Intelligent computing service platform for digital teachers and teaching resource processing method

Abstract

The application discloses a digital teacher-oriented intelligent computing service platform and a teaching resource processing method. The platform comprises five modules of resource library management, classroom content collection and analysis, intelligent teaching plan generation, knowledge graph construction and personalized recommendation, and can realize the following functions: structured storage and dynamic management of teaching resources, multi-picture classroom content collection, live broadcast and voice transcription and subtitle generation, automatic generation of a structured teaching plan based on a verbatim script, hierarchical modeling of knowledge points and dynamic construction of a knowledge graph, generation of a personalized learning path and pushing of adaptive resources in combination with student learning behaviors and knowledge mastery. The application integrates multi-link intelligent teaching functions, solves the problems of low resource management efficiency, complicated teaching plan generation, lagging classroom analysis and single resource recommendation of existing education platforms, improves the teaching efficiency of digital teachers, provides personalized learning support for students, and is suitable for various teaching scenes such as traditional classrooms, online education and practical teaching.

Description

Technical Field

[0001] This invention relates to the fields of educational informatization and artificial intelligence, and in particular to an intelligent computing service platform and teaching resource processing method for digital teachers. Background Technology

[0002] With the development of educational informatization, smart education has become an important means to promote the modernization of education. Traditional teaching support systems are no longer adequate to meet the teaching needs of the new era, and digital teachers have become the core carrier for teaching model innovation. Currently, the teaching process suffers from problems such as scattered and unstructured teaching resources, inefficient manual lesson plan generation, resource recommendations failing to match students' personalized learning needs, a lack of intelligent means for collecting and analyzing classroom content, and difficulty in quickly locating and reviewing knowledge points. These issues lead to low efficiency in teachers' teaching resource management, time-consuming and labor-intensive lesson preparation, students' inability to receive accurate learning resource recommendations, and a significant reduction in the interactivity and effectiveness of remote and practical teaching.

[0003] The existing educational platforms' resource libraries cannot form effective connections, making it difficult to achieve digital and intelligent management; lesson plan generation can only complete simple text organization and cannot automatically generate structured and standardized lesson plans based on the teaching process; resource recommendations are mostly single content pushes, without personalized adaptation based on students' learning behavior and knowledge mastery; classroom recording systems also struggle to achieve real-time capture and switching of multiple screens, have insufficient accuracy in speech transcription, and require significant manual intervention for subsequent analysis and extraction of teaching content.

[0004] To address the aforementioned issues, there is an urgent need to construct an intelligent computing service platform for digital teachers that integrates intelligent resource management, intelligent lesson plan generation, classroom content collection and analysis, and intelligent processing of knowledge points. Through artificial intelligence, knowledge graphs, big data analysis, and multimodal processing technologies, this platform can achieve efficient management of teaching resources, intelligent recording and analysis of the teaching process, and personalized delivery of teaching content, thereby improving teaching efficiency and learning outcomes. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent computing service platform and teaching resource processing method for digital teachers. It integrates functions such as intelligent resource management, classroom content collection and analysis, intelligent lesson plan generation, intelligent knowledge point processing, and personalized resource recommendation. This solves the problems of inefficient resource management, cumbersome lesson plan generation, lagging classroom content analysis, and limited resource recommendations in existing education platforms. It improves the teaching efficiency of digital teachers, provides personalized learning support for students, and is adaptable to various teaching scenarios such as traditional classrooms, online education, and practical teaching.

[0006] To achieve the above objectives, the present invention provides the following solution: An intelligent computing service platform for digital teachers includes a resource library management module, a classroom content acquisition and analysis module, an intelligent lesson plan generation module, a knowledge graph construction module, and a personalized recommendation module. The resource library management module is used for structured storage, intelligent classification and dynamic management of various types of teaching resources, to build a multi-dimensional resource library system and realize cross-library association of resources; The classroom content acquisition and analysis module is used to realize the real-time acquisition, live streaming and storage of multi-screen classroom content, and to complete the accurate transcription of classroom audio content and the real-time generation of video subtitles; The intelligent lesson plan generation module is used to automatically generate structured and standardized lesson plans based on the verbatim transcript of classroom speech transcription, and supports the editing and export of lesson plans; The knowledge graph construction module is used to automatically label and hierarchically model knowledge points in teaching resources, construct a dynamically evolving knowledge graph, and realize the association and tracing of knowledge points. The personalized recommendation module is used to generate personalized learning paths in real time based on students' learning behavior data and knowledge mastery, combined with the relationships in the knowledge graph, and push suitable teaching resources.

[0007] Furthermore, the resource library management module includes a course resource library, a courseware resource library, a lesson plan library, an equipment and experimental resource model library, and a job knowledge library. Each resource library supports uploading resources in multiple formats, updating resources, version management, and cross-library sharing. The resource library management module is also equipped with an intelligent tagging unit and a resource association unit. The intelligent tagging unit uses artificial intelligence algorithms to automatically tag knowledge points of uploaded teaching resources, and the resource association unit uses knowledge graphs to achieve the association and matching of teaching resources in different resource libraries.

[0008] Furthermore, the knowledge graph of the resource library management module supports dynamic evolution driven by learning behavior. Specifically, it adjusts the edge weights between knowledge point nodes in the knowledge graph based on at least one learning behavior data, such as the student's answer accuracy, the number of times the knowledge point is repeatedly studied, and the time spent on the question. For knowledge points that students have mastered, it reduces the recommendation priority of their extended paths and weakens irrelevant learning paths.

