CN-122000028-A - Intelligent pediatric nursing teaching AI inquiry system and method based on digital background

CN122000028ACN 122000028 ACN122000028 ACN 122000028ACN-122000028-A

Abstract

The invention discloses an intelligent pediatric nursing teaching AI inquiry system and method based on a digital background, and relates to the field of intersection of education informatization and artificial intelligence. The system comprises a front-end interaction module, a voice processing module, a teaching knowledge base, an intelligent question-answering engine, a teaching evaluation and feedback module, a background management and data analysis module and a virtual simulation operation linkage module, wherein the teaching knowledge base fuses a sentry card and pediatric characteristic knowledge of traditional Chinese medicine, supports knowledge patterns and vectorization retrieval, and the intelligent question-answering engine adopts a retrieval enhancement generation technology and combines a standardized question-answering dialogue state machine to ensure answer accuracy. The method comprises the steps of case selection, multi-mode interaction, intelligent question answering, process evaluation, virtual simulation linkage, self-adaptive recommendation and the like. The invention realizes the simulated consultation training, the fine procedural evaluation and the teaching ecological closed-loop optimization, improves the specialty, the safety and the individuation of pediatric nursing teaching, and cultures the core capacity of the power-assisted nursing talents.

Inventors

Shang Qingjuan
ZHU CHUNFENG
CUI YAMIN
LIU HONGXIA
ZOU XIAOHUI

Assignees

山东中医药高等专科学校

Dates

Publication Date: 20260508
Application Date: 20260127

Claims (10)

