CN-121999666-A - Stomatology teaching system integrating knowledge graph and multi-mode intelligent interaction

CN121999666ACN 121999666 ACN121999666 ACN 121999666ACN-121999666-A

Abstract

The invention provides an oral medical teaching system integrating knowledge graphs and multi-mode intelligent interaction, which comprises an oral medical knowledge graph construction module and a multi-mode teaching resource semantic association engine, wherein the oral medical knowledge graph construction module is used for constructing a knowledge graph containing oral medical teaching entities and semantic relations based on acquired data, and the multi-mode teaching resource semantic association engine is used for carrying out semantic annotation on teaching resources and knowledge point nodes in the knowledge graph, and establishing a closed loop teaching link of knowledge points, explanation, cases and exercises. By constructing the knowledge graph, after the nodes of the knowledge graph are associated with teaching resources, a closed loop teaching link is established, so that the associated knowledge points in the teaching resources are associated, and the purposes of teaching resource information association and systematic knowledge network construction are achieved.

Inventors

XU MINGMING
CAO ZHANQIANG
JIAO JIAN
WU YUJIA
DENG XULIANG
WANG YING
HE YING
DU CHENLIN

Assignees

北京大学口腔医学院

Dates

Publication Date: 20260508
Application Date: 20260410

Claims (10)

1. An oral medicine teaching system integrating knowledge graph and multi-mode intelligent interaction, which is characterized by comprising: The stomatology knowledge graph construction module comprises a stomatology knowledge graph construction module, the method comprises the steps of constructing a knowledge graph containing oral medical teaching entities and semantic relations based on acquired data; the multi-mode teaching resource semantic association engine is used for carrying out semantic annotation on teaching resources and knowledge point nodes in the knowledge graph, and establishing a closed loop teaching link of knowledge points, explanation, cases and practice; The learning path generation engine is used for dynamically generating an optimal learning sequence according to the semantic relation in the learner image and the knowledge graph; The operation and real-time feedback module is used for collecting operation parameters of a learner in the virtual environment, judging operation normalization and feeding back associated knowledge points in the knowledge graph based on the operation data of the learner; the teaching feedback and reasoning path visualization module is used for generating a visual reasoning path based on the knowledge graph and performing difference comparison analysis when a student diagnoses errors; And the dynamic evaluation and knowledge evolution module is used for optimizing teaching effects and a knowledge system, tracking student behavior data, updating a student knowledge mastering model, automatically updating a knowledge graph based on feedback of students and teachers, and realizing self-evolution of the system.
2. The system for oral medical teaching of integrating knowledge-graph interaction with multimodal intelligence of claim 1, further comprising: And the AI virtual teaching aid module is used for providing intelligent questions and answers and coaching based on the entity and semantic relation of the knowledge graph.
3. The system for oral medical teaching integrating knowledge-graph and multi-modal intelligent interaction of claim 1, wherein the system comprises a plurality of sensors, The method for semantic annotation of teaching resources and knowledge point nodes in the knowledge graph comprises the following steps: Extracting operation action characteristics from teaching videos in teaching resources by using a 3D convolutional neural network, combining a voice recognition text, extracting medical entities by using BiLSTM-CRF models, and performing vector similarity matching with knowledge graph entities; Modeling a 3D grid topological structure by using a graph neural network for a 3D model in teaching resources, extracting semantic features of an anatomical structure, and performing graph embedding alignment with anatomical entities in a knowledge graph; extracting a main complaint-diagnosis-treatment triplet from a text by adopting a named entity recognition and relation extraction model for typical cases in teaching resources, and mapping the main complaint-diagnosis-treatment triplet to a disease-symptom-therapy sub-graph in a knowledge graph; And for the examination question bank in the teaching resources, the question text is converted into a logic expression through a semantic analyzer, and the concept-attribute-relation structure in the knowledge graph is matched.
4. The system of claim 1, wherein the means for semantically labeling teaching resources with knowledge point nodes in the knowledge graph comprises: Introducing a semantic consistency scoring function to quantify the matching degree of teaching resources and knowledge points, and setting a threshold filtering low-confidence label, wherein the semantic consistency scoring function is as follows: , Wherein, the For the purpose of teaching resources, Is a knowledge point node in the knowledge graph, Is a resource Is used to determine the multi-modal fusion embedded vector, Is a node Is embedded in the knowledge graph of the (2), For the cosine similarity it is the cosine similarity, To associate an entity to a node from a resource The most significant sense path existence probability in the knowledge graph, For the context consistency score, 、 And Is a weight coefficient.
5. The system of claim 2, wherein the means for dynamically generating an optimal learning sequence based on the learner representation and the semantic relationship in the knowledge graph comprises: constructing a student portrait vector; Extracting first-repair dependency relationships of semantic relationships in the knowledge graph to form a directed acyclic graph; defining a learning path optimization objective function based on the learner representation vector and the directed acyclic graph; searching the shortest path from the current mastering set to the target capability set on the knowledge graph based on the learning path optimization objective function and the path generation algorithm; and constructing an optimal learning sequence based on the shortest path.
6. The oral medical teaching system integrating knowledge-graph and multi-modal intelligent interaction according to claim 5, the heuristic function of the path generation algorithm is characterized in that: , Wherein, the Is a node To target capability set Is provided for the shortest path distance of (a), As a priority coefficient for points of weakness, In order for the knowledge to be able to master the vector, Is a node pair Is a grasping probability.
7. The system of claim 2, wherein the means for collecting the learner operation parameters in the virtual environment, determining the operation normalization, and feeding back the knowledge points associated with the knowledge graph based on the learner operation data comprises: Collecting parameters in stomatology; defining a local deviation index based on the acquired parameters so as to judge whether the current moment is illegal or not; Carrying out global track consistency test on the local deviation index; inquiring a normative principle node related to the current operation in the knowledge graph when the checking result triggers feedback; And feeding back the operation based on the standard principle of the node.
8. The system of claim 7, wherein the defining a local deviation index based on the collected parameters comprises: , Wherein, the For the included angle between the tool axis and the tooth long axis, Is the depth of cut, For the spatial trajectory of the tool tip, In order to achieve a cutting speed, the cutting speed, 、 And As a standard reference value for the operating procedure, 、 And In order to allow for a range of errors, 、 And As the weight coefficient of the light-emitting diode, Is a local deviation index.
9. The system of claim 2, wherein the system for generating visual reasoning paths and performing contrast analysis based on the knowledge-graph when a learner diagnoses errors comprises: Defining a diagnostic rationality scoring function to perform logic consistency verification of the knowledge graph, thereby judging whether the student diagnoses errors; the rationality scoring function is: , Wherein, the For the knowledge graph edge set involved in the path of the learner diagnosis, For the recommended level of the relationship in the clinical guideline, For the number of supporting documents or the level of evidence for the relationship, For specific diagnostic evidence or diagnostic elements, Are scored for rationality.
10. An oral medicine teaching method integrating knowledge graph and multi-mode intelligent interaction is characterized by comprising the following steps: constructing a knowledge graph containing the stomatology teaching entity and the semantic relation based on the acquired data; And carrying out semantic annotation on teaching resources and knowledge point nodes in the knowledge graph, and establishing a closed loop teaching link of knowledge points, explanation, cases and exercises.