[0009] Furthermore, the classroom content acquisition and analysis module supports real-time acquisition, preview, and layout customization of at least six video streams, including teacher panoramas, teacher tracking, student panoramas, student close-ups, director footage, and courseware images. The classroom content acquisition and analysis module is equipped with a hardware access unit and a storage unit. The hardware access unit adopts an architecture combining fixed and movable cameras to achieve low-latency live streaming of local video streams. The storage unit supports dual modes of local storage and synchronous storage on the cloud platform for recorded content.

[0010] Furthermore, the classroom content acquisition and analysis module also includes a speech-to-text unit, a real-time subtitle generation unit, and a video processing unit; the speech-to-text unit transcribes classroom lectures into editable verbatim transcripts based on speech recognition technology; the real-time subtitle generation unit synchronously adds the transcribed text to the recorded / live video and supports online subtitle editing; the video processing unit supports knowledge point annotation, knowledge point video slicing, and intelligent annotation of the recorded classroom video, the intelligent annotation including text descriptions and arrow indicators.

[0011] Furthermore, the intelligent lesson plan generation module includes a verbatim script preprocessing unit, a key teaching node detection unit, a paragraph clustering and merging unit, a structured tagging unit, and a template adaptation and export unit. The verbatim script preprocessing unit cleanses the verbatim script, removes redundant content, and breaks it down into sentence-level units. The key teaching node detection unit identifies teaching segment boundaries through explicit transition word recognition and implicit semantic analysis, and classifies teaching behaviors. The paragraph clustering and merging unit merges short paragraphs of the same type of teaching behavior and verifies the rationality of the paragraphs through logical coherence checks. The structured tagging unit adds teaching attribute tags to each paragraph. The template adaptation and export unit automatically fills the processed paragraphs according to a preset lesson plan template, generates a standardized lesson plan, and supports exporting it in an editable format.

[0012] Furthermore, the explicit transition word identification includes detecting sequence words and topic switching words to determine the paragraph start point; the implicit semantic analysis uses a pre-trained model to calculate sentence semantic similarity to identify topic change points; the teaching behaviors include explanatory, interactive, and activity-based; and the teaching attribute tags include paragraph type, knowledge point association, and difficulty level.

[0013] Furthermore, the knowledge graph construction module constructs a three-level knowledge point system, which includes first-level subject modules, second-level unit themes, and third-level specific ability points. Each level has a standardized range of quantities, description requirements, and correlation dimensions. The knowledge graph construction module is also equipped with a knowledge point coordinate system, which constructs the three-dimensional coordinates of knowledge points through horizontal logical position, vertical difficulty level, and three-dimensional application depth to achieve knowledge point positioning.

[0014] Furthermore, the personalized recommendation module includes a learning behavior analysis unit, a context-aware unit, and a learning path generation unit; the learning behavior analysis unit collects and analyzes students' learning behavior data to determine students' mastery of knowledge points; the context-aware unit adjusts resource recommendation strategies by combining temporal context and environmental context, where the temporal context includes pre-class preparation and post-class review, and the environmental context includes device type and learning time period; the learning path generation unit calculates the optimal learning path in real time through reinforcement learning to achieve dynamic reconstruction of the learning path.

[0015] This invention also provides a teaching resource processing method for digital teachers, applied to the aforementioned intelligent computing service platform for digital teachers, comprising the following steps: Structured management of teaching resources: Upload various types of teaching resources to the resource library management module, complete automatic knowledge point annotation and cross-library association of resources, and dynamically adjust the node association relationship of the knowledge graph based on student learning behavior data; Intelligent acquisition of classroom content: Real-time acquisition and live streaming of multi-view video streams are achieved through multiple cameras, and classroom speech transcription and real-time video subtitle generation are completed simultaneously. The recorded content adopts a dual storage mode of local and cloud platform. Automatic generation of structured lesson plans: The text is preprocessed and cleaned from the speech-to-text transcript. The teaching content is segmented by detecting key teaching nodes. After the segments are clustered and merged and labeled with structured tags, standardized lesson plans are generated according to preset templates. Knowledge point video slicing and annotation: Mark knowledge points in classroom recording videos, generate knowledge point video slices and add intelligent annotations to enable quick location of knowledge points; Personalized learning resource recommendation: Collect and analyze students' learning behavior data, combine the relationships in the knowledge graph to determine students' mastery of knowledge points, generate personalized learning paths based on contextual information, and push suitable teaching resources.

[0016] According to specific embodiments provided by the present invention, the intelligent computing service platform and teaching resource processing method for digital teachers provided by the present invention disclose the following technical effects: (1) Intelligent and structured resource management: Construct a multi-dimensional teaching resource database system, realize automatic labeling and cross-database association of resources through intelligent tagging and knowledge graphs, and at the same time, knowledge graphs support dynamic evolution driven by learning behavior, which solves the problems of scattered resources and inefficient management in traditional education platforms and improves the utilization rate of teaching resources; (2) Integrated classroom content collection and analysis: It supports real-time collection, live streaming and storage of multiple screens, realizes accurate speech transcription, real-time subtitle generation and knowledge point video slicing, and combines intelligent annotation technology to enable students to quickly locate key knowledge points, solving the problems of incomplete classroom content collection and low efficiency of knowledge point review, and is suitable for various scenarios such as traditional classroom, practical teaching and remote teaching. (3) Automated and standardized lesson plan generation: Based on the verbatim transcript of classroom speech transