Intelligent auxiliary system and method for gastrointestinal nursing teaching

Abstract

The invention provides an intelligent auxiliary system and method for gastrointestinal nursing teaching, and relates to the field of intelligent assistance, and the method comprises the steps: firstly, extracting a part where a user answers mistakenly from an evaluation test report, and carrying out the content segmentation and semantic coding of the part through employing a semantic analysis and processing algorithm based on artificial intelligence; and meanwhile, semantic coding is performed on the learning demand description, so that the type of the learning material is intelligently recommended according to rapid scanning semantic query representation between the semantic coding features of the learning demand of the user and the semantic coding features of the wrong content segments answered by the user. Therefore, the learning demand of the user and the specific content of the answer error can be deeply understood, and the semantic interaction information between the learning demand and the answer error can be captured, so that the knowledge blind spot and the learning demand of the user can be accurately identified, the learning material suitable for the user can be accurately recommended in real time, and the pertinence and effectiveness of recommendation are improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent assistance, and more specifically, to an intelligent assistance system and method for teaching gastrointestinal nursing. Background Art

[0002] Gastrointestinal care refers to a series of medical care activities for the digestive system (including the esophagus, stomach, small intestine, large intestine, etc.), which aims to prevent, diagnose and treat digestive system diseases and ensure the health and normal function of the patient's digestive system. Through systematic gastrointestinal care training, medical staff can not only master the basic knowledge and latest developments of gastrointestinal care, but also improve their clinical practice and emergency response capabilities, thereby reducing medical errors and improving patient safety and satisfaction.

[0003] However, with the increasing demand for medical education and training, traditional gastrointestinal nursing teaching methods can no longer meet the personalized and efficient needs of modern medical education. Especially in the professional field of gastrointestinal nursing, medical staff need to master a lot of professional knowledge and operational skills, and traditional teaching models often cannot provide personalized guidance and assistance for each student's specific situation.

[0004] Therefore, an intelligent auxiliary solution for gastrointestinal nursing teaching is desired, which can provide personalized guidance and recommendations to help each user conduct targeted learning, thereby improving the overall teaching quality and effectiveness. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed.

[0006] According to one aspect of the present application, an intelligent auxiliary system for teaching gastrointestinal nursing is provided, which includes: User registration module, used for user account registration; An evaluation test module, used for publishing evaluation test questions and receiving answers input by users, and generating an evaluation test report based on the evaluation test questions and the answers input by users; A learning requirement collection module is used to collect learning requirement descriptions input by users; A personalized learning assistance module, for recommending learning materials based on the assessment test report and the learning needs description; Among them, the personalized learning assistance module includes: an error part extraction unit, which is used to extract the part of the user's incorrect answer from the evaluation test report; an error content segmentation encoding unit, which is used to perform content segmentation and semantic encoding on the part of the user's incorrect answer to obtain a sequence of semantic encoding features of the user's incorrect content fragments; a learning need description encoding unit, which is used to semantically encode the learning need description to obtain the user's learning need semantic encoding features; a learning need-error content semantic query encoding unit, which is used to quickly scan the sequence of the user's learning need semantic encoding features and the user's answer error content fragment semantic encoding features to obtain the learning need-error content query response semantic features; a learning material recommendation unit, which is used to obtain the recommendation result of the type of learning material based on the learning need-error content query response semantic features.

[0007] Furthermore, the error content segmentation encoding unit includes: A user answer incorrect content segmentation subunit, used to segment the content of the user answer incorrectly based on the question as a unit to obtain a set of user answer incorrect content segments; The incorrect content segment semantic coding subunit is used to semantically code each incorrect user answer content segment in the set of incorrect user answer content segments to obtain a sequence of semantic coding feature vectors of incorrect user answer content segments as a sequence of semantic coding features of incorrect user answer content segments.

[0008] Furthermore, the erroneous content fragment semantic encoding subunit is used to: input each user answer erroneous content fragment in the set of user answer erroneous content fragments into an erroneous content fragment semantic encoder containing a Bert model to obtain a sequence of semantic encoding feature vectors of the user answer erroneous content fragments as semantic encoding features of user learning needs.

[0009] Furthermore, the learning need description encoding unit is used to: input the learning need description into a learning need description semantic encoder including a Bert model to obtain a user learning need semantic encoding feature vector as the user learning need semantic encoding feature.

[0010] Furthermore, the learning requirement-error content semantic query encoding unit includes: A learning need-error content semantic mutual information calculation subunit, used to calculate the mutual information between the user learning need semantic encoding feature vector and each user answer error content segment semantic encoding feature vector in the sequence of the user answer error content segment semantic encoding feature vector to obtain a sequence of learning need-error content semantic fast semantic matching factors; A learning need-error content semantic subsequence determination subunit, used to determine a subsequence of semantic encoding feature vectors of incorrect content segments answered by the user based on a sequence of learning need-error content semantic fast semantic matching factors; The learning need-error content query response subunit is used to perform cross-domain query response encoding on a subsequence of the user's learning need semantic encoding feature vector and the user's answer error content fragment semantic encoding feature vector to obtain a learning need-error content query response semantic encoding vector as the learning need-error content query response semantic feature.

[0011] Furthermore, the learning requirement-error content semantic subsequence determination subunit is used to: Identify a first learning requirement-error content semantic maximum value and a second learning requirement-error content semantic maximum value from the sequence of the learning requirement-error content semantic fast semantic matching factors; Based on the positions of the first learning requirement-erroneous content semantic maximum and the second learning requirement-erroneous content semantic maximum in the sequence of learning requirement-erroneous content semantic fast semantic matching factors, a subsequence of the semantic encoding feature vectors of the user's incorrect answer content fragments that is quickly matched is determined from the sequence of semantic encoding feature vectors of the user's incorrect answer content fragments.

[0012] Furthermore, the learning need-error content query response sub-unit is used to: perform a linear transformation on the user learning need semantic encoding feature vector to obtain a user learning need semantic query vector and a user learning need semantic value vector, and use a subsequence of the semantic encoding feature vector of the user's answer error content fragment as a subsequence of the key vector, and input the query vector, value vector and the subsequence of the key vector into a cross-domain query encoding module based on a converter structure to obtain the learning need-error content query response semantic encoding vector.

[0013] Furthermore, the learning material recommendation unit is used to: input the learning requirement-error content query response semantic encoding vector into a classifier-based learning material recommender to obtain the recommendation result, which is a type label of the learning material.

[0014] According to another aspect of the present application, an intelligent auxiliary method for teaching gastrointestinal nursing is provided, which includes: Extract the incorrect answers from the evaluation test report; Performing content segmentation and semantic coding on the part where the user gives an incorrect answer to obtain a sequence of semantic coding features of the content segment where the user gives an incorrect answer; Semantically encode the learning demand description to obtain the semantic encoding features of the user's learning demand; Performing a quick scan semantic query encoding on the sequence of the semantic coding features of the user's learning needs and the semantic coding features of the incorrect content segments answered by the user to obtain a learning needs-incorrect content query response semantic feature; Based on the learning need-error content query response semantic features, a recommendation result of the type of learning material is obtained.

[0015] Furthermore, content segmentation and semantic coding are performed on the part of the user's incorrect answer to obtain a sequence of semantic coding features of the content segment of the user's incorrect answer, including: Segmenting the content of the part in which the user gives an incorrect answer based on the question to obtain a set of content segments in which the user gives an incorrect answer; Semantically encode each incorrect user answer content segment in the set of incorrect user answer content segments to obtain a sequence of semantic encoding feature vectors of incorrect user answer content segments as a sequence of semantic encoding features of incorrect user answer content segments.

[0016] Compared with the prior art, the intelligent auxiliary system and method for teaching gastrointestinal nursing provided by the present application first extracts the part of the user's wrong answer from the evaluation test report, and uses artificial intelligence-based semantic analysis and processing algorithms to perform content segmentation and semantic coding on it, and semantically encodes the learning needs description at the same time, so as to intelligently recommend the type of learning materials based on the semantic coding features of the user's learning needs and the semantic coding features of the content fragments of each user's wrong answer. In this way, it is possible to deeply understand the user's learning needs and the specific content of the wrong answer and capture the semantic interaction information between the two, so as to accurately identify the user's knowledge blind spots and learning needs, thereby accurately and in real time recommending learning materials suitable for the user, improving the pertinence and effectiveness of the recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 This is a system block diagram of an intelligent assistance system for teaching gastrointestinal nursing according to an embodiment of the present application.

[0019] Figure 2 It is a block diagram of a personalized learning assistance module in an intelligent assistance system for teaching gastrointestinal nursing according to an embodiment of the present application.

[0020] Figure 3This is a data flow diagram of a personalized learning assistance module in an intelligent assistance system for gastrointestinal nursing instruction according to an embodiment of the present application.

[0021] Figure 4 It is a block diagram of an error content segmentation encoding unit in an intelligent assistance system for gastrointestinal nursing instruction according to an embodiment of the present application.

[0022] Figure 5 The present invention is a block diagram of a learning requirement-error content semantic query encoding unit in an intelligent assistance system for gastrointestinal nursing instruction according to an embodiment of the present application.

[0023] Figure 6 The present invention is a flowchart of an intelligent auxiliary method for teaching gastrointestinal nursing according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] Various exemplary embodiments, features and aspects of the present application 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.

[0025] 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.

[0026] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0027] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0028] Gastrointestinal care refers to a series of medical care activities for the digestive system (including the esophagus, stomach, small intestine, large intestine, etc.). However, in the traditional teaching methods of gastrointestinal care, especially in the professional field of gastrointestinal care, medical staff need to master a lot of professional knowledge and operational skills, and traditional teaching methods often cannot provide targeted guidance and support according to the specific needs of each student.

[0029] In response to the above technical problems, the present application proposes an intelligent auxiliary system for teaching gastrointestinal nursing. Figure 1 FIG. 1 is a system block diagram of an intelligent auxiliary system for teaching gastrointestinal nursing according to an embodiment of the present application. Figure 1 As shown, the intelligent assistance system 100 for teaching gastrointestinal nursing according to the embodiment of the present application includes: a user registration module 110, which is used for user account registration; an evaluation test module 120, which is used to publish evaluation test questions and receive answers entered by the user, and generate an evaluation test report based on the evaluation test questions and the answers entered by the user; a learning needs collection module 130, which is used to collect learning needs descriptions entered by the user; and a personalized learning assistance module 140, which is used to recommend learning materials based on the evaluation test report and the learning needs description.

[0030] In the above-mentioned intelligent assistance system 100 for teaching gastrointestinal nursing, the user registration module 110 is used for user account registration. It should be understood that in the intelligent assistance system for teaching gastrointestinal nursing, user registration can help the system identify the identities of different users, such as caregivers, patients or medical professionals. By assigning a unique account to each user, the system can track the user's learning progress, record learning history, analyze behavioral data, and provide tailored nursing knowledge and suggestions. At the same time, the system can adjust the teaching content according to the learning progress and feedback of each user to ensure the effective absorption of their knowledge.

[0031] Among them, the user registration module provides a safe and convenient way for users to create personal accounts so that they can access various resources and services on the platform. The design and implementation of the user registration module are described in detail below.

[0032] Before you start designing the user registration process, you first need to clarify who the target user group of the system is. For the gastrointestinal nursing teaching platform, the main users may include but are not limited to working medical staff, medical students, and the public who are interested in gastrointestinal health. Based on different user backgrounds, the type of information collected during the registration process will also be different. For example, for professionals, in addition to basic personal information (such as name, email address), you may also need to know their work unit, position level, existing qualifications, etc.; for ordinary users, it is more focused on understanding their interests and learning purposes.

[0033] Next, it is time to design the specific registration interface. A good registration page should be concise and clear to avoid confusing or annoying users. Usually, the whole process can be divided into several simple steps: the first is to fill in basic information, including username and password settings, etc.; the second is the verification step, which confirms the authenticity of the identity by sending a verification code to the email or mobile phone number provided by the user; the last step may be to choose the subscription service type or accept the terms of use, etc. In order to improve the conversion rate, you can also consider providing a direct login option for social accounts.

[0034] Security is an aspect that cannot be ignored in the user registration module. First of all, users should be forced to use a more complex combination of passwords and be reminded to change them regularly. In addition, all transmitted data must be encrypted to prevent it from being intercepted during transmission. Strict security measures must also be taken on the server side, such as using the HTTPS protocol to ensure communication security, and deploying firewalls and intrusion detection systems to resist external attacks. For sensitive information such as medical records, relevant laws and regulations must also be followed to ensure compliance with storage and management.

[0035] In order to allow new users to successfully complete the registration and integrate into the community as soon as possible, the system can guide them to perform some preliminary operations immediately after successful registration, such as completing personal information, participating in the novice guidance course, etc. This will not only help users familiarize themselves with the platform functions, but also make them feel a warm and friendly atmosphere, increasing retention rate. At the same time, sending a welcome email to new users is also a good choice, which can include platform introductions, usage guide links, etc.