Description

Stomatology teaching system integrating knowledge graph and multi-mode intelligent interaction Technical Field The invention belongs to the technical field of artificial intelligence, and particularly relates to an oral medicine teaching system integrating knowledge graph and multi-mode intelligent interaction. Background The current oral medical education mainly depends on traditional modes such as teaching materials, PPT demonstration, experiment courses, case discussion and the like, and has the problems of knowledge fragmentation, insufficient teaching individuation, weak clinical thinking training and the like. In recent years, some studies have attempted to use knowledge maps for medical information organization or auxiliary diagnosis, but have not been applied deeply to medical education scenes. The prior art mainly comprises: And the static teaching resource library is used for storing courseware, videos and question libraries in a centralized manner, and lacks semantic association and intelligent recommendation. Single function simulation systems, such as dental simulation head models or simple VR training systems, provide only an operating environment, lacking intelligent feedback and knowledge guidance. And the basic online learning platform supports course playing and testing, but cannot dynamically adjust the content according to the level of a learner, and lacks deep cognitive support. General AI question answering tools answer medical questions using large language models, but with "hallucination" risks, lack of fact constraints and evidence-based support. These schemes have the common problems of isolated knowledge, weak interactivity, unreliable feedback, lack of individuation and the like, and are difficult to meet the increasing demands of modern oral medical education on systemicity, intellectualization and safety. The prior art has the following remarkable defects: semantic association is lacking among teaching resources, and students have difficulty in establishing a systematic knowledge network; the operation training system lacks real-time and accurate knowledge feedback, and can not realize the closed loop of 'doing middle school'; The intelligent guidance tool lacks medical accuracy guarantee and has misguidance risks; The learning feedback mechanism is opaque, and students cannot understand the logic basis of the wrong decision; The teaching process lacks individualization and self-adaptation ability, and is difficult to teach in accordance with the material. Disclosure of Invention Aiming at the problems in the prior art, the invention provides an oral medical teaching system integrating knowledge graph and multi-mode intelligent interaction, which at least partially solves the problems of isolated teaching information and lack of association in the prior art. In a first aspect, an embodiment of the present disclosure provides an oral medical teaching system that fuses knowledge-graph and multi-modal intelligent interactions, including: The stomatology knowledge graph construction module comprises a stomatology knowledge graph construction module, the method comprises the steps of constructing a knowledge graph containing oral medical teaching entities and semantic relations based on acquired data; the multi-mode teaching resource semantic association engine is used for carrying out semantic annotation on teaching resources and knowledge point nodes in the knowledge graph, and establishing a closed loop teaching link of knowledge points, explanation, cases and practice; The learning path generation engine is used for dynamically generating an optimal learning sequence according to the semantic relation in the learner image and the knowledge graph; The operation and real-time feedback module is used for collecting operation parameters of a learner in the virtual environment, judging operation normalization and feeding back associated knowledge points in the knowledge graph based on the operation data of the learner; the teaching feedback and reasoning path visualization module is used for generating a visual reasoning path based on the knowledge graph and performing difference comparison analysis when a student diagnoses errors; And the dynamic evaluation and knowledge evolution module is used for optimizing teaching effects and a knowledge system, tracking student behavior data, updating a student knowledge mastering model, automatically updating a knowledge graph based on feedback of students and teachers, and realizing self-evolution of the system. In a second aspect, an embodiment of the present disclosure further provides an oral medical teaching method for integrating knowledge-graph and multi-modal intelligent interaction, including: constructing a knowledge graph containing the stomatology teaching entity and the semantic relation based on the acquired data; And carrying out semantic annotation on teaching resources and knowledge point nodes in the knowledge graph, and est