Generative AI-based nursing case intelligent teaching question-answering robot and application thereof

Abstract

The invention relates to a generative AI-based nursing case intelligent teaching question and answer robot and application thereof.A current teaching task is executed through an RAG enhanced retrieval module, question and answer knowledge is retrieved from a nursing knowledge base based on keywords in user question and answer information, and LLM is used for performing enhanced retrieval on the question and answer knowledge; the learning dynamic evaluation module performs learning performance evaluation on answer information in user question and answer information by using LLM, and the workflow management module pushes teaching tasks with corresponding learning strength to the RAG enhanced retrieval module in a matching manner to perform personalized learning path and branch scheduling. Clinical nursing cases can be created based on the generative AI, the nursing key points are answered and answered in a guiding mode, the clinical nursing thinking of students is improved in repeated practice, and the nursing teaching effect is improved. And nursing cases of different specialities are generated according to a rule enhancement formula, so that students are gradually guided to strengthen learning, and the autonomous learning ability and clinical thinking of the students are improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of intelligent nursing robots, and in particular to a nursing case intelligent teaching question-answering robot based on generative AI, an application method thereof, and an electronic device. Background Art

[0002] The Guidelines for the Use of Generative Artificial Intelligence in Education and Research encourage the use of generative artificial intelligence (AI) in education and urge its strengthening supervision. This shows that the application of generative AI in education is receiving more and more attention and attention. However, although some studies have explored the changes, advantages and disadvantages, and countermeasures brought by generative AI to education, there are relatively few studies focusing on nursing teaching.

[0003] Although teaching robots are gradually being used in nursing work, the working principle of the existing nursing teaching robots is mainly: the nursing teaching robots realize interactive teaching with students through the integrated question-answering system. Its working principle mainly includes the following aspects:

[0004] 1. Natural language processing: The robot uses natural language processing technologies, such as speech recognition and semantic understanding, to parse and understand students’ questions;

[0005] 2. Knowledge retrieval and matching: Based on understanding the student's question, the robot retrieves relevant information from the built-in knowledge base and finds the answer that matches the question;

[0006] 3. Answer generation and feedback: The robot generates accurate and complete answers based on the retrieved knowledge and provides feedback to students through voice or text;

[0007] 4. Continuous learning and optimization: The robot uses machine learning technology to continuously learn and optimize from interactions with students, improving the accuracy and efficiency of question and answering.

[0008] Traditional nursing teaching robots mainly generate corresponding question-and-answer pairs based on user questions. This method does not have the ability to dynamically evaluate the results of user question-and-answer training, strengthen training based on the evaluation results, and provide guiding answers to increase the difficulty of questions and answers. It cannot allow students to dynamically manage the learning process in a mode of gradually increasing difficulty, and enhance students' depth of understanding of nursing teaching. Summary of the invention

[0009] In order to solve the above problems, the present application proposes a nursing case intelligent teaching question-answering robot based on generative AI and its application method and electronic equipment.

[0010] On the one hand, this application proposes a nursing case intelligent teaching question-answering robot based on generative AI, which is connected to the user-side APP in communication, including:

[0011] Input module, used to collect and input user question and answer information;

[0012] RAG enhanced retrieval module, used to perform the current teaching task, retrieve question and answer knowledge from the nursing knowledge base based on the keywords in the user's question and answer information, and use the LLM large language model to perform enhanced retrieval on the question and answer knowledge, generate corresponding question data and feed it back to the output module;

[0013] The learning dynamic evaluation module is used to use the LLM large language model to evaluate the learning performance of the answer information in the user's question and answer information and output the corresponding performance evaluation results to the workflow management module;

[0014] The workflow management module is used to match and push teaching tasks of corresponding learning intensity to the RAG enhanced retrieval module according to the performance evaluation results, and perform personalized learning path and branch scheduling;

[0015] Nursing knowledge base, used to store knowledge about question and answer pairs related to nursing teaching;

[0016] The output module is used to output the problem data to the user-side APP.