[0036] Considering that users may forget their passwords or encounter other problems that prevent them from logging in normally, it is very necessary to add a password retrieval function to the registration module. This function is usually implemented in two ways: one is to reset the password by answering preset security questions; the other is to send a one-time link to the user's registered email or mobile phone, and click it to jump to the password change page. No matter which method is used, it should be easy to operate and safe and reliable.

[0037] In this way, the design of the user registration module can comprehensively consider multiple factors such as user experience, security, privacy protection, etc. By carefully polishing the details, it can provide users with a convenient and reliable entrance, laying a solid foundation for the subsequent learning journey.

[0038] In the above-mentioned intelligent assistance system 100 for teaching gastrointestinal nursing, the evaluation test module 120 is used to publish evaluation test questions and receive answers entered by the user, and generate an evaluation test report based on the evaluation test questions and the answers entered by the user. It should be understood that publishing evaluation test questions and receiving answers entered by the user can evaluate the user's learning effect, discover potential knowledge gaps, and provide personalized feedback to the user. Furthermore, through the evaluation test, the system can detect the user's knowledge mastery in the field of gastrointestinal nursing, as well as the depth of understanding of nursing operation specifications and treatment plans. This not only helps to ensure that users master core knowledge, but also helps the system identify which areas require further training and improvement, and provide more targeted teaching resources.

[0039] In order to measure the user's current knowledge level and skill mastery, and then provide personalized learning suggestions for each user, it is necessary to design an assessment test module. The following is a detailed description of how to implement this module.

[0040] First of all, the first step in building an evaluation test module is to comprehensively sort out the knowledge system in the field of gastrointestinal nursing. This includes but is not limited to anatomy, physiology, common diseases and their treatment methods, the latest research progress, etc. Based on these contents, a comprehensive and structured question bank is developed, covering different levels of knowledge points from basic theory to advanced application. The types of questions can be diversified, such as multiple-choice questions, true-or-false questions, short-answer questions and even case analysis questions to meet the needs of different learning objectives. In addition, considering the importance of operational skills in gastrointestinal nursing practice, some virtual simulation exercises or questions after video demonstrations can also be set to evaluate the user's actual operational ability.

[0041] Next, once the user completes registration and logs into the system, they will be directed to the assessment test interface. At this point, the system will automatically filter out the most appropriate initial test set based on the user's basic information (such as years of work experience, existing qualifications, etc.). Of course, users are also allowed to manually select specific topics for testing based on their personal needs. Once the user starts answering, all answers will be recorded instantly and will receive scoring feedback immediately after submission.

[0042] Furthermore, as users complete each round of testing, the system background will use advanced data analysis technology to process the collected data. The technologies involved here mainly include natural language processing (NLP), machine learning algorithms, etc. For example, when dealing with short-answer questions, NLP can help identify key words, while machine learning models can understand the contextual semantics, thereby giving more accurate performance assessments. More importantly, by learning from a large amount of user data, the system can continuously optimize its evaluation criteria so that each evaluation result is as close to the actual situation as possible.

[0043] Furthermore, the generated evaluation report includes overall score, knowledge point coverage, strengths and weaknesses analysis, learning path recommendation and personalized prompts. Among them, the overall score shows the user's performance overview in the entire test. The knowledge point coverage lists which areas the user has mastered well and which ones need to be strengthened. The strengths and weaknesses analysis points out which specific questions the user performed poorly on and provides relevant suggestions. Based on the above analysis results, the learning path recommendation proposes the next learning plan, including recommended courses, books, online resource links, etc. Personalized prompts provide additional guidance based on the user's personal characteristics. For example, if a user is found to be weak in memorizing knowledge, it is recommended to use more visual memory methods; if they have difficulties in solving complex problems, they are encouraged to participate in group discussions and other activities.

[0044] It is worth mentioning that in order to ensure the effectiveness and practicality of the evaluation report, this application sets up a user feedback mechanism to encourage opinions on problems encountered during the experience so as to adjust the improvement strategy in a timely manner. In addition, it is worth noting that although there will be a comprehensive round of evaluation when entering the system for the first time, as time goes by and as the user's learning progress changes, the evaluation test module will also adjust its function accordingly. For example, a periodic review mechanism can be set up to allow users to retake tests of similar difficulty levels at regular intervals to monitor their growth trajectory. At the same time, the difficulty of the content of subsequent tests can also be dynamically adjusted according to the user's daily learning behavior (such as the length of time watching video tutorials, the speed of completing homework, etc.), so as to truly teach students in accordance with their aptitude.

[0045] In this way, users can clearly understand their own strengths and weaknesses, and promote the overall improvement of the comprehensive quality of medical staff, ultimately achieving the goal of improving the quality of patient care.

[0046] In the above-mentioned intelligent assistance system 100 for teaching gastrointestinal nursing, the learning needs collection module 130 is used to collect the learning needs description input by the user. It should be understood that, considering that each user's background, experience, job responsibilities and learning goals are different, the knowledge field involved in gastrointestinal nursing is vast, covering multiple aspects from basic nursing to clinical treatment. By collecting the user's learning needs description, the system can better understand the user's specific needs, and then adjust the course content, recommend suitable learning resources, and provide timely feedback and guidance during the teaching process.

[0047] In building an intelligent assistance system for gastrointestinal nursing teaching, the learning needs collection module can gain in-depth understanding of each user's specific learning goals, interests, and personal preferences, thereby providing users with a more personalized and targeted learning experience. The following details how to effectively implement the learning needs collection module.

[0048] Among them, after users complete the registration and have a preliminary understanding of the functions of the system, they will be guided to a specially designed learning needs collection interface. Here, users can express their learning wishes in a variety of ways. A common method is to provide an open text box that allows users to freely enter the knowledge areas they want to acquire, the specific problems they want to solve, or the specific technologies they are interested in. The advantage of this form is that it is highly flexible and can cover almost all types of requests; however, it also requires users to have a certain degree of self-reflection and be able to clearly express their needs. In order to help those who may not be sure where to start, the system can also preset some prompt questions, such as: "Which skills do you most want to improve at present?", "Have you encountered any particularly difficult situations that you hope to improve through learning?", etc., to stimulate users' thinking.

[0049] In addition to text descriptions, the module can also introduce graphical selection tools, considering the high efficiency of visual information transmission and the ease of understanding. For example, using charts to show the relationship network between different topics, users can intuitively see which areas are interrelated and indicate their interests by clicking on the corresponding nodes. In addition, a slider or star rating system can be set to allow users to quantify their interest in a topic, which helps the system to more accurately grasp the user's true thoughts.

