[0025] In a specific implementation scenario, taking the target question as an example (where each question in the question bank has a corresponding answer), the answer to the target question can be directly obtained from the question bank. Alternatively, taking the target question as an example (where it is entered manually or by scanning), as a possible scenario, the target question may not have a corresponding answer. In this case, the target question can be first input into a question-answering model such as a large language model to obtain the answer output by the question-answering model. Taking a large language model as the question-answering model as an example, a prompt text can be constructed based on the target question, such as "Please answer the following questions and provide the solution steps [specific content of the target question]", etc. The specific content of the prompt text is not limited here. Then, the prompt text is input into the large language model, and the answer to the target question can be obtained by utilizing the general understanding ability of the large language model even without a corresponding answer. It should be noted that large language models can include, but are not limited to, open-source large models such as LLAMA and Bloom; or, large language models can be obtained by fine-tuning the parameters of open-source large models based on specific training data; or, large language models can be custom-designed models. The specific source of large language models is not limited here.

[0026] In a specific implementation scenario, a knowledge analysis model, such as a large language model, can be used to analyze at least one of the target question and the answer to obtain the target knowledge points required to solve the problem. Taking the large language model as an example, prompt text can be constructed based on the target question and the answer, such as "Please analyze the knowledge points required to solve the problem based on the following questions and answers [specific content of the target question] [specific content of the answer]", etc. This leverages the general understanding capabilities of the large language model to obtain the target knowledge points needed to solve the problem. It should be noted that the specific source of the large language model can be found in the aforementioned descriptions, and will not be repeated here.

[0027] In a specific implementation scenario, after obtaining the target knowledge point, a generative explanation model, such as a large language model, can be used to combine the target question and its solution to generate the explanation content for the target question. Taking the large language model as an example, prompt text can be constructed based on the target question, the solution, and the target knowledge point, such as "The question [specific content of the target question] examines the [target knowledge point], and its answer is [specific content of the solution]; please explain this question in a simple and easy-to-understand way." This utilizes the general understanding capabilities of the large language model to obtain the explanation content for the target question. As a possible example, the prompt text can also specify an outline for the explanation content, such as "Read the question—Analyze—Answer—Summary," to generate the explanation content according to the target outline. Of course, the above example is only one possible example of an outline and does not limit the specific content of the outline. It should be noted that the specific source of the large language model can be found in the aforementioned descriptions, which will not be repeated here.

[0028] In one implementation scenario, key information from the target question, determined through analysis, can be acquired and displayed on the interactive interface. Furthermore, key information is highlighted using preset markers within the interface. This approach, by displaying the target question on the interactive interface while simultaneously highlighting key information with preset markers, guides user thinking during the explanation process, thereby enhancing the effectiveness of the explanation.

[0029] In a specific implementation scenario, as mentioned earlier, the explanation content can be generated according to a defined outline. Taking an outline that includes "reading the question—analyzing—answering—summarizing" as an example, the target question can be displayed on the interactive interface from beginning to end. For example, in the "reading the question" stage, the interactive interface can display the target question in its entirety, highlighting key information with preset markers; or, in the "analyzing," "answering," and "summarizing" stages following the "reading the question" stage, the interactive interface can simultaneously display the question area and the whiteboard area, displaying the target question in the question area with preset markers highlighting key information, while the whiteboard area can display at least some of the explanation statements from the explanation content.

[0030] In a specific implementation scenario, as a possible interactive example, during the "answering questions" phase, the left side of the interactive interface can display the target question, and the right side can display a virtual teacher. Furthermore, preset markers such as boxes can be used to highlight key information in the target question. Of course, the specific style of these preset markers is not limited to this. For example, preset markers can also include, but are not limited to, text background highlighting, underlining, and bolding; these are not limited here. The dashed box on the left side of the interactive interface represents the question area, where the target question can be displayed, while the right side is the whiteboard area, where at least some of the explanation statements can be displayed. In addition, the virtual teacher can be overlaid on the whiteboard area; or overlaid on the question area; or displayed in any area of ​​the interactive interface that does not overlap with either the whiteboard area or the question area; these are not limited here. Of course, the above examples are merely possible layout examples of the interactive interface in practical applications and do not limit the specific layout of the interactive interface.

[0031] In a specific implementation scenario, taking the target question "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example, the key information "39 trees" and "4 more than two-thirds" can be extracted and highlighted with preset markers in the interactive interface. Of course, the above example is only one possible example. If the target question is another question, the same logic can be applied. Further examples will not be provided here.

[0032] Step S12: Determine whether a question command has been detected.

[0033] In one implementation scenario, as a possible interactive example, the interface interacting with the virtual teacher can display a first question control. It can then determine whether a trigger command for the first question control has been detected, and in response, generate a question command. This method, by independently setting the first question control in the interactive interface, facilitates asking questions at any time during the explanation process by triggering the first question control, thereby further improving the timeliness of feedback.

[0034] In another implementation scenario, as another possible example of interaction, if the virtual teacher's explanation of the target topic is at least partially displayed on the interactive interface, a second question control can be displayed on the interface based on the selection of the explanation statement. This allows for the generation of a question command in response to the detection of a trigger command on the second question control. This method, which displays the second question control on the interactive interface when any explanation statement is selected, facilitates asking questions about the explanation statement during the explanation process, thereby improving the relevance of the questions.