1. An intelligent pediatric nursing teaching AI consultation system based on a digital background, comprising: the front-end interaction module is used for rendering and displaying the digital human image of the doctor at the user terminal, receiving the voice or text input of the user, and synchronously playing the voice answer generated by the system and the mouth animation for driving the digital human image; The voice processing module is in communication connection with the front-end interaction module and is used for converting voice input of a user into text and converting text questions and answers generated by the system into voice streams; The system comprises a teaching knowledge base module, a semantic search module and a semantic search module, wherein the teaching knowledge base module stores structured pediatric nursing teaching knowledge data, wherein the knowledge data at least comprises pediatric nursing course standard content reconstructed based on post capability requirements, typical cases related to nursing skill big race standards, infant development guide professional skill grade evaluation standards, and pediatric nursing special knowledge and operation specifications of traditional Chinese medicine; the intelligent question-answering engine module is respectively in communication connection with the voice processing module and the teaching knowledge base module, and is used for receiving the converted user query text, carrying out retrieval enhancement generation based on the teaching knowledge base module, and generating text answers conforming to pediatric nursing teaching standards and medical accuracy in combination with a preset dialogue management strategy, wherein the dialogue management strategy comprises a dialogue state machine based on standardized patient question-answering flows and is used for guiding a user to complete nursing flows from nursing evaluation, nursing diagnosis to nursing measures and health guidance; The teaching evaluation and feedback module is in communication connection with the front-end interaction module and the intelligent question-answering engine module and is used for recording and analyzing interaction process data of a user and a system in real time, wherein the interaction process data at least comprises question logic integrity, keyword extraction accuracy, communication term normalization and matching degree of standard answers according to a teaching knowledge base; The background management and data analysis module is used for carrying out teaching case management, knowledge base updating, student learning track checking and teaching effect macroscopic analysis based on whole student interaction data at the teacher end so as to support continuous optimization of a teaching mode.
2. The digital context based intelligent pediatric care teaching AI consultation system of claim 1, wherein the teaching knowledge base module further comprises a dynamically updated typical case library containing a plurality of standardized pediatric care consultation scenarios, each scenario corresponding to a structured FAQ template comprising standard question sets, allowed similar question variants, and standard answer or answer generation logic conforming to a care specification, wherein the intelligent question and answer engine module preferentially matches and invokes the FAQ templates during a conversation based on current conversation state to provide standardized, evaluable interactive training.
3. The intelligent pediatric care teaching AI consultation system based on digital context according to claim 2, wherein the assessment model of the teaching assessment and feedback module incorporates a quantitative scoring function for calculating the score S of a single consultation exercise: The method comprises the steps of C logic representing a query logic integrity score, C keyword representing a keyword extraction accuracy score, C communication representing a communication term normalization score, C accuracy representing an answer matching score, and when a user answers a system query by a role playing nurse, similarity calculation is carried out on answers and standard answers of a knowledge base, w 1 ,w 2 ,w 3 ,w 4 is a weight coefficient of each score, w 1 +w 2 +w 3 +w 4 =1, and the weight coefficient can be adjusted by a teacher end according to teaching importance.
4. The intelligent pediatric nursing teaching AI consultation system based on the digital background according to claim 1, wherein the intelligent question-answer and engine module adopts a search enhancement generation method specifically comprising: (a) Vectorizing the user query text to obtain a query vector V q ; (b) In a vector database of the teaching knowledge base module, calculating the similarity between V q and all knowledge segment vectors, and searching Top-K most relevant knowledge segments D 1 ,D 2 ,...,D K ; (c) The user query text, the retrieved related knowledge segment D 1 ,D 2 ,...,D K and the current dialogue history context together form a prompt word, and the prompt word is input into a large language model; (d) The large language model generates final answer text based on the prompt words, wherein the answer text needs to reference or follow the content of the searched relevant knowledge segments and accords with the context of nursing communication.
5. The digital context-based intelligent pediatric nursing teaching AI consultation system according to claim 1, further comprising a virtual simulation operation linkage module, wherein when specific nursing operations are involved in answers or suggestions of the teaching assessment and feedback module generated by the intelligent question and answer engine module, the virtual simulation operation linkage module is triggered, a corresponding three-dimensional virtual simulation operation interface is presented at a user terminal, a user is guided to conduct simulation operation exercises, and operation result data are fed back to the teaching assessment and feedback module.
6. The digital context based intelligent pediatric care teaching AI consultation system of claim 1, wherein the front-end interaction module supports multi-modal input, allows a user to upload local photos or schematic drawings of an infant in addition to speech and text, and the intelligent question-answering engine module integrates a multi-modal large model capable of comprehensive analysis and answering in combination with image information and text dialogue context.
7. The intelligent pediatric care teaching AI consultation system based on digital context of claim 1, wherein the background management and data analysis module provides a teaching weaknesses analysis function based on group learning data by: (a) Collecting interactive process data and evaluation scores of all students on specific teaching cases or knowledge points; (b) For student groups with scores lower than a threshold value, carrying out cluster analysis on the frequently-found error types or missing inquiry links; (c) The overall error rate E k for a particular knowledge point is calculated, Where N error,k is the number of students with errors at knowledge point k, N total,k is the total number of students with contact to knowledge point k; (d) And (3) presenting the analysis result to a teacher in a visual chart form, automatically marking the knowledge points with high error rate, and prompting that the teaching needs to be enhanced or the knowledge base content needs to be optimized.
8. An intelligent pediatric nursing teaching AI consultation method based on a digital background, applied to the system according to any of claims 1-7, characterized by comprising the following steps: S1, starting a system, and loading and displaying a designated doctor digital person image and an initial greeting by a front-end interaction module; S2, receiving a consultation exercise request initiated by a user, and determining a pediatric nursing case theme of the exercise; S3, initializing a dialogue state machine based on the selected case, wherein the intelligent question-answering engine module generates a guiding problem or plays a patient/family to make a statement according to the current state and in combination with the teaching knowledge base module; s4, receiving voice or text input of a user, processing the voice or text input by an intelligent question-answering engine module after the voice input is converted by a voice processing module, combining knowledge retrieval and dialogue history to generate an answer meeting the nursing standard, and feeding back the answer in a digital human voice and animation mode through a front-end interaction module; S5, a teaching evaluation and feedback module monitors and records the interaction process of the steps S3 and S4 in real time, and evaluation characteristics are extracted; S6, repeating the steps S3 to S5 until the dialogue state machine judges that the inquiry flow is finished or the user initiatively ends; s7, generating and pushing a multidimensional evaluation report and a learning suggestion by the teaching evaluation and feedback module based on the complete interaction record; and S8, the background management and data analysis module gathers the current and historical interaction data and updates the group learning analysis report.
9. The method according to claim 8, wherein in step S3, the system supports a sentry demonstration fusion training mode, and the teacher end can configure the emphasis point of each exercise in the background, namely, when the sentry is emphasized, the dialogue flow and the evaluation standard are closely abutted to the clinical real post requirements, when the sentry is emphasized, the system simulates the standardized patient site assessment scene and the scoring rule of the nursing skill big race, and when the sentry is emphasized, the question and answer content and the evaluation point are unfolded around the assessment point of the infant development guide professional skill level certificate.
10. The method of claim 8, further comprising an adaptive learning path recommendation step: constructing a user capacity image according to the evaluation result of the teaching evaluation and feedback module on the multiple exercises of the user history, and identifying the strong and weak items of the user capacity image; The calculation consideration factors of the recommendation priority P rec comprise the historical score S hist of the corresponding knowledge point of the user, the group error rate E k of the knowledge point and the importance weight I k of the knowledge point in the whole course system, wherein the calculation formulas are as follows: ; And prompting the recommendation result to the user through the front-end interaction module.