[0050] In order to make the information collection process more interactive and interesting, and to obtain more comprehensive data, small games or quizzes can be added during the learning needs collection stage. For example, a series of scenario simulation questions can be designed, which include several different gastrointestinal care scenarios. Users need to choose the answer that best suits their current status based on their actual situation. In this way, users can not only complete the task in a relaxed and pleasant atmosphere, but also learn more about their work experience, knowledge level, etc., thus laying the foundation for the subsequent recommendation of more accurate learning content.

[0051] After users submit their learning needs descriptions, the data needs to be processed and analyzed. First, natural language processing technology is used to parse open texts, extract key words and semantic features, and then identify the user's main concerns. Then, combined with the user's performance in other parts (such as evaluation test results), a user portrait model is constructed using machine learning algorithms. This model comprehensively considers multiple dimensions such as the user's professional background, existing knowledge structure, and learning style preferences, and can more accurately predict the user's learning trends and development directions in the future.

[0052] In this way, based on each user's unique learning needs description, the learning needs collection module can effectively capture the personalized needs characteristics of each user and tailor a learning path for them accordingly.

[0053] In the above-mentioned intelligent assistance system 100 for teaching gastrointestinal nursing, the personalized learning assistance module 140 is used to recommend learning materials based on the evaluation test report and the learning needs description. It should be understood that the evaluation test report can reflect the user's mastery of a specific knowledge field. At the same time, the learning needs description provides the system with the user's specific learning goals and areas of interest. In this way, the system can first identify the user's weaknesses in certain knowledge areas by analyzing the user's evaluation test report; then, combined with the learning needs description provided by the user, the system will match these needs with the existing learning resource library and screen out the most suitable materials; finally, the system pushes these personalized recommended learning resources to the user to support his further learning. This personalized recommendation method based on the evaluation test report and the learning needs description can accurately meet the user's learning needs, improve learning efficiency, shorten the learning curve, and ensure that users obtain continuous and systematic knowledge updates and skills improvement in the field of gastrointestinal nursing.

[0054] That is, by recommending learning materials suitable for each user's level and background, it is possible to solve the user's learning problems and improve the user's skill level in a targeted manner. However, traditional learning material recommendation systems are usually based on fixed rules or simple keyword matching, and are unable to make personalized recommendations based on the semantic association between the user's specific needs and knowledge blind spots, resulting in poor accuracy and relevance of the recommendation results. In addition, traditional recommendation systems provide static recommendation lists and are unable to dynamically adjust the recommended content based on the user's learning progress and feedback, making it difficult to adapt to the user's changing learning needs.

[0055] Based on this, in the learning needs collection module, the technical concept of the present application is to extract the part of the user's wrong answer from the evaluation test report, and use artificial intelligence-based semantic analysis and processing algorithms to perform content segmentation and semantic coding on the part of the user's wrong answer, and at the same time perform semantic coding on the learning needs description, so as to intelligently recommend the type of learning materials based on the semantic coding features of the user's learning needs and the fast scanning semantic query representation between the semantic coding features of the content fragments of each user's wrong answer. In this way, it is possible to deeply understand the user's learning needs and the specific content of the wrong answer and capture the semantic interaction information between the two, so as to accurately identify the user's knowledge blind spots and learning needs, thereby accurately and in real time recommending learning materials suitable for the user, improving the pertinence and effectiveness of the recommendation.

[0056] Figure 2It is a block diagram of a personalized learning assistance module in an intelligent assistance system for teaching gastrointestinal nursing according to an embodiment of the present application. Figure 3 FIG. 1 is a data flow diagram of a personalized learning assistance module in an intelligent assistance system for gastrointestinal nursing teaching according to an embodiment of the present application. Figure 2 and Figure 3 As shown, the personalized learning assistance module 140 includes: an error part extraction unit 141, which is used to extract the part of the user's incorrect answer from the evaluation test report; an error content segmentation encoding unit 142, which is used to perform content segmentation and semantic encoding on the part of the user's incorrect answer to obtain a sequence of semantic encoding features of the user's incorrect content fragments; a learning need description encoding unit 143, which is used to semantically encode the learning need description to obtain the user's learning need semantic encoding features; a learning need-incorrect content semantic query encoding unit 144, which is used to quickly scan the sequence of the user's learning need semantic encoding features and the user's incorrect content fragment semantic encoding features to obtain the learning need-incorrect content query response semantic features; a learning material recommendation unit 145, which is used to obtain a recommendation result of the type of learning material based on the learning need-incorrect content query response semantic features.

[0057] Specifically, in the error part extraction unit 141, the part where the user answered incorrectly is extracted from the evaluation test report. It should be understood that the correct answer to each question represents the standard understanding of the knowledge point, while the wrong answer reveals the user's possible understanding bias, memory loopholes or operational errors. By analyzing the type and distribution of wrong answers, the system can determine which knowledge points or skills are the user's weak links. For example, if the questions that the user frequently makes mistakes in multiple tests are about gastrointestinal endoscopy operation specifications, the system can identify that the user has difficulties in understanding or operating in this field. In this way, extracting the part where the user answered incorrectly from the evaluation test report can help the system identify the user's weak links in specific knowledge fields or skill operations, thereby providing users with more accurate learning support and personalized improvement suggestions. By analyzing the user's wrong answers, the system can provide real-time feedback and guide the user to strengthen these deficient parts, thereby improving the learning effect and the mastery of nursing skills.

[0058] Figure 4 FIG. 1 is a block diagram of an error content segmentation encoding unit in an intelligent auxiliary system for gastrointestinal nursing teaching according to an embodiment of the present application. Figure 4As shown, in an embodiment of the present application, the error content segmentation encoding unit 142 includes: a user answer error content segmentation subunit 1421, which is used to perform content segmentation on the part of the user's incorrect answer in units of questions to obtain a set of user answer error content segments; an error content segment semantic encoding subunit 1422, which is used to perform semantic encoding on each user answer error content segment in the set of user answer error content segments to obtain a sequence of user answer error content segment semantic encoding feature vectors as a sequence of user answer error content segment semantic encoding features.