[0017] As an optional implementation scheme of the present application, optionally, the execution of the current teaching task, based on the keywords in the user's question and answer information, retrieves question and answer knowledge from the nursing knowledge base and uses the LLM large language model to perform enhanced retrieval on the question and answer knowledge, generates corresponding question data and feeds it back to the output module, including:

[0018] Parsing the user question and answer information, and identifying and extracting question keywords from the parsed results;

[0019] Build and configure question retrieval logic;

[0020] Combining the question keywords and the question search logic, constructing search prompt words and inputting them into the LLM large language model;

[0021] Retrieving the question-answer knowledge matching the search prompt word from the nursing knowledge base through the LLM large language model to achieve enhanced retrieval;

[0022] The retrieved question data is fed back to the output module.

[0023] As an optional implementation scheme of the present application, optionally, the use of the LLM large language model to evaluate the learning performance of the answer information in the user question and answer information and outputting the corresponding performance evaluation result to the workflow management module includes:

[0024] Receive the user question and answer information and extract the user's answer information to the current teaching task;

[0025] Parsing the answer information, and identifying and extracting answer keywords from the parsed results;

[0026] Build and configure answer evaluation logic;

[0027] Combining the answer keywords and the answer evaluation logic, constructing answer evaluation prompt words and inputting them into the LLM large language model;

[0028] Using the LLM large language model, based on the answer evaluation prompt words, the user's answer information to the current teaching task is evaluated for learning performance and a corresponding performance evaluation result is output;

[0029] The performance evaluation result is fed back to the workflow management module.

[0030] As an optional implementation scheme of the present application, optionally, according to the performance evaluation result, the teaching tasks of corresponding learning intensity are matched and pushed to the RAG enhanced retrieval module to perform personalized learning path and branch scheduling, including:

[0031] Pre-constructing and storing learning flows of different learning intensities, wherein: the learning flow includes the personalized learning path and branch scheduling;

[0032] Receive the performance evaluation result input by the learning dynamic evaluation module, and determine whether the performance evaluation result reaches the learning intensity required by the current teaching task:

[0033] If it is reached, the teaching task of the next learning intensity is pushed to the RAG enhanced retrieval module to execute the learning flow of the corresponding learning intensity;

[0034] Otherwise, continue to execute the learning flow of the current teaching task.

[0035] As an optional implementation scheme of the present application, optionally, the learning dynamic evaluation module is also used to:

[0036] Parse the answer information in the user's question and answer information to obtain the answer information therein;

[0037] Using the LLM large language model, the answer information is compared and evaluated with the standard answer in the nursing knowledge base, and the evaluation result of the user's answer to the current teaching task is output;

[0038] The LLM large language model gives corresponding nursing learning suggestions based on the answer evaluation results and outputs them to the output module.

[0039] As an optional implementation scheme of the present application, optionally, the learning dynamic evaluation module is also used to:

[0040] Collecting evaluation results of users' responses to all teaching tasks;

[0041] All the evaluation results of the answers are counted and analyzed through the LLM large language model, and a corresponding evaluation report is generated;

[0042] The evaluation report is output to the output module.

[0043] On the other hand, the present application proposes an application method of a nursing case intelligent teaching question-answering robot based on generative AI, comprising the following steps:

[0044] Log in to the generative AI-based nursing case intelligent teaching question-answering robot through the user-side APP, and the generative AI-based nursing case intelligent teaching question-answering robot starts to perform the current teaching task:

[0045] The RAG enhanced retrieval module performs the current teaching task;

[0046] The input module collects and inputs user question and answer information;

[0047] The RAG enhanced retrieval module retrieves question-and-answer knowledge from the nursing knowledge base based on the keywords in the user's question-and-answer information and uses the LLM large language model to perform enhanced retrieval on the question-and-answer knowledge, generates corresponding question data and feeds it back to the output module;

[0048] The learning dynamic evaluation module is used to use the LLM large language model to evaluate the learning performance of the answer information in the user's question and answer information and output the corresponding performance evaluation results to the workflow management module;

[0049] The workflow management module is used to match and push teaching tasks of corresponding learning intensity to the RAG enhanced retrieval module according to the performance evaluation results, and perform personalized learning path and branch scheduling;

[0050] The output module is used to output the problem data to the user-side APP.

[0051] In another aspect, the present application further provides an electronic device, comprising:

[0052] processor;

[0053] a memory for storing processor-executable instructions;

[0054] Among them, the processor is configured to implement the application method of the nursing case intelligent teaching question-and-answer robot based on generative AI when executing the executable instructions.