[0035] It should be noted that the above examples are merely a few possible instances of generating question commands in practical applications, and do not limit other possible ways of generating question commands.

[0036] Step S13: In response to the detection of a question instruction, drive the virtual teacher to pause the current explanation and wait to receive the question statement until it is determined that the question statement has been input, and obtain the answer statement generated by the large language model based on the question statement.

[0037] In one implementation scenario, upon detecting a question instruction, the current explanation can be paused and recorded so that it can be resumed from the current point later. For example, taking the target question "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example, the explanation could include, "First, we need to calculate what two-thirds of the number of trees Class 6(1) planted is. 39 multiplied by two-thirds equals what? That's right, it should be 26…". At this point, a question instruction can be generated by triggering either the first or second question control, allowing the explanation to pause at "That's right, it should be 26", meaning the current explanation is "That's right, it should be 26". Of course, other situations can be deduced similarly, and will not be listed here.

[0038] In one implementation scenario, after detecting a question command, the system can wait for the user to input the question via voice, typing, or other methods until the question is considered complete. As a possible interaction example, the user-inputted question can be displayed on the interface. After confirming its accuracy, the user can click the "Send" button, thus confirming the completion of the question input. Taking voice input as an example, the user's voice can be recognized, and the recognized text can be displayed in real-time on the interface. After confirming its accuracy, the user can click the "Send" button. Of course, the above examples are merely one possible scenario in practical applications and do not limit other possible methods. For instance, if no new input is detected within a preset time (e.g., 5 seconds, 6 seconds), the completion of the question input can be confirmed; or, if a preset reminder word indicating the end of the question (e.g., "above," "over") is detected, the completion of the question input can be confirmed. The specific method for confirming whether the question input is complete is not limited here.

[0039] In one implementation scenario, after receiving a question, an answer can be generated based on at least the question. For example, the question can be processed using a question-and-answer model such as a large language model to obtain an answer. Alternatively, as another possible implementation, the answer can be generated based on the question and relevant statements in the current explanation. For example, the question and relevant content in the current explanation can be input into a question-and-answer model such as a large oracle model to obtain an answer. Specifically, the relevant content in the current explanation may include, but is not limited to, explanation statements already broadcast before the current explanation; no specific limitation is made on the relevant content here. Taking the target question "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example, the question could be "Why multiply 39 by two-thirds?", which would generate the answer "Because the question states that Class 6(2) planted 4 more trees than two-thirds of Class 6(1). We need to first calculate what two-thirds of Class 6(1) is, and then add 4 more trees to get the total number of trees planted by Class 6(2)." Of course, the above example is only one possible example in actual application. Other situations can be deduced by analogy, and will not be listed here.

[0040] Step S14: Drive the virtual teacher to read the answer to the question.

[0041] In one implementation scenario, after receiving the answer statement, the virtual teacher can be driven to read and recite the answer to the question. Furthermore, as a possible implementation, while the virtual teacher is reading and reciting the answer statement, the virtual teacher's interface also displays the answer statement. As a possible interactive example, while the virtual teacher is reading and reciting the answer statement, the virtual teacher's interface also displays at least a first sub-interface for displaying the answer statement. Based on this, after the virtual teacher reads and recites the answer to the question, the virtual teacher can be driven to resume the explanation from the current point. Taking the aforementioned question, "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example, we can resume the explanation from the current point, "That's right, it should be 26." For example, we can continue explaining, "Since Class 6(2) planted 4 more trees than two-thirds of Class 6(1), the number of trees planted by Class 6(2) should be 26 plus 4 equals 30." Of course, the above example is only one possible case in actual application and does not limit other possible cases. We will not give examples of each case here.

[0042] In another implementation scenario, differing from the aforementioned implementation, as another possible approach, after the virtual teacher reads the answer to the question, it can first determine whether a new question instruction has been detected. A new question instruction represents a follow-up question to the latest answer. If no new question instruction is detected, it can be determined that the interactive Q&A with the virtual teacher has ended; if a new question instruction is detected, it can be determined that the interactive Q&A with the virtual teacher has not ended. This method, by determining whether a new question instruction has been detected before resuming the explanation, can further enhance interactivity.

[0043] In a specific implementation scenario, as a possible interactive example, when driving a virtual teacher to read and answer questions, the interactive interface with the virtual teacher can also display at least a first sub-interface for displaying the answers. This first sub-interface further includes a first follow-up question control indicating continued questioning and a second follow-up question control indicating the end of follow-up questioning. Based on this, it can be determined whether a trigger command for the first or second follow-up question control has been detected. In response to detecting a trigger command for the first follow-up question control, it can be determined that the Q&A interaction with the virtual teacher is complete, and a follow-up question command is generated to await new questions. Similarly, in response to detecting a trigger command for the second follow-up question control, it can be determined that the Q&A interaction with the virtual teacher is complete. This method, by setting follow-up question controls in the interactive interface, greatly improves the convenience of asking follow-up questions.

[0044] In a specific implementation scenario, if the interaction between the user and the virtual teacher is not yet complete, the system can wait to receive a new question until it is determined that the new question has been input. The system then obtains a new answer generated by the large language model based on at least the new question and the latest answer. This allows the virtual teacher to read the new answer, and the system re-detects whether the interaction between the user and the virtual teacher is complete, until it is confirmed that the interaction is finished. For example, a prompt text can be constructed based on the new question and the latest answer, such as "Please try to answer the question [the specific content of the new question] based on the previous answer [the specific content of the latest answer]," and this prompt text is input into the large language model to utilize its general understanding capabilities to generate the new answer. As a possible example, let's take the aforementioned target question, "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example. The answer statement, "Because the question states that Class 6(2) planted 4 more trees than two-thirds of Class 6(1), we need to first calculate what two-thirds of Class 6(1) is, and then add 4 more trees to get the total number of trees planted by Class 6(2)," can be updated if a new question instruction is detected. The system waits to receive new questions until it determines that the new question has been entered. For example, if a new question is received, such as "So how do you calculate two-thirds of 39?", a new answer can be generated: "Two-thirds of 39 means dividing 39 into three equal parts and taking two of them. Therefore, to calculate two-thirds of 39, first divide 39 by 3, which gives 13, meaning each part is 13. Then multiply 13 by 2, which gives 26. In other words, two-thirds of 39 equals 26." If the user has no further questions, a second follow-up question control can be triggered to end the follow-up question. Of course, if the user still has questions, the first follow-up question control can be triggered to continue the follow-up question. This method, upon detecting a new question, waits for the user to input a new question and obtains a corresponding new answer to drive the virtual teacher to read the new answer. It then returns to the step of determining whether a new question has been detected, continuing until the follow-up question is confirmed to end. Therefore, it supports multiple rounds of interaction for questions, helping to further improve interactivity.

[0045] Step S15: In response to the detection that the interactive question-and-answer session with the virtual teacher has ended, obtain the new explanation content generated by the large language model based on the interactive question-and-answer session.

[0046] In this embodiment of the disclosure, the new explanation content starts from the point where the explanation was interrupted before the interactive Q&A session and re-explains the target question.

[0047] In one implementation scenario, prompt text can be constructed based on at least the interactive question-and-answer format. This prompt text instructs the large language model to generate new narration content from the point where the narration was interrupted, referencing the interactive question-and-answer format. It should be noted that the specific selection of the large language model can be found in the aforementioned descriptions, and will not be repeated here. Based on this, the prompt text can be sent to the large language model, and the new narration content generated by the large language model can be retrieved.

[0048] In one implementation scenario, let's take the aforementioned target question, "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example. If the answer to the question, "What method can we use to find the number of trees planted by Class 6(2)?" is correct, such as "finding the product of an integer and a fraction," then the explanation can be adjusted to something like, "That's right, therefore, based on the fact that Class 6(2) planted 4 more trees than two-thirds of Class 6(1), we can first find that two-thirds of the number of trees planted by Class 6(1) is 39." *(2 / 3) = 26, then add 4 to 26 to get 30, which is the number of trees planted by Class 6(2)”; or, if the answer to the question “What method can be used to find the number of trees planted by Class 6(2)” is incorrect, such as “finding the quotient of two numbers”, then the explanation can be adjusted to new content, such as “Think again, the question says that the number of trees planted by Class 6(2) is 4 more than two-thirds of the number planted by Class 6(1), so is two-thirds of the number planted by Class 6(1) multiplied by two-thirds, or is two-thirds of the number planted by Class 6(1) divided by two-thirds”. Of course, the above example is only one possible example in actual application and does not limit other possible methods.

[0049] It should be noted that, as another possible implementation, in practical applications, the virtual teacher can also be driven to deliver lectures according to a script for the target topic. That is, unlike the aforementioned implementation, the virtual teacher explains the target topic according to a fixed script. In this case, in response to the detection that the interactive Q&A with the virtual teacher has ended, the virtual teacher can be driven to resume explaining the target topic from the point of interruption in the lecture script before the interactive Q&A. Taking the aforementioned target question, "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example, the script could include: "How can we solve the problem of finding the number of trees planted by Class 6(2)? We can use the product of an integer and a fraction to solve this problem. Since Class 6(2) planted 4 more trees than two-thirds of Class 6(1), we can first find that two-thirds of the number of trees planted by Class 6(1) is 39 * (2 / 3) = 26. Then we add 4 to 26 to get 30, which is the number of trees planted by Class 6(2)." If the question "What method can be used to solve the problem of finding the number of trees planted by Class 6(2)?" is answered correctly, such as "finding the product of an integer and a fraction", or incorrectly, such as "finding the quotient of two numbers", then the explanation can continue from where it was interrupted, following the original script: "We can use the product of an integer and a fraction to solve this problem. Since the number of trees planted by Class 6(2) is 4 more than two-thirds of the number planted by Class 6(1), we can first find that two-thirds of the number of trees planted by Class 6(1) is 39*(2 / 3) = 26. Then we add 4 to 26 to get 30, which is the number of trees planted by Class 6(2)." Of course, the above example is only one possible example in actual application and does not limit other possible methods. In this method, the virtual teacher is driven to deliver a lecture according to the script for the target question. In response to the detection that the interaction and Q&A with the virtual teacher has ended, the virtual teacher is driven to resume the lecture from the point in the script where the Q&A was interrupted. Compared with generating new lecture content in real time, this method can greatly reduce the consumption of computing resources, network resources and other resources.

[0050] Furthermore, the interactive question-and-answer format described in this embodiment can be either a virtual teacher asking a question and the user answering it, or a user asking a question and the virtual teacher answering it, or a combination of both; no limitation is made here. Additionally, the specific method of interactive question-and-answer can include at least one of text, voice, handwriting, and drawing / circling formats; that is, any one or a combination of these methods can be used, without limitation.

[0051] Step S16: Drive the virtual teacher to broadcast new explanation content.

[0052] In one implementation scenario, during the explanation, the virtual teacher's pace can be followed, with the indicated portion of the explanation content displayed synchronously in the whiteboard area of ​​the interactive interface. It should be noted that, to improve interactive comfort, the virtual teacher's presentation can include not only content substantively related to the topic but also filler words and other content without substantial meaning or unrelated to the topic. The whiteboard area, however, can only display content substantively related to the explanation. This method, by synchronously displaying at least a portion of the explanation content in the whiteboard area of ​​the interactive interface, following the virtual teacher's pace, improves the consistency between the audio explanation and the whiteboard display.

[0053] In one implementation scenario, during the explanation of a problem, questions generated based on reference information can be obtained. This reference information includes at least one of the following: key information from the target problem, or key steps in the solution to the target problem. Based on this, the questions generated from the reference information can be output and prompts for answers. This method, by generating questions from reference information and then outputting and prompting for answers, proactively asks questions of the user, which helps to further enhance interactivity.

[0054] In a specific implementation scenario, the methods for obtaining key information can be found in the relevant descriptions, which will not be repeated here.

[0055] In a specific implementation scenario, the key steps in the solution can be determined by combining the target knowledge points. The methods for obtaining the target knowledge points can be found in the aforementioned descriptions and will not be repeated here. After obtaining the target knowledge points, a step analysis model, such as a large language model, can be used to process the target knowledge points and the solution to determine the key steps in the solution. For example, taking a large language model as the step analysis model, a prompt text can be constructed based on the target knowledge points and the solution, such as "The answer to the question is [the specific content of the solution], the knowledge point tested in the question is [the target knowledge point], please determine the key steps in the answer." This prompt text can then be input into the large language model to obtain the key steps determined by the large language model. It should be noted that the specific source of the large language model can be found in the aforementioned descriptions and will not be repeated here. Taking the target question "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example, the target knowledge point is "multiplication of fractions and integers", and the answer is "39 ​​* (2 / 3) = 26 (trees), 26 + 4 = 30 (trees), therefore Class 6(2) planted 30 trees". Therefore, the key step is "39 ​​* (2 / 3) = 26 (trees)". Of course, the above example is only one possible example in actual application. Other situations can be deduced by analogy, and will not be listed here.

[0056] In a specific implementation scenario, taking the aforementioned target question "Class 6(1) planted 39 trees, and Class 6(2) planted 4 more trees than two-thirds of Class 6(1). How many trees did Class 6(2) plant?" as an example, by combining the key information and the key steps in the solution, the question statement "What method can be used to find the number of trees planted by Class 6(2)?" can be generated. In addition to generating the question statement, several options for the question statement can also be generated and output along with the question statement, with a prompt to choose from several options. Taking the aforementioned target question as an example, for the question statement "What method can be used to find the number of trees planted by Class 6(2)?", several options can also be generated such as "find the product of an integer and a fraction", "find the quotient of two numbers", and "don't know". Of course, the above example is only one possible example, and other situations can be deduced by analogy, which will not be listed here.

[0057] In a specific implementation scenario, the timing of actively generating and outputting a question statement based on reference information, along with prompts for answering the question statement, can be any time. Specifically, it can be done by actively asking the question after the virtual teacher reads the question, and then analyzing it. For example, the question statement "What method can be used to find the number of trees planted by Class 6(2)?" can be output after the virtual teacher reads the target question. Alternatively, the question statement can be actively output during the "analysis" stage. For example, for the question statement "Because the question says that the number of trees planted by Class 6(2) is 4 more than two-thirds of the number planted by Class 6(1). We need to first calculate what two-thirds of the number planted by Class 6(1) is, and then add 4 more trees to get the total number of trees planted by Class 6(2)", if no follow-up question is detected, the question statement "So think about it, what should the total number of trees planted by Class 6(2) be?" can be actively generated and output. Of course, the above examples are just a few possible examples of the timing of actively asking questions, and the timing of occurrence will not be listed one by one here.

[0058] In a specific implementation scenario, after obtaining the question statement generated based on reference information, a virtual teacher can be driven to read and recite the question statement, thus outputting and prompting answers. Alternatively, as a possible interactive example, upon obtaining the question statement generated based on reference information, a second sub-interface can be displayed on the interactive interface, showing the question statement generated based on reference information. Furthermore, both methods can be implemented simultaneously. Of course, the above examples are merely one possible scenario in practical applications, and will not be elaborated upon further. The above methods, by driving the virtual teacher to read and recite the question statement generated based on reference information, and by displaying and prompting answers to the question statement on the interactive interface upon obtaining it, further enhance interactivity.

[0059] In practical applications, if no questioning instruction is detected, step S17 can be executed, i.e., continue executing step S11. In other words, the process of driving the virtual teacher to explain the target question can proceed uninterrupted even if no questioning instruction is detected.

[0060] Step S17: Continue with step S11.