[0059] Specifically, in the user's incorrect answer content segmentation subunit 1421, the part of the user's incorrect answer is segmented in units of questions to obtain a collection of user's incorrect answer content segments. It should be understood that, considering that the part of the user's incorrect answer contains each question answer information, and specific questions have different error points, that is, each question usually corresponds to a specific knowledge point or skill point. Therefore, the user's error is pinpointed to a specific question and knowledge point, so as to more accurately locate the user's knowledge blind spot, so as to improve the accuracy of subsequent learning material recommendations. In the technical solution of the present application, the part of the user's incorrect answer is segmented in units of questions to obtain a collection of user's incorrect answer content segments. In this way, the user's misunderstanding and cognitive bias can be more accurately captured, providing reliable data support for subsequent personalized learning recommendations.

[0060] Specifically, in the erroneous content segment semantic coding subunit 1422, semantic coding is performed on each erroneous content segment of the user's answer in the set of erroneous content segments of the user's answer to obtain a sequence of semantic coding feature vectors of the erroneous content segments of the user's answer as a sequence of semantic coding features of the erroneous content segments of the user's answer. It should be understood that in order to further capture and refine the key and important semantic expressions of erroneous content contained in each erroneous content segment of the user's answer, in the technical solution of the present application, semantic coding is performed on each erroneous content segment of the user's answer in the set of erroneous content segments of the user's answer to mine and extract deeper semantic meanings, thereby obtaining a sequence of semantic coding feature vectors of erroneous content segments of the user's answer.

[0061] In an embodiment of the present application, the erroneous content segment semantic encoding subunit 1422 is used to: input each user's erroneous answer content segment in the set of the user's erroneous answer content segments into the erroneous content segment semantic encoder containing the Bert model to obtain a sequence of semantic encoding feature vectors of the user's erroneous answer content segments as user learning demand semantic encoding features. In this way, by processing the user's erroneous answer content segments using a semantic encoder containing the Bert model and encoding them through a multi-layer Transformer structure, rich semantic information in the context of the user's erroneous answer content can be captured, so that the generated features have high semantic expression capabilities.

[0062] In an embodiment of the present application, the learning need description encoding unit 143 is used to: input the learning need description into a learning need description semantic encoder containing a Bert model to obtain a user learning need semantic encoding feature vector as the user learning need semantic encoding feature. It should be understood that, similarly, considering that the learning need description usually contains key semantic information such as the user's intentions, goals and specific requirements, therefore, in order to capture deeper semantic information to more accurately understand the user's real needs, in the technical solution of the present application, the learning need description is semantically encoded to obtain a user learning need semantic encoding feature vector. In particular, in a specific embodiment of the present application, the learning need description is input into a learning need description semantic encoder containing a Bert model for processing to extract the contextual semantic information in the description, so that the user's specific needs and goals can be more accurately identified.

[0063] Specifically, in the learning needs-wrong content semantic query encoding unit 144, the sequence of the semantic encoding features of the user's learning needs and the semantic encoding features of the wrong content segments answered by the user is quickly scanned for semantic query encoding to obtain the learning needs-wrong content query response semantic features. It should be understood that considering that the user's learning needs semantic encoding feature vector contains the semantic information of the specific goals and intentions expressed by the user in the learning needs description, it reflects the user's learning interests and preferences. The sequence of semantic encoding feature vectors of the wrong content segments answered by the user reflects the user's misunderstanding, lack of knowledge or misunderstanding of the user on a specific knowledge point. Therefore, in order to extract the most relevant semantic information and features of the user's learning needs semantic encoding features from the sequence of semantic encoding features of the wrong content segments answered by the user, so as to screen out highly relevant wrong content segments, thereby accurately identifying the user's knowledge blind spots and needs, and providing personalized learning resource recommendations, in the technical solution of the present application, the sequence of the semantic encoding features of the user's learning needs and the semantic encoding features of the wrong content segments answered by the user are quickly scanned for semantic query encoding to obtain the learning needs-wrong content query response semantic features.

[0064] Figure 5 FIG. 1 is a block diagram of a learning requirement-error content semantic query encoding unit in an intelligent auxiliary system for gastrointestinal nursing teaching according to an embodiment of the present application. Figure 5 As shown, in an embodiment of the present application, the learning need-wrong content semantic query encoding unit 144 includes: a learning need-wrong content semantic mutual information calculation subunit 1441, which is used to calculate the mutual information between the user learning need semantic encoding feature vector and each user answer wrong content segment semantic encoding feature vector in the sequence of the user answer wrong content segment semantic encoding feature vector to obtain a sequence of learning need-wrong content semantic fast semantic matching factors; a learning need-wrong content semantic subsequence determination subunit 1442, which is used to determine a subsequence of the user answer wrong content segment semantic encoding feature vector based on the sequence of the learning need-wrong content semantic fast semantic matching factors; a learning need-wrong content query response subunit 1443, which is used to perform cross-domain query response encoding on the subsequence of the user learning need semantic encoding feature vector and the user answer wrong content segment semantic encoding feature vector to obtain a learning need-wrong content query response semantic encoding vector as the learning need-wrong content query response semantic feature.

[0065] Specifically, in the learning need-wrong content semantic mutual information calculation subunit 1441, the mutual information between the semantic encoding feature vectors of the user learning need and the semantic encoding feature vectors of the user answer wrong content segments in the sequence of semantic encoding feature vectors of the user answer wrong content segments is calculated to obtain a sequence of learning need-wrong content semantic fast semantic matching factors. It should be understood that in the technical solution of the present application, mutual information is used to evaluate the semantic similarity between different feature vectors. The larger the mutual information value, the more information overlaps between the two feature vectors, that is, to some extent, the concepts or information represented by the two feature vectors are highly correlated.

[0066] More specifically, the calculation process of the learning demand-error content semantic mutual information calculation subunit 1441 can be expressed as follows: ; ; ; ; in, A sequence of semantic encoding feature vectors of incorrect content segments answered by the user, , , and are the first, second, and third in the sequence of semantic encoding feature vectors of the user's wrong answer content fragments. and The semantic encoding feature vector of the user’s wrong answer content fragment, is the semantic encoding feature vector of the user's learning needs, yes function, is the semantic encoding transformation feature vector of user learning needs, yes The corresponding semantic encoding transformation feature vector of the user's wrong answer content fragment, express The information entropy of express The information entropy of express and The joint entropy between express and Therefore, by calculating the mutual information, it is not only possible to identify the semantic encoding feature vector of the user's answer error content segment that is most relevant to the user's learning demand semantic encoding feature vector, but also to provide an objective standard for subsequent rapid matching in a quantitative manner.