Description

Intelligent pediatric nursing teaching AI inquiry system and method based on digital background Technical Field The invention relates to the field of cross application of education informatization and artificial intelligence technology, in particular to an intelligent teaching system and method, and especially relates to an intelligent pediatric nursing teaching AI inquiry system and method which are special for pediatric nursing courses of nursing professions in higher vocational education under the digital education background and integrate digital personal technology, intelligent voice interaction, knowledge enhancement generation, teaching evaluation and data analysis. Background With the deep advancement of digital strategies and the rapid development of artificial intelligence technology, hybrid teaching modes and intelligent teaching tools are widely applied to the field of professional education, particularly medical nursing education. At present, auxiliary teaching by using an online platform, virtual simulation and an intelligent question-answering system has become an important means for improving teaching effects and relieving practical training resource pressure. Currently, online medical consultation or teaching auxiliary systems in the market mostly adopt text chat robots or prerecorded videos to carry out knowledge transfer. Although a 'digital person' integrating voice interaction and a simple animation image appears, the core function of the 'digital person' still stays on information broadcasting or question answering of a fixed flow. These systems lack structural knowledge system support combined with the depth of professional care teaching, and their dialogue logic is common, so that they cannot simulate the complex and dynamic care evaluation communication scene (such as pediatric consultation) in real clinic. When facing open unstructured student questions, a system based on a general large language model easily generates content illusions, gives answers which do not accord with medical specifications or teaching outlines, has misleading risks, and cannot meet the strict requirements of high-standard nursing teaching on accuracy and safety. Many so-called teaching software is only a digital display of knowledge points and cannot effectively realize the 'sentry course' communication. They cannot organically integrate the core capability requirements of clinical posts, the scoring criteria of nursing skill games, and the assessment points of professional skill level certificates into an interactive training and assessment system. When students practice, the students lack definite professional ability to guide standard and standardized operation flow, faults exist in the training process, clinical actual demands and authoritative evaluation systems, so that training pertinence is not strong, and teaching effects are difficult to quantitatively measure. The traditional online learning system evaluation is mostly dependent on post-class choice question test or subjective evaluation of teachers, and cannot perform procedural and formative evaluation on core literacy such as clinical thinking process, communication skills, in-situ decision-making capability and the like of students. The prior art is difficult to capture and analyze whether the logic of students in the simulated inquiry is complete, whether words are professional or not, and whether communication is concentric or not in real time, so that instant, specific and data-driven personalized feedback and improvement suggestions cannot be provided. Teaching stays in the shallow layer of "exercise-answer", lacking a deep "assessment-feedback-optimization" closed loop. Most teaching platforms only record surface layer data such as simple login times, video watching time length and the like, and lack effective mining and analysis capability on deep interactive data (such as dialogue texts, operation sequences and decision paths) generated in the teaching process. The teacher can not quickly and accurately identify weak links of teaching contents and common cognitive error areas of student groups from group data, so that teaching optimization depends on experience rather than evidence, and accurate teaching of data driving and continuous iterative upgrading of a hybrid teaching mode are prevented. The current training system often breaks the inquiry communication training and the skill training. The inquiry system only manages 'inquiry' and the virtual simulation software only manages 'do', and the inquiry system and the virtual simulation software lack of organic linkage. Students cannot experience natural consistency from communication assessment to execution operation in completing a complete care process. In addition, the interactive mode of the system is single, and the multimode input of voice, text, images (such as skin rash photos of children) and the like cannot be effectively integrated, so that the reality and the complexity o