[0061] In one implementation scenario, as mentioned earlier, the explanation content can be generated based on the outline. After the "analysis" stage is completed, the virtual teacher can be driven to explain the content of the "answer" stage. Of course, the answer steps can also be displayed synchronously in the whiteboard area during this process. Taking the aforementioned target question as an example, the virtual teacher can be driven to explain the following content of the "answer" stage: "39*(2 / 3)=26 (trees), 26+4=30 (trees), Answer: Class 6(2) planted 30 trees." Of course, the above example is only a possible example in actual application. Other situations can be deduced by analogy, and will not be listed here.

[0062] In one implementation scenario, as mentioned earlier, the explanation content can be generated based on the outline. After the "answer" stage is completed, the virtual teacher can be driven to explain the "summary" stage content. Of course, the summary content can also be displayed synchronously in the whiteboard area during this process. Taking the aforementioned target question as an example, the virtual teacher can be driven to explain the following "summary" stage content: "This question allows us to practice fraction multiplication and simple addition. First, find the known quantities: the number of trees planted by Class 6(1) and the number of trees that Class 6(2) has more than Class 6(1). Then, based on the relationship described in the question, calculate part of the quantity using fraction multiplication, and then calculate the final result using addition." Of course, the above example is only a possible example in the actual application process. Other situations can be deduced by analogy, and will not be listed here.

[0063] Therefore, based on the above-described process steps, the interactivity and inspiration of learning can be greatly enhanced. Furthermore, the support for instant feedback and user questions significantly improves learning effectiveness and motivation. Compared to traditional teaching methods and existing AI teaching systems, this embodiment of the disclosure is more effective in improving learning efficiency, stimulating learning interest, and meeting personalized learning needs. In addition, although this embodiment uses a primary school mathematics teaching scenario as an example, it does not mean that the method and process described in this embodiment can only be used in primary school mathematics teaching scenarios. Depending on the application needs, it can also be applied to teaching scenarios of other subjects such as physics, chemistry, Chinese, and English. Therefore, the application scenario is not limited here.

[0064] The above scheme drives a virtual teacher to explain the target problem and determines whether a question command is detected. Upon detecting a question command, the virtual teacher pauses its current explanation and waits to receive the question until the question is input. It also obtains at least one answer statement generated based on the question, and then broadcasts the answer. Afterwards, upon detecting the completion of the Q&A session with the virtual teacher, a large language model generates new explanation content based on the Q&A. This new explanation resumes the explanation of the target problem from the point of interruption before the Q&A session, driving the virtual teacher to broadcast the new content. This allows for interactive Q&A with the virtual teacher regarding points of confusion, significantly improving interactivity compared to traditional teaching. Furthermore, if a question command is detected during the virtual teacher's explanation, the explanation is paused, and the virtual teacher answers the question, ensuring timely answers to any questions that arise during the explanation, maximizing feedback timeliness. The generation of new explanation content based on the Q&A session allows for timely adjustments to the explanation. Therefore, it maximizes interactivity and timely feedback, especially in elementary school mathematics teaching scenarios.

[0065] Please see Figure 2 , Figure 2This is a schematic diagram of the framework of an embodiment of the interactive question-and-answer device 20 of this application. The interactive question-and-answer device 20 includes: a first driving module 21, a first judging module 22, a first waiting module 23, a second driving module 24, an explanation acquisition module 25, and a third driving module 26. The first driving module 21 is used to drive the virtual teacher to explain the target question; the first judging module 22 is used to judge whether a question instruction is detected; the first waiting module 23 is used to, in response to the detection of a question instruction, drive the virtual teacher to pause the current explanation and wait to receive a question statement until the question statement is judged to be input, and acquire the answer statement generated by the large language model based on the question statement; the second driving module 24 is used to drive the virtual teacher to broadcast the answer statement to the question statement; the explanation acquisition module 25 is used to, in response to the detection that the interactive question and answer with the virtual teacher is completed, acquire the new explanation content generated by the large language model based on the interactive question and answer, wherein the new explanation content re-explains the target question from the point where the explanation was interrupted before the interactive question and answer; the third driving module 26 is used to drive the virtual teacher to broadcast the new explanation content.

[0066] In the above scheme, the interactive question-and-answer device 20 drives the virtual teacher to explain the target problem and determines whether a question command is detected. Upon detecting a question command, the virtual teacher pauses its current explanation and waits to receive a question until the question is input. It also acquires at least one answer statement generated based on the question and then plays the answer statement. Afterwards, upon detecting the completion of the interactive question-and-answer session with the virtual teacher, a large language model generates new explanation content based on the interactive question and answer. This new explanation content re-explains the target problem from the point of interruption before the interactive question and answer session, driving the virtual teacher to play the new explanation content. This allows for interactive question-and-answer sessions with the virtual teacher to address points of confusion, thus significantly improving interactivity compared to traditional teaching. Furthermore, if a question command is detected during the virtual teacher's explanation, the explanation is paused, and the virtual teacher answers the question, ensuring timely answers to any questions that arise during the explanation, maximizing feedback timeliness. The generation of new explanation content based on the interactive question and answer session allows for timely adjustments to the explanation. Therefore, it maximizes interactivity and feedback timeliness, especially in elementary school mathematics teaching scenarios.