[0067] More specifically, in the learning demand-wrong content semantic subsequence determination subunit 1442, based on the sequence of the learning demand-wrong content semantic fast semantic matching factors, a subsequence of the semantic encoding feature vector of the user's wrong answer content segment is determined. In an embodiment of the present application, the learning demand-wrong content semantic subsequence determination subunit 1442 is used to: identify the first learning demand-wrong content semantic maximum and the second learning demand-wrong content semantic maximum from the sequence of the learning demand-wrong content semantic fast semantic matching factors; based on the positions of the first learning demand-wrong content semantic maximum and the second learning demand-wrong content semantic maximum in the sequence of the learning demand-wrong content semantic fast semantic matching factors, determine the subsequence of the semantic encoding feature vector of the user's wrong answer content segment that is quickly matched from the sequence of the semantic encoding feature vector of the user's wrong answer content segment. It should be understood that the process of identifying the first learning demand-wrong content semantic maximum and the second learning demand-wrong content semantic maximum from the sequence of the learning demand-wrong content semantic fast semantic matching factors is actually to find the two semantic encoding feature vectors of the user's wrong answer content segment that are most semantically related to the user's learning demand semantic encoding feature vector. Based on the positions of the first learning requirement-incorrect content semantic maximum and the second learning requirement-incorrect content semantic maximum in the sequence of semantic fast semantic matching factors, a subsequence of the semantic encoding feature vectors of the incorrect content fragments of the user's answer that is quickly matched is determined from the sequence of semantic encoding feature vectors of the incorrect content fragments of the user's answer.

[0068] More specifically, the learning requirement-error content semantic subsequence determination subunit 1442 process can be expressed as follows: ; ; ; ; ; in, is a sequence of learning demand-error content semantic fast semantic matching factors, is the number of learning requirement-error content semantics fast semantics matching factors in the sequence of learning requirement-error content semantics fast semantics matching factors, To get the maximum value in the sequence, The first learning requirement is the maximum value of the semantics of the error content. The second learning requirement is the maximum semantic value of error content. To find the position of the maximum value in the sequence, To first determine the quick match position, Determine the quick match position for the second, represents the semantic encoding feature vector of the user's incorrect answer content segment corresponding to the first determined quick matching position, represents the semantic encoding feature vector of the user's incorrect answer content segment corresponding to the second determined quick matching position, It is a subsequence of semantic encoding feature vectors of incorrect user answer content segments. This process can be regarded as a preliminary feature selection or filtering. In this way, the system can focus resources on processing those semantic encoding feature vectors of incorrect user answer content segments that are most likely to have meaningful interactions with the semantic encoding feature vectors of user learning needs, thereby improving processing efficiency and the quality of results. The selection of the subsequence of semantic encoding feature vectors of incorrect user answer content segments for fast matching also reflects the system's judgment on the importance of input data, which helps to reduce noise interference and improve the generalization ability of the model.

[0069] More specifically, in the learning need-wrong content query response subunit 1443, a subsequence of the user learning need semantic coding feature vector and the user answer wrong content segment semantic coding feature vector is cross-domain query response encoded to obtain a learning need-wrong content query response semantic coding vector as the learning need-wrong content query response semantic feature. In an embodiment of the present application, the learning need-wrong content query response subunit 1443 is used to: perform a linear transformation on the user learning need semantic coding feature vector to obtain a user learning need semantic query vector and a user learning need semantic value vector, and use a subsequence of the user answer wrong content segment semantic coding feature vector as a subsequence of the key vector, and input the query vector, value vector and the subsequence of the key vector into a cross-domain query encoding module based on a converter structure to obtain the learning need-wrong content query response semantic coding vector. It should be understood that the user learning needs semantic encoding feature vector is linearly transformed to generate a user learning needs semantic query vector and a user learning needs semantic value vector, and the selected user answer error content fragment semantic encoding feature vector subsequence is used as a subsequence of the key vector, and then these vectors are input into the cross-domain query encoding module based on the converter structure. The core of the converter model lies in the self-attention mechanism, which allows the model to dynamically focus on different parts of the sequence when processing sequence data, without having to process each element sequentially like traditional recurrent neural networks. In the technical solution of the present application, the user learning needs semantic query vector is responsible for finding the key vector most relevant to itself, while the user learning needs semantic value vector carries the specific information of these related parts.

[0070] Specifically, the calculation process of the learning requirement-error content query response subunit 1443 can be expressed as follows: ; ; ; in, and are the query embedding matrix and the query bias vector respectively, is the user's learning demand semantic query vector, and are the value embedding matrix and value bias vector respectively, is the semantic value vector of user learning needs, Indicates The semantic encoding feature vector of the user’s wrong answer content fragment, yes The transposed vector of yes Length, yes function, is matrix multiplication, is the learning demand-error content query response semantic encoding vector. In this way, the transformer model can effectively capture and utilize the long-distance dependencies between the subsequences of the user learning demand semantic encoding feature vector and the user answer error content segment semantic encoding feature vector to generate an accurate query response semantic encoding vector.

[0071] In an embodiment of the present application, the learning material recommendation unit 145 is used to: input the learning demand-error content query response semantic coding vector into a learning material recommender based on a classifier to obtain the recommendation result, and the recommendation result is a type label of the learning material. That is, the learning demand-error content query response semantic features obtained by semantic querying the sequence of the user's learning demand semantic coding feature vector and the user's answer error content segment semantic coding feature vector are classified and processed to intelligently recommend the type of learning materials. In this way, the user's learning needs and the specific content of the wrong answer can be deeply understood and the semantic interaction information between the two can be captured to accurately identify the user's knowledge blind spots and learning needs, so as to accurately and in real time recommend learning materials suitable for the user, thereby improving the pertinence and effectiveness of the recommendation. It is worth mentioning that in a specific embodiment of the present application, the learning demand-error content query response semantic coding vector is input into a learning material recommender based on a classifier. The classifier can be a supervised learning model, such as a support vector machine (SVM), a decision tree, a random forest, or a deep learning model (such as a neural network). These models have been trained to determine which category the user's learning needs belong to based on the input learning needs-error content query response semantic encoding vector, and predict a category label that represents the type of user's learning needs. For example, the prediction result may be "gastrointestinal endoscopy operation skills" or "gastrointestinal disease treatment plan". According to these category labels, the system matches the corresponding learning materials (such as relevant tutorials, operation videos, case studies, etc.). Each category label corresponds to a specific set of learning materials, and the system pushes the resources that best meet the needs of users through a recommendation mechanism. These recommended materials include not only standard tutorials and learning videos, but also personalized adjustments based on the user's error content, such as providing additional exercises, simulated operations or learning tips to help users further strengthen their weak links. This process ensures that the recommended learning resources are highly relevant and help users effectively learn and improve in the areas where they need it most.