[0055] Technical effects of the present invention:

[0056] This application performs the current teaching task through the RAG enhanced retrieval module, retrieves the question and answer knowledge from the nursing knowledge base based on the keywords in the user's question and answer information, and uses the LLM large language model to enhance the retrieval of the question and answer knowledge, generates corresponding question data and feeds it back to the output module; uses the LLM large language model to evaluate the learning performance of the answer information in the user's question and answer information through the learning dynamic evaluation module, and outputs the corresponding performance evaluation results to the workflow management module; through the workflow management module, according to the performance evaluation results, the teaching tasks of the corresponding learning intensity are matched and pushed to the RAG enhanced retrieval module for personalized learning paths and branch scheduling. It can create clinical nursing cases and guided question and answer nursing points based on generative AI, so that students can improve their clinical nursing thinking and improve the teaching effect of nursing through repeated practice. Generate nursing cases of different specialties according to rule enhancement, so as to gradually guide students to strengthen their learning and improve their autonomous learning ability and clinical thinking.

[0057] Further features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure and, together with the description, serve to explain the principles of the disclosure.

[0059] Figure 1 It is a schematic diagram showing the structure of the robot application system of the present invention;

[0060] Figure 2 An example of an AI dialogue for question-and-answer teaching of pain care according to the present invention is shown;

[0061] Figure 3 Shown is a schematic flow chart of the application method of the present invention;

[0062] Figure 4 It is a schematic diagram showing the application of the electronic device of the present invention. DETAILED DESCRIPTION

[0063] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0064] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0065] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present disclosure.

[0066] LLM large language model, referred to as "LLM" or "model", can be a technology developed by users themselves or requesting a third-party API, calling a third-party LLM large language model (such as Baidu's Wenxinyiyan), which is not limited in this embodiment.

[0067] The present invention can be applied to clinical nursing teaching of thyroid cancer, for example, or other subjects.

[0068] Example 1

[0069] like Figure 1 As shown, on the one hand, this application proposes a nursing case intelligent teaching question-answering robot based on generative AI, which is connected to the user-side APP in communication, including:

[0070] Input module, used to collect and input user question and answer information;

[0071] RAG enhanced retrieval module, used to perform the current teaching task, retrieve question and answer knowledge from the nursing knowledge base based on the keywords in the user's question and answer information, and use the LLM large language model to perform enhanced retrieval on the question and answer knowledge, generate corresponding question data and feed it back to the output module;

[0072] The learning dynamic evaluation module is used to use the LLM large language model to evaluate the learning performance of the answer information in the user's question and answer information and output the corresponding performance evaluation results to the workflow management module;

[0073] The workflow management module is used to match and push teaching tasks of corresponding learning intensity to the RAG enhanced retrieval module according to the performance evaluation results, and perform personalized learning path and branch scheduling;

[0074] Nursing knowledge base, used to store knowledge about question and answer pairs related to nursing teaching;

[0075] The output module is used to output the problem data to the user-side APP.

[0076] The principle of the present invention is as follows:

[0077] The current teaching task is performed through the RAG enhanced retrieval module. Based on the keywords in the user's question and answer information, the question and answer knowledge is retrieved from the nursing knowledge base and the LLM large language model is used to enhance the retrieval of the question and answer knowledge, generate the corresponding question data and feed it back to the output module; the learning dynamic evaluation module uses the LLM large language model to evaluate the answer information in the user's question and answer information, and outputs the corresponding performance evaluation results to the workflow management module; the workflow management module pushes the teaching tasks of the corresponding learning intensity to the RAG enhanced retrieval module according to the performance evaluation results, and performs personalized learning paths and branch scheduling. It can create clinical nursing cases and guided question and answer nursing points based on generative AI, so that students can improve their clinical nursing thinking and improve the teaching effect of nursing through repeated practice. Nursing cases of different specialties are generated according to rule enhancement, so as to gradually guide students to strengthen their learning and improve their autonomous learning ability and clinical thinking.

[0078] Combined with Figure 1 As shown, the core technologies of the intelligent teaching question-answering robot proposed in the present invention include LLM (large language model), RAG (retrieval-based generation technology, enhanced retrieval module) and workflow management module (input / output module, which is a conventional interface or module of the robot, and can refer to the input / output of existing robots. The APP and the AI ​​robot generally use API communication, which can be combined with the data communication and interaction mechanism between the existing AI teaching APP and its background). In the specific implementation process, the system is divided into two teaching modes:

[0079] Mode 1: Completely rely on LLM to generate case background, question content and standard answers

[0080] Mode 1: LLM-based case generation: LLM is used to dynamically generate cases that simulate clinical scenarios. The system automatically generates relevant case contexts based on keywords entered by the user (such as "surgery" or "thyroid cancer"), thereby providing users with a comprehensive situational simulation. The generated cases include background descriptions, nursing questions, and standard answers to ensure that students can obtain learning content from multiple perspectives.