[0067] In some disclosed embodiments, the interactive question-and-answer device 20 further includes a second judgment module for judging whether a new question instruction is detected; wherein, the new question instruction represents a follow-up question instruction on the latest answer statement; the interactive question-and-answer device 20 further includes a first determination module for determining that the interactive question-and-answer session with the virtual teacher is completed in response to the absence of the new question instruction; the interactive question-and-answer device 20 further includes a second determination module for determining that the interactive question-and-answer session with the virtual teacher is not completed in response to the detection of the new question instruction.

[0068] In some disclosed embodiments, the interactive question-and-answer device 20 further includes a second waiting module, which waits to receive a new question statement in response to detecting that the interactive question-and-answer session with the virtual teacher is not yet complete, until it is determined that the new question statement has been input, and obtains a large language model to generate a new answer statement based at least on the new question statement and the latest answer statement; the interactive question-and-answer device 20 further includes a fourth driving module, which drives the virtual teacher to broadcast the new answer statement, and re-detects whether the interactive question-and-answer session with the virtual teacher is complete, until it is confirmed that the interactive question-and-answer session is complete.

[0069] In some disclosed embodiments, when the virtual teacher is driving the presentation of Q&A statements, the interactive interface with the virtual teacher also displays at least a first sub-interface for displaying the Q&A statements. This first sub-interface also includes a first follow-up question control indicating continued questioning and a second follow-up question control indicating the end of follow-up questioning. The interactive device 20 further includes a third judgment module for determining whether a trigger command for the first or second follow-up question control is detected. The interactive device 20 also includes a first generation module for determining, in response to detecting a trigger command for the first follow-up question control, that the Q&A interaction with the virtual teacher is not yet complete, and generating a follow-up question command to await receiving new questions. The interactive device 20 also includes a confirmation end module for determining, in response to detecting a trigger command for the second follow-up question control, that the Q&A interaction with the virtual teacher is complete.

[0070] In some publicly available embodiments, the interactive interface with the virtual teacher displays a first question control, and the interactive device 20 further includes a fourth judgment module for judging whether a trigger command for the first question control is detected; the interactive device 20 also includes a second generation module for generating a question command in response to the detection of a trigger command for the first question control.

[0071] In some disclosed embodiments, the explanation content of the virtual teacher explaining the target question is at least partially displayed on the interactive interface. The question-answering interactive device 20 also includes a control display module, which is used to display a second question control on the interactive interface based on the selection instruction of the explanation statement. The question-answering interactive device 20 also includes a third generation module, which is used to generate a question instruction in response to detecting the trigger instruction of the second question control.

[0072] In some disclosed embodiments, the interactive question-and-answer device 20 further includes a statement acquisition module for acquiring question statements generated based on reference information; wherein the reference information includes at least one of the following: key information in the target question, key steps in the solution to the target question; the interactive question-and-answer device 20 further includes an output prompt module for outputting and prompting answers to question statements generated based on reference information.

[0073] In some disclosed embodiments, the output prompt module is specifically used to perform at least one of the following: drive the virtual teacher to broadcast the question statement generated based on the reference information; when the question statement generated based on the reference information is obtained, prompt the second sub-interface on the interactive interface, and display the question statement generated based on the reference information on the second sub-interface.

[0074] In some disclosed embodiments, the interactive device 20 further includes a synchronous display module, which is used to synchronously display at least part of the explanation content in the whiteboard area of ​​the interactive interface, following the explanation progress of the virtual teacher.

[0075] In some disclosed embodiments, the interactive question-and-answer device 20 further includes an information analysis module for acquiring key information in the target question determined by analyzing the target question; the interactive question-and-answer device 20 also includes an information marking module for displaying the target question on the interactive interface and highlighting key information on the interactive interface with preset markings.

[0076] In some disclosed embodiments, the explanation acquisition module includes a construction submodule for constructing prompt text based at least on interactive question and answer, the prompt text instructing the large language model to generate new explanation content from the point of interruption in problem-solving based on the interactive question and answer, and the explanation acquisition module includes a sending and receiving submodule for sending the prompt text to the large language model and acquiring the new explanation content generated by the large language model.

[0077] Please see Figure 3 , Figure 3 This is a schematic diagram of an embodiment of the electronic device 30 of this application. The electronic device 30 includes a memory 31 and a processor 32 coupled to each other. The memory 31 stores program instructions, and the processor 32 is used to execute the program instructions to implement the steps in any of the above-described interactive lecture method embodiments. For details, please refer to the foregoing disclosed embodiments, which will not be repeated here.

[0078] Specifically, processor 32 controls itself and memory 31 to implement the steps in any of the above-described interactive lecture method embodiments. Processor 32 can also be referred to as a CPU (Central Processing Unit). Processor 32 may be an integrated circuit chip with signal processing capabilities. Processor 32 can also be a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor can be a microprocessor or any conventional processor. Furthermore, processor 32 can be implemented using integrated circuit chips.

[0079] In the above scheme, the electronic device 30 drives the virtual teacher to explain the target problem and determines whether a question command is detected. Upon detecting a question command, the virtual teacher pauses its current explanation and waits to receive the question until the question is input. It also acquires at least one answer statement generated based on the question and then plays the answer statement. Afterwards, upon detecting the completion of the interactive Q&A with the virtual teacher, a large language model generates new explanation content based on the interactive Q&A. This new explanation content re-explains the target problem from the point of interruption before the interactive Q&A, driving the virtual teacher to play the new explanation. This allows for interactive Q&A with the virtual teacher regarding points of confusion, thus significantly improving interactivity compared to traditional teaching. Furthermore, if a question command is detected during the virtual teacher's explanation, the explanation is paused, and the virtual teacher answers the question, ensuring timely answers to any questions that arise during the explanation, maximizing feedback timeliness. The generation of new explanation content based on the interactive Q&A allows for timely adjustments to the explanation. Therefore, it maximizes interactivity and feedback timeliness, especially in elementary school mathematics teaching scenarios.