[0072] Here, in the case where the sequence of the user's learning need semantic encoding feature vector and the user's incorrect answer content fragment semantic encoding feature vector respectively represent the encoded semantic features of the learning need description and the user's incorrect answer content fragment, when performing semantic query encoding based on the fast scanning mechanism, considering the differences in fast scanning semantic queryability caused by differences in source semantic features, the learning need-incorrect content query response semantic encoding vector will also have richness in the response semantic encoding query distribution, resulting in repetition of unstructured probability mapping, affecting the accuracy of the classification results.

[0073] Therefore, preferably, when the learning need-error content query response semantic encoding vector is input into a classifier-based learning material recommender to obtain a recommendation result, the learning need-error content query response semantic encoding vector is optimized. Specifically, the optimization includes the following steps: The length of the learning requirement-error content query response semantic encoding vector Subtract the zero norm of the learning requirement-error content query response semantic encoding vector to determine the number of zero eigenvalues ​​of the learning requirement-error content query response semantic encoding vector , and the number of zero eigenvalues Subtract one to get the learning requirement - error content query response semantic encoding zero differentiable value ; Semantically encode the learning requirement-error content query response to zero differentiable value Multiply and divide by the number of zero eigenvalues ​​respectively Get the first learning requirement - error content query response semantic encoding field fitting value and the second learning requirement - error content query response semantic encoding field fitting value ; Calculate the square root of the sum of squares of all eigenvalues ​​of the learning requirement-error content query response semantic encoding vector to obtain the learning requirement-error content query response semantic encoding mode representation value ,in, The first semantic encoding vector representing the learning requirement-error content query response eigenvalues, represents the length of the learning requirement-error content query response semantic encoding vector, Representation learning needs - error content query response semantic encoding pattern representation value; Encoding zero differentiable values ​​with the learning requirement-error content query response semantics The power function of each eigenvalue of the learning demand-error content query response semantic encoding vector is calculated as an exponent , and multiply the first learning demand-error content query response semantic encoding field fitting value To obtain the first learning requirement - error content query response semantic encoding intermediate representation vector ; The learning requirement-error content query response semantic encoding vector and the second learning requirement-error content query response semantic encoding field fitting value and the learning requirement-error content query response semantic encoding mode representation value Perform point multiplication to obtain the second learning demand-error content query response semantic encoding intermediate representation vector ; Calculate the first learning requirement - error content query response semantic encoding intermediate representation vector The second learning requirement - error content query response semantic encoding intermediate representation vector The weighted points minus To obtain the optimized learning demand-error content query response semantic encoding vector, where and represents the weighted hyperparameter, It means point multiplication by position. Indicates adding by position point.

[0074] Finally, the optimized learning requirement-error content query response semantic encoding vector is input into a classifier-based learning material recommender to obtain the recommendation result.

[0075] That is, based on the vector field fitting law of high-dimensional manifold, the isolated zeros in the learning need-error content query response semantic coding vector are used as the differentiable dimension of the high-dimensional manifold, so as to fix the normal direction pointing to the distribution center near the local coordinates represented by the eigenvalues ​​of the learning need-error content query response semantic coding vector. In this way, the invalid repetition in the probability mapping process caused by the unstructured characteristics of the rich distribution of the learning need-error content query response semantic coding vector can be avoided by effectively aligning the characteristic distribution pattern of the learning need-error content query response semantic coding vector to the probability density differentiable field, thereby improving the accuracy of the recommendation results obtained by the classifier-based learning material recommender.

[0076] Therefore, the personalized learning auxiliary module 140 is explained, which first extracts the part of the user's wrong answer from the evaluation test report, and uses the semantic analysis and processing algorithm based on artificial intelligence to perform content segmentation and semantic coding, and semantically encodes the learning needs description, so as to intelligently recommend the type of learning materials based on the semantic coding features of the user's learning needs and the fast scanning semantic query representation between the semantic coding features of the content fragments of each user's wrong answer. In this way, it is possible to deeply understand the user's learning needs and the specific content of the wrong answer and capture the semantic interaction information between the two, so as to accurately identify the user's knowledge blind spots and learning needs, thereby accurately and in real time recommending learning materials suitable for the user, improving the pertinence and effectiveness of the recommendation.

[0077] In summary, the intelligent assistance system 100 for teaching gastrointestinal nursing based on the embodiment of the present application is explained, which first allows the user to create a personal account through the user registration module, and then publishes a series of evaluation test questions through the evaluation test module and receives the answers entered by the user, and generates a personalized evaluation test report based on these answers. At the same time, the learning needs collection module is responsible for collecting the user's specific learning needs description. Finally, the personalized learning assistance module will comprehensively evaluate the test report and the user's learning needs description to provide the user with customized learning material recommendations, thereby forming a complete process from user registration to personalized learning support. This series of steps ensures that each user can obtain educational resources that meet their abilities and needs, thereby improving the pertinence and effectiveness of the recommendations.

[0078] As described above, the intelligent assistance system 100 for teaching gastrointestinal nursing according to the embodiment of the present application can be implemented in various terminal devices. In one example, the intelligent assistance system 100 for teaching gastrointestinal nursing can be integrated into the terminal device as a software module and / or a hardware module. For example, the intelligent assistance system 100 for teaching gastrointestinal nursing can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the intelligent assistance system 100 for teaching gastrointestinal nursing can also be one of the many hardware modules of the terminal device.

[0079] Alternatively, in another example, the intelligent assistance system 100 for teaching gastrointestinal care and the terminal device may also be separate devices, and the intelligent assistance system 100 for teaching gastrointestinal care may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0080] Figure 6 FIG. 1 is a flow chart of an intelligent auxiliary method for teaching gastrointestinal nursing according to an embodiment of the present application. Figure 6As shown, according to the embodiment of the present application, the intelligent assistance method for teaching gastrointestinal nursing includes: S110, extracting the part of the user's incorrect answer from the evaluation test report; S120, performing content segmentation and semantic encoding on the part of the user's incorrect answer to obtain a sequence of semantic encoding features of the user's incorrect answer content fragments; S130, semantic encoding the learning need description to obtain the user's learning need semantic encoding features; S140, performing fast scanning semantic query encoding on the sequence of the user's learning need semantic encoding features and the user's incorrect answer content fragment semantic encoding features to obtain the learning need-incorrect content query response semantic features; S150, based on the learning need-incorrect content query response semantic features, obtaining a recommendation result of the type of learning material.