[0081] Mode 2: Combine RAG technology to obtain case background and standard nursing measures from the knowledge base, and generate targeted questions by LLM

[0082] This modular design effectively meets diverse teaching needs. It can not only build content using generative models, but also provide more targeted learning paths based on the existing knowledge base.

[0083] Mode 2 is a combination of RAG technology and knowledge base, which uses RAG technology to retrieve specific case backgrounds and standard care measures. This information is first provided by the knowledge base, and then LLM generates targeted questions based on this information.

[0084] like Figure 2 As shown in the figure, RAG enhances LLM's understanding of domain knowledge, enabling it to combine existing knowledge resources, more accurately model problems and ask appropriate questions. The answers it generates are more desirable and can provide users with better teaching question-answering knowledge (deep breathing and tightening the incision, which enhances the answer content).

[0085] Mode 1 is more traditional, and users can learn and understand it by combining it with the existing AI robot's intelligent question-and-answer mode, so it will not be elaborated here.

[0086] The present invention will be mainly described in conjunction with Mode 2 (the two modes can be determined by the administrator according to the learning level of the student user, such as the grade of study, age, etc.).

[0087] The workflow management module provides personalized learning paths and branch scheduling with different learning intensities (learning flow, which can be built in a low-code way. It contains multiple nodes and processes. It is mainly used to build nursing teaching tasks, and conduct nursing teaching, assessment and analysis at each node according to the process. The specific teaching content and nodes can be set by the administrator): dynamic evaluation of students' learning performance based on LLM. The system will analyze the students' answers after each round of questions, and adjust the difficulty and content focus of subsequent questions accordingly. This mechanism is implemented through the workflow management module. The workflow module manages the students' learning process through branch scheduling, and pre-defines different learning flows. As the questions and answers progress, the system will dynamically enter the corresponding flow to ensure that the students' learning path is highly personalized and coherent. This branch scheduling mechanism effectively improves the adaptability and responsiveness of the system.

[0088] The implementation details of the present invention will be further described below.

[0089] As an optional implementation scheme of the present application, optionally, the execution of the current teaching task, based on the keywords in the user's question and answer information, retrieves question and answer knowledge from the nursing knowledge base and uses the LLM large language model to perform enhanced retrieval on the question and answer knowledge, generates corresponding question data and feeds it back to the output module, including:

[0090] Parsing the user question and answer information, and identifying and extracting question keywords from the parsed results;

[0091] Build and configure question retrieval logic;

[0092] Combining the question keywords and the question search logic, constructing search prompt words and inputting them into the LLM large language model;

[0093] Retrieving the question-answer knowledge matching the search prompt word from the nursing knowledge base through the LLM large language model to achieve enhanced retrieval;

[0094] The retrieved question data is fed back to the output module.

[0095] User question and answer information includes students’ answers to questions asked by the robot, as well as other students’ question and answer information, etc. The specific information is subject to the interaction.

[0096] In the enhanced search of the RAG enhanced search module, we first analyze the keywords in the user's question, and then combine them with the search logic constructed by the administrator to build prompt words to remind the large language model to answer. The model locates the search from the library based on the logic and keywords in the prompt words, and the enhanced search obtains the corresponding question data.

[0097] At the beginning of the teaching task, the robot can collect the subjects answered by students and extract question keywords (such as thyroid cancer + nursing) based on the subjects (such as thyroid cancer clinical nursing learning).

[0098] Specific:

[0099] 1. Construct search prompt words

[0100] Keyword extraction: First, automatically extract keywords from the questions entered by the user. These keywords should be directly related to the core points of nursing knowledge, such as disease names, nursing operations, drug names, etc.

[0101] Question type identification: Use natural language processing technology to identify the type of question (such as definition inquiry, operation steps, cause analysis, etc.) in order to construct search prompts more accurately.