[0080] Please see Figure 4 , Figure 4 This is a schematic diagram of the framework of an embodiment of the learning machine 40 of this application. The learning machine 40 includes: a display device 41, a speaker device 42, a pickup device 43, and a control device 44. The display device 41, the speaker device 42, and the pickup device 43 are electrically connected to the control device 44, and the control device 44 is the electronic device in the above-described electronic device embodiment. For details, please refer to the aforementioned electronic device embodiment, which will not be repeated here.

[0081] In one implementation scenario, the display device 41 can be used to display an interactive interface. The specific content of the interactive interface can be found in the relevant descriptions in the foregoing disclosed embodiments, and will not be repeated here. Furthermore, if the display device 41 is a touch screen, it can also receive touch commands such as touch clicks to interact with the user.

[0082] In one implementation scenario, the speaker 42 can be used to play the virtual teacher's audio and other audio that interacts with the user (e.g., audio expressing encouragement, praise, etc.).

[0083] In one implementation scenario, the microphone 43 can be used to receive voice input, as detailed in the relevant descriptions in the foregoing disclosed embodiments, which will not be repeated here.

[0084] In one implementation scenario, the learning machine 40 may also include, but is not limited to, a camera device (not shown), and sensing devices such as light sensors and distance sensors (not shown). For example, the camera device can collect information such as the user's facial expressions and movements to analyze the learning state (e.g., fatigue, excitement) and then use this information to provide targeted interaction. Alternatively, the sensing device can collect information such as ambient light and student posture to provide relevant prompts (e.g., "Poor posture detected, please adjust your posture," "Dim light detected, please increase ambient brightness to protect your eyesight," etc.). Of course, the above examples are just some of the devices that the learning machine 40 may integrate; we will not list all the electronic devices that can be integrated into the learning machine 40 here.

[0085] The learning machine 40 in the above solution includes a display device 41, a speaker 42, a microphone 43, and a control device 44. The display device 41, speaker 42, and microphone 43 are electrically connected to the control device 44, and the control device 44 is the electronic device in the above-described electronic device embodiment. Therefore, compared with traditional teaching, it can greatly improve interactivity. Furthermore, if a question is detected during the virtual teacher's explanation, the explanation is paused, and the virtual teacher answers the question, allowing for timely responses to questions arising during the explanation process. This helps to maximize the timeliness of feedback. New explanation content is generated based on the interactive questions and answers, allowing for timely adjustments to the explanation. Therefore, it can maximize interactivity and timely feedback, especially in elementary school mathematics teaching scenarios.

[0086] Please see Figure 5 , Figure 5 This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium 50 of this application. The computer-readable storage medium 50 stores program instructions 51 that can be executed by a processor. The program instructions 51 are used to implement the steps in any of the above-described interactive lecture method embodiments.

[0087] The above scheme involves a computer-readable storage medium 50 driving a virtual teacher to explain a target problem and determining whether a question command has been detected. In response to a detected question command, the virtual teacher pauses its current explanation and waits to receive a question until the question is input. It also acquires at least one answer statement generated based on the question, and then drives the virtual teacher to read the answer. Afterwards, in response to the completion of the interactive Q&A with the virtual teacher, a large language model generates new explanation content based on the interactive Q&A. This new explanation content re-explains the target problem from the point of interruption before the interactive Q&A, driving the virtual teacher to read the new content. This allows for interactive Q&A with the virtual teacher regarding points of confusion, significantly improving interactivity compared to traditional teaching. Furthermore, if a question command is detected during the virtual teacher's explanation, the explanation is paused, and the virtual teacher answers the question, ensuring timely answers to any questions that arise during the explanation, maximizing feedback timeliness. The generation of new explanation content based on the interactive Q&A allows for timely adjustments to the explanation. Therefore, it maximizes interactivity and feedback timeliness, especially in elementary school mathematics teaching scenarios.

[0088] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0089] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0090] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus implementations described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0091] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0092] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0093] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0094] If the technical solution of this application involves personal information, the product using this technical solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using this technical solution has obtained the user's separate consent before processing the sensitive personal information, and also meets the requirement of "express consent". For example, at personal information collection devices such as cameras, clear and prominent signs are set up to inform users that they have entered the scope of personal information collection and that personal information will be collected. If an individual voluntarily enters the collection scope, it is deemed that they have agreed to the collection of their personal information; or on the personal information processing device, with clear signs / information informing users of the personal information processing rules, authorization is obtained from the individual through pop-up information or by asking the individual to upload their personal information; wherein, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed.

Claims

1. A question-and-answer interactive method, characterized in that, include: Drive a virtual teacher to explain the target question; Determine if a question command has been detected; In response to the detection of the question instruction, the virtual teacher is driven to pause the current explanation and wait to receive the question statement until it is determined that the question statement has been input, and to obtain the answer statement generated by the large language model based on the question statement. The virtual teacher is driven to read the answer to the question. In response to the detection that the interactive question-and-answer session with the virtual teacher has ended, the large language model generates new explanation content based on the interactive question-and-answer session; wherein, the new explanation content re-explains the target topic from the point where the explanation was interrupted before the interactive question-and-answer session. The virtual teacher is then driven to deliver the new lecture content.

2. The method according to claim 1, characterized in that, After driving the virtual teacher to broadcast the answer to the question, the method further includes: Determine whether a new question instruction has been detected; wherein, the new question instruction represents a follow-up question instruction on the latest answer statement; In response to the absence of the detected new question instruction, it is determined that the interaction and question-and-answer session with the virtual teacher has ended. In response to the detection of the new question instruction, it is determined that the interactive question-and-answer session with the virtual teacher is not yet complete.

3. The method according to claim 1 or 2, characterized in that, The method further includes: In response to detecting that the interaction between the virtual teacher and the question and answer is not completed, wait to receive a new question statement until it is determined that the new question statement has been input, and obtain the large language model to generate a new answer statement based at least on the new question statement and the latest answer statement; The system drives the virtual teacher to read the new Q&A statement and re-checks whether the Q&A interaction with the virtual teacher is complete until it is confirmed that the Q&A interaction is complete.

4. The method according to claim 1 or 2, characterized in that, When the virtual teacher reads the Q&A statement, the interactive interface with the virtual teacher also displays at least a first sub-interface for displaying the Q&A statement. This first sub-interface also includes a first follow-up question control indicating continued questioning and a second follow-up question control indicating the end of follow-up questioning. The method further includes: Determine whether a trigger command for the first follow-up control or a trigger command for the second follow-up control has been detected; In response to detecting a trigger command for the first follow-up question control, it is determined that the interactive question and answer between the virtual teacher and the virtual teacher is not yet complete, and a follow-up question command is generated to wait for receiving new questions. In response to the detection of a trigger command for the second follow-up question control, it is determined that the interactive question-and-answer session with the virtual teacher has ended.

5. The method according to claim 1, characterized in that, The interaction interface with the virtual teacher displays a first question control, and the method further includes: Determine whether a trigger command for the first question control has been detected; In response to detecting a trigger command for the first question control, the question command is generated.

6. The method according to claim 1, characterized in that, The virtual teacher's explanation of the target question is at least partially displayed on the interactive interface, and the method further includes: Based on the selection command of the explanatory statement, a second question control is displayed on the interactive interface; In response to detecting a trigger command for the second question control, the question command is generated.

7. The method according to claim 1, characterized in that, The method further includes: Obtain a question statement generated based on reference information; wherein the reference information includes at least one of the following: key information in the target question, key steps in the solution to the target question; Output and prompt answers to the questions generated based on the reference information.

8. The method according to claim 7, characterized in that, The output and prompt for answers to the questions generated based on the reference information includes at least one of the following: The virtual teacher is driven to broadcast questions generated based on the reference information; When a question statement generated based on the reference information is obtained, a second sub-interface is prompted on the interactive interface, and the question statement generated based on the reference information is displayed on the second sub-interface.

9. The method according to claim 1, characterized in that, The method further includes: Following the virtual teacher's explanation progress, at least part of the explanation content is simultaneously displayed in the whiteboard area of ​​the interactive interface.

10. The method according to claim 1, characterized in that, The method further includes: Obtain key information from the target question as determined by analysis of the target question; The target question is displayed on the interactive interface, and the key information is highlighted on the interactive interface with preset markers.

11. The method according to claim 1, characterized in that, The process of obtaining new explanatory content generated by the large language model based on the interactive question-and-answer session includes: Based at least on the interactive question and answer, a prompt text is constructed; wherein the prompt text instructs the large language model to generate new explanatory content from the point where the explanation is interrupted, with reference to the interactive question and answer; Send the prompt text to the large language model and obtain the new explanation content generated by the large language model.

12. A question-and-answer interactive device, characterized in that, include: The first driving module is used to drive the virtual teacher to explain the target question; The first judgment module is used to determine whether a question command has been detected. The first waiting module is used to respond to the detection of the question instruction, drive the virtual teacher to pause the current explanation, and wait to receive the question statement until it is determined that the question statement has been input, and obtain the answer statement generated by the large language model based on the question statement. The second driving module is used to drive the virtual teacher to broadcast the answer statements to the question statements; The explanation acquisition module is used to, in response to the detection that the interactive question and answer between the virtual teacher has ended, acquire new explanation content generated by the large language model based on the interactive question and answer; wherein, the new explanation content re-explains the target topic from the point where the explanation was interrupted before the interactive question and answer; The third driving module is used to drive the virtual teacher to broadcast the new explanation content.

13. An electronic device, characterized in that, It includes at least a memory and a processor coupled to each other, the memory storing program instructions, and the processor executing the program instructions to implement the interactive presentation method according to any one of claims 1 to 11.

14. A learning machine, characterized in that, It includes a display device, a speaker device, a pickup device, and a control device, wherein the display device, the speaker device, and the pickup device are electrically connected to the control device, and the control device is the electronic device as described in claim 13.

15. A computer-readable storage medium, characterized in that, The system stores program instructions that can be executed by a processor, the program instructions being used to implement the interactive question-and-answer method according to any one of claims 1 to 11.

Images