[0081] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned intelligent auxiliary method for teaching gastrointestinal nursing have been described in the above reference. Figures 1 to 5 The invention has been described in detail in the description of the intelligent assistance system for teaching gastrointestinal care, and therefore, its repeated description will be omitted.

[0082] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0083] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0084] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0085] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0086] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0087] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The second and other words are used to indicate names, but not to indicate any particular order.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit of the technical solution of the present invention.

Claims

1. An intelligent auxiliary system for teaching gastrointestinal nursing, characterized in that: include: User registration module, used for user account registration; An evaluation test module, used for publishing evaluation test questions and receiving answers input by users, and generating an evaluation test report based on the evaluation test questions and the answers input by users; A learning requirement collection module is used to collect learning requirement descriptions input by users; A personalized learning assistance module, for recommending learning materials based on the assessment test report and the learning needs description; Among them, the personalized learning assistance module includes: an error part extraction unit, which is used to extract the part of the user's incorrect answer from the evaluation test report; an error content segmentation encoding unit, which is used to perform content segmentation and semantic encoding on the part of the user's incorrect answer to obtain a sequence of semantic encoding features of the user's incorrect content fragments; a learning need description encoding unit, which is used to semantically encode the learning need description to obtain the user's learning need semantic encoding features; a learning need-error content semantic query encoding unit, which is used to quickly scan the sequence of the user's learning need semantic encoding features and the user's answer error content fragment semantic encoding features to obtain the learning need-error content query response semantic features; a learning material recommendation unit, which is used to obtain the recommendation result of the type of learning material based on the learning need-error content query response semantic features.

2. The intelligent auxiliary system for teaching gastrointestinal nursing according to claim 1, characterized in that: The error content segmentation encoding unit comprises: A user answer incorrect content segmentation subunit, used to segment the content of the user answer incorrectly based on the question as a unit to obtain a set of user answer incorrect content segments; The incorrect content segment semantic coding subunit is used to semantically code each incorrect user answer content segment in the set of incorrect user answer content segments to obtain a sequence of semantic coding feature vectors of incorrect user answer content segments as a sequence of semantic coding features of incorrect user answer content segments.

3. The intelligent auxiliary system for teaching gastrointestinal nursing according to claim 2, characterized in that: The erroneous content fragment semantic encoding subunit is used to: input each user answer erroneous content fragment in the set of user answer erroneous content fragments into the erroneous content fragment semantic encoder containing the Bert model to obtain a sequence of semantic encoding feature vectors of the user answer erroneous content fragments as the semantic encoding features of user learning needs.

4. The intelligent auxiliary system for teaching gastrointestinal nursing according to claim 3, characterized in that: The learning need description encoding unit is used to: input the learning need description into a learning need description semantic encoder including a Bert model to obtain a user learning need semantic encoding feature vector as the user learning need semantic encoding feature.

5. The intelligent auxiliary system for teaching gastrointestinal nursing according to claim 4, characterized in that: The learning requirement-error content semantic query encoding unit includes: A learning need-error content semantic mutual information calculation subunit, used to calculate the mutual information between the user learning need semantic encoding feature vector and each user answer error content segment semantic encoding feature vector in the sequence of the user answer error content segment semantic encoding feature vector to obtain a sequence of learning need-error content semantic fast semantic matching factors; A learning need-error content semantic subsequence determination subunit, used to determine a subsequence of semantic encoding feature vectors of incorrect content segments answered by the user based on a sequence of learning need-error content semantic fast semantic matching factors; The learning need-error content query response subunit is used to perform cross-domain query response encoding on a subsequence of the user's learning need semantic encoding feature vector and the user's answer error content fragment semantic encoding feature vector to obtain a learning need-error content query response semantic encoding vector as the learning need-error content query response semantic feature.

6. The intelligent auxiliary system for teaching gastrointestinal nursing according to claim 5, characterized in that: The learning requirement-error content semantic subsequence determination subunit is used to: Identify a first learning requirement-error content semantic maximum value and a second learning requirement-error content semantic maximum value from the sequence of the learning requirement-error content semantic fast semantic matching factors; Based on the positions of the first learning requirement-erroneous content semantic maximum and the second learning requirement-erroneous content semantic maximum in the sequence of learning requirement-erroneous content semantic fast semantic matching factors, a subsequence of the semantic encoding feature vectors of the user's incorrect answer content fragments that is quickly matched is determined from the sequence of semantic encoding feature vectors of the user's incorrect answer content fragments.

7. The intelligent auxiliary system for teaching gastrointestinal nursing according to claim 6, characterized in that: The learning need-error content query response subunit is used to: perform a linear transformation on the user learning need semantic coding feature vector to obtain a user learning need semantic query vector and a user learning need semantic value vector, and use a subsequence of the semantic coding feature vector of the user's incorrect answer content fragment as a subsequence of the key vector, and input the query vector, value vector and the subsequence of the key vector into a cross-domain query coding module based on a converter structure to obtain the learning need-error content query response semantic coding vector.

8. The intelligent auxiliary system for teaching gastrointestinal nursing according to claim 7, characterized in that: The learning material recommendation unit is used to: input the learning requirement-error content query response semantic encoding vector into a classifier-based learning material recommender to obtain the recommendation result, which is a type label of the learning material.

9. An intelligent auxiliary method for teaching gastrointestinal nursing, characterized in that: include: Extract the incorrect answers from the evaluation test report; Performing content segmentation and semantic coding on the part where the user gives an incorrect answer to obtain a sequence of semantic coding features of the content segment where the user gives an incorrect answer; Semantically encode the learning demand description to obtain the semantic encoding features of the user's learning demand; Performing a quick scan semantic query encoding on the sequence of the semantic coding features of the user's learning needs and the semantic coding features of the incorrect content segments answered by the user to obtain a learning needs-incorrect content query response semantic feature; Based on the learning need-error content query response semantic features, a recommendation result of the type of learning material is obtained.

10. The intelligent auxiliary method for teaching gastrointestinal nursing according to claim 9, characterized in that: The part of the user's incorrect answer is segmented and semantically encoded to obtain a sequence of semantic encoding features of the user's incorrect answer content segment, including: Segmenting the content of the part in which the user gives an incorrect answer based on the question to obtain a set of content segments in which the user gives an incorrect answer; Semantically encode each incorrect user answer content segment in the set of incorrect user answer content segments to obtain a sequence of semantic encoding feature vectors of incorrect user answer content segments as a sequence of semantic encoding features of incorrect user answer content segments.

Images