[0102] Retrieval logic construction: Based on the extracted keywords and question types, the retrieval logic is constructed. This may include the combination of keywords, synonym replacement, and the addition of contextual related words (for example, searching for knowledge related to post-care of thyroid cancer), etc., to form a series of retrieval prompts that can fully cover the user's intentions.

[0103] Dynamic adjustment: According to the feedback from the LLM large language model and the actual needs of users, the search prompt word construction strategy is dynamically adjusted to improve the accuracy and efficiency of the search.

[0104] 2. Enter the LLM large language model for retrieval

[0105] Model interface call: Input the constructed search prompt words into the LLM large language model through the API interface or other methods.

[0106] Model processing: The LLM large language model uses its powerful natural language understanding and generation capabilities to deeply analyze the input search prompt words, and retrieves relevant information from the nursing knowledge base based on its internal knowledge representation and reasoning mechanism.

[0107] Sorting of search results: The model sorts the search results based on factors such as the relevance of the search results to the search prompt terms, the authority and timeliness of the information, so that the information that best meets the user's needs is presented first.

[0108] 3. Implement enhanced retrieval

[0109] Result screening and filtering: From the search results returned by the LLM large language model, further screen and filter out information that is irrelevant to user needs or of low quality to ensure that accurate and useful nursing knowledge is provided to users.

[0110] Result integration and presentation: Integrate the filtered search results and present them to users in a structured manner (such as lists, tables, charts, etc.) to facilitate users' quick understanding and application.

[0111] It can also intelligently recommend relevant nursing knowledge, learning resources or expert advice based on the user's search history and preferences to further enrich the user's learning experience.

[0112] The specific interactive output can be output in the form of text or graphics, as shown in the attached Figure 2 As shown, the present invention is output in text.

[0113] Through the above detailed implementation content, we can make full use of the powerful capabilities of the LLM large language model to achieve efficient and accurate retrieval of the nursing knowledge base, providing strong support for nursing education and clinical practice. The retrieval results are output to the APP by the output module for viewing.

[0114] As an optional implementation scheme of the present application, optionally, the use of the LLM large language model to evaluate the learning performance of the answer information in the user question and answer information and outputting the corresponding performance evaluation result to the workflow management module includes:

[0115] Receive the user question and answer information and extract the user's answer information to the current teaching task;

[0116] Parsing the answer information, and identifying and extracting answer keywords from the parsed results;

[0117] Build and configure answer evaluation logic;

[0118] Combining the answer keywords and the answer evaluation logic, constructing answer evaluation prompt words and inputting them into the LLM large language model;

[0119] Using the LLM large language model, based on the answer evaluation prompt words, the user's answer information to the current teaching task is evaluated for learning performance and a corresponding performance evaluation result is output;

[0120] The performance evaluation result is fed back to the workflow management module.

[0121] The system will analyze the students' answers after each round of questions and adjust the difficulty and content focus of subsequent questions accordingly. This mechanism is implemented through the workflow management module, which manages the students' learning process through branch scheduling and pre-defines different learning flows. As the questions and answers progress, the system will dynamically enter the corresponding flow to ensure that the students' learning path is highly personalized and coherent. This branch scheduling mechanism effectively improves the adaptability and responsiveness of the system.

[0122] Here, LLM is used based on the constructed answer evaluation prompts to allow the model to evaluate the students' answer information, determine the gap between the students' answer information and the standard answers in the system library, and output the corresponding learning performance evaluation results (the output structure of the evaluation results is determined based on the answer evaluation logic. For example, the administrator can specify the answer comparison object and evaluation criteria in the answer evaluation logic, and the administrator can determine the specific logical formula).

[0123] LLM intelligently evaluates the answer information input by the user based on the evaluation prompt words, extracts and compares the answer keywords according to the logical formula in the evaluation logic, and outputs the evaluation results corresponding to each knowledge. Subsequently, the student's performance evaluation results for the current teaching task can be output (in a table format, with the evaluation results of each knowledge item filled in the table).

[0124] As an optional implementation scheme of the present application, optionally, according to the performance evaluation result, the teaching tasks of corresponding learning intensity are matched and pushed to the RAG enhanced retrieval module to perform personalized learning path and branch scheduling, including:

[0125] Pre-constructing and storing learning flows of different learning intensities, wherein: the learning flow includes the personalized learning path and branch scheduling;

[0126] Receive the performance evaluation result input by the learning dynamic evaluation module, and determine whether the performance evaluation result reaches the learning intensity required by the current teaching task:

[0127] If it is reached, the teaching task of the next learning intensity is pushed to the RAG enhanced retrieval module to execute the learning flow of the corresponding learning intensity;

[0128] Otherwise, continue to execute the learning flow of the current teaching task.

[0129] When constructing different learning flows, learning flows corresponding to different learning intensities can be constructed in batches according to the teaching difficulty, wherein: the learning flow includes the personalized learning path and branch scheduling. Each learning flow contains several teaching nodes, and each node is deployed with a corresponding learning path (content) and branch scheduling (that is, learning on the corresponding learning path is carried out according to different learning difficulties, and when the corresponding node is reached, learning is carried out according to the learning branch, similar to a tree structure, each node is configured with a learning intensity, and different difficulties are reached (for example, if the intensity score is 5, the teaching task of the next level of learning intensity is pushed to the user (the learning flow with a score of 6 is automatically selected (RAG enhanced retrieval module performs the selection) for branch scheduling learning))).

[0130] As an optional implementation scheme of the present application, optionally, the learning dynamic evaluation module is also used to:

[0131] Parse the answer information in the user's question and answer information to obtain the answer information therein;

[0132] Using the LLM large language model, the answer information is compared and evaluated with the standard answer in the nursing knowledge base, and the evaluation result of the user's answer to the current teaching task is output;

[0133] The LLM large language model gives corresponding nursing learning suggestions based on the answer evaluation results and outputs them to the output module.

[0134] The present invention also has real-time feedback and context retention functions: the intelligent teaching question-answering robot uses LLM to conduct in-depth analysis of students' answers and provides detailed improvement suggestions by comparing them with standard answers. In order to improve the effectiveness of feedback, the system retains the context of each round of questions and answers to ensure the coherence and pertinence of feedback.

[0135] The nursing knowledge base is provided with standard answers to corresponding questions. After the robot collects the answer information of the student's answer, it can be input into the learning dynamic evaluation module, so that it can use the LLM large language model to compare and evaluate the answer information with the standard answer in the nursing knowledge base, and output the evaluation result of the user's answer to the current teaching task. In the specific operation, the LLM large language model can be used to compare and analyze the keywords in the answer information and the standard answer in the nursing knowledge base, evaluate the matching degree of the keywords and the number of keywords that reach the preset matching degree, so as to evaluate the similarity between the answer information and the standard answer in the nursing knowledge base. For example, 3 keywords are extracted from the answer information and the standard answer in the nursing knowledge base. Through keyword similarity analysis, if at least 2 keywords are found to be consistent or similar, it is determined that the student's answer is close, and the system can determine that its answer meets the teaching requirements (qualified). On the contrary, it is determined to be unqualified, and the LLM can automatically recommend the answer information of the corresponding question and answer (that is, nursing learning suggestions) to the student. In this process, LLM extracts the standard answer to the question from the library and outputs it to the user to strengthen its learning.

[0136] The above-mentioned LLM can evaluate the answer information and the standard answers in the nursing knowledge base based on the question and answer evaluation logic input by the administrator (such as keyword similarity calculation and qualification threshold).

[0137] As an optional implementation scheme of the present application, optionally, the learning dynamic evaluation module is also used to:

[0138] Collecting evaluation results of users' responses to all teaching tasks;

[0139] All the evaluation results of the answers are counted and analyzed through the LLM large language model, and a corresponding evaluation report is generated;

[0140] The evaluation report is output to the output module.

[0141] The present invention also has a comprehensive evaluation mechanism function service: after the entire question-answering round is over, the mechanism uses LLM to analyze all the students' answers as a whole and generate a comprehensive evaluation report. The evaluation report includes the students' mastery of each knowledge point and their overall learning performance, and provides targeted improvement suggestions to help students better understand what they have learned and improve their practical ability.

[0142] The LLM can analyze the evaluation results of the students' answers to each question as a whole to determine in which aspects of nursing knowledge the students have passed or failed in this learning process, and can make statistics (according to the above evaluation results). The statistical results can be entered into the report template to generate an evaluation report.

[0143] During the overall analysis, nursing learning suggestions can be annotated for unqualified knowledge to remind students to learn the nursing learning suggestions for the unqualified parts. Finally, LLM can also collect their overall learning performance to make overall improvement suggestions. For example, based on the calculation of the pass rate, LLM can automatically output nursing learning suggestions corresponding to the degree of learning performance, which weak points to start with, and so on.

[0144] Obviously, those skilled in the art should understand that the implementation of all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above-mentioned embodiments. Those skilled in the art can understand that the implementation of all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above-mentioned embodiments. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (Flash Memory), a hard disk (Hard Disk Drive, abbreviated as: HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above-mentioned types of memory.

[0145] Example 2

[0146] like Figure 3 As shown, based on the implementation principle of Example 1, on the other hand, the present application proposes an application method of a nursing case intelligent teaching question-answering robot based on generative AI, comprising the following steps:

[0147] S1. Log in to the intelligent teaching question-answering robot for nursing cases based on generative AI through the user-side APP;

[0148] S2. The generative AI-based nursing case intelligent teaching question-answering robot starts to perform the current teaching task:

[0149] The RAG enhanced retrieval module performs the current teaching task;

[0150] The input module collects and inputs user question and answer information;

[0151] The RAG enhanced retrieval module retrieves question-and-answer knowledge from the nursing knowledge base based on the keywords in the user's question-and-answer information and uses the LLM large language model to perform enhanced retrieval on the question-and-answer knowledge, generates corresponding question data and feeds it back to the output module;

[0152] The learning dynamic evaluation module uses the LLM large language model to evaluate the answer information in the user's question and answer information, and outputs the corresponding performance evaluation results to the workflow management module;

[0153] The workflow management module pushes the teaching tasks of corresponding learning intensity to the RAG enhanced retrieval module according to the performance evaluation results, and performs personalized learning path and branch scheduling;

[0154] The output module outputs the problem data to the user-side APP.

[0155] Please understand the above steps in conjunction with the method principles of Example 1, and will not be repeated in this example.

[0156] The modules or steps of the present invention described above can be implemented by a general-purpose computing system, they can be concentrated on a single computing system, or distributed on a network composed of multiple computing systems, and optionally, they can be implemented by a program code executable by a computing system, so that they can be stored in a storage system and executed by the computing system, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0157] Example 3

[0158] like Figure 4 As shown, further, in another aspect, the present application also proposes an electronic device, including:

[0159] processor;

[0160] a memory for storing processor-executable instructions;

[0161] Among them, the processor is configured to implement the application method of the nursing case intelligent teaching question-and-answer robot based on generative AI as described in Example 2 when executing the executable instructions.

[0162] The electronic device of the disclosed embodiment includes a processor and a memory for storing processor executable instructions. The processor is configured to implement the application method of the generative AI-based nursing case intelligent teaching question-answering robot described in the previous embodiment 2 when executing the executable instructions.

[0163] Here, it should be noted that the number of processors can be one or more. At the same time, the electronic device of the embodiment of the present disclosure may also include an input system and an output system. Among them, the processor, memory, input system and output system may be connected through a bus or in other ways, which are not specifically limited here.

[0164] The memory, as a computer-readable storage medium, can be used to store software programs, computer executable programs and various modules, such as the program or module corresponding to the application method of the intelligent teaching question-answering robot for nursing cases based on generative AI in the embodiment of the present disclosure. The processor executes various functional applications and data processing of the electronic device by running the software programs or modules stored in the memory.

[0165] The input system can be used to receive input numbers or signals. The signal can be a key signal related to user settings and function control of the device / terminal / server. The output system can include display devices such as display screens.

[0166] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A nursing case intelligent teaching question-answering robot based on generative AI, characterized in that: Communication connection with the user-side APP, including: Input module, used to collect and input user question and answer information; RAG enhanced retrieval module, used to perform the current teaching task, retrieve question and answer knowledge from the nursing knowledge base based on the keywords in the user's question and answer information, and use the LLM large language model to perform enhanced retrieval on the question and answer knowledge, generate corresponding question data and feed it back to the output module; The learning dynamic evaluation module is used to use the LLM large language model to evaluate the learning performance of the answer information in the user's question and answer information and output the corresponding performance evaluation results to the workflow management module; The workflow management module is used to match and push teaching tasks of corresponding learning intensity to the RAG enhanced retrieval module according to the performance evaluation results, and perform personalized learning path and branch scheduling; Nursing knowledge base, used to store knowledge about question and answer pairs related to nursing teaching; The output module is used to output the problem data to the user-side APP.

2. The generative AI-based nursing case intelligent teaching question-answering robot and its application according to claim 1 is characterized in that: The execution of the current teaching task, based on the keywords in the user's question and answer information, retrieves question and answer knowledge from the nursing knowledge base and uses the LLM large language model to perform enhanced retrieval on the question and answer knowledge, generates corresponding question data and feeds it back to the output module, including: Parsing the user question and answer information, and identifying and extracting question keywords from the parsed results; Build and configure question retrieval logic; Combining the question keywords and the question search logic, constructing search prompt words and inputting them into the LLM large language model; Retrieving the question-answer knowledge matching the search prompt word from the nursing knowledge base through the LLM large language model to achieve enhanced retrieval; The retrieved question data is fed back to the output module.

3. The generative AI-based nursing case intelligent teaching question-answering robot according to claim 1 is characterized in that: The method of using the LLM large language model to evaluate the learning performance of the answer information in the user's question and answer information and outputting the corresponding performance evaluation result to the workflow management module includes: Receive the user question and answer information and extract the user's answer information to the current teaching task; Parsing the answer information, and identifying and extracting answer keywords from the parsed results; Build and configure answer evaluation logic; Combining the answer keywords and the answer evaluation logic, constructing answer evaluation prompt words and inputting them into the LLM large language model; Using the LLM large language model, based on the answer evaluation prompt words, the user's answer information to the current teaching task is evaluated for learning performance and a corresponding performance evaluation result is output; The performance evaluation result is fed back to the workflow management module.

4. The generative AI-based nursing case intelligent teaching question-answering robot according to claim 1 is characterized in that: According to the performance evaluation results, the teaching tasks of corresponding learning intensity are matched and pushed to the RAG enhanced retrieval module to perform personalized learning paths and branch scheduling, including: Pre-constructing and storing learning flows of different learning intensities, wherein: the learning flow includes the personalized learning path and branch scheduling; Receive the performance evaluation result input by the learning dynamic evaluation module, and determine whether the performance evaluation result reaches the learning intensity required by the current teaching task: If it is reached, the teaching task of the next learning intensity is pushed to the RAG enhanced retrieval module to execute the learning flow of the corresponding learning intensity; Otherwise, continue to execute the learning flow of the current teaching task.

5. The generative AI-based nursing case intelligent teaching question-answering robot according to claim 1 is characterized in that: The learning dynamic evaluation module is also used to: Parse the answer information in the user's question and answer information to obtain the answer information therein; Using the LLM large language model, the answer information is compared and evaluated with the standard answer in the nursing knowledge base, and the evaluation result of the user's answer to the current teaching task is output; The LLM large language model gives corresponding nursing learning suggestions based on the answer evaluation results and outputs them to the output module.

6. The generative AI-based nursing case intelligent teaching question-answering robot according to claim 5 is characterized in that: The learning dynamic evaluation module is also used for: Collecting evaluation results of users' responses to all teaching tasks; All the evaluation results of the answers are counted and analyzed through the LLM large language model, and a corresponding evaluation report is generated; The evaluation report is output to the output module.

7. An application method of the generative AI-based nursing case intelligent teaching question-answering robot according to any one of claims 1 to 8, characterized in that: The steps include: Log in to the generative AI-based nursing case intelligent teaching question-answering robot through the user-side APP, and the generative AI-based nursing case intelligent teaching question-answering robot starts to perform the current teaching task: The RAG enhanced retrieval module performs the current teaching task; The input module collects and inputs user question and answer information; The RAG enhanced retrieval module retrieves question-and-answer knowledge from the nursing knowledge base based on the keywords in the user's question-and-answer information and uses the LLM large language model to perform enhanced retrieval on the question-and-answer knowledge, generates corresponding question data and feeds it back to the output module; The learning dynamic evaluation module is used to use the LLM large language model to evaluate the learning performance of the answer information in the user's question and answer information and output the corresponding performance evaluation results to the workflow management module; The workflow management module is used to match and push teaching tasks of corresponding learning intensity to the RAG enhanced retrieval module according to the performance evaluation results, and perform personalized learning path and branch scheduling; The output module is used to output the problem data to the user-side APP.

8. An electronic device, characterized in that include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the application method of the nursing case intelligent teaching question-and-answer robot based on generative AI as described in claim 7 when executing the executable instructions.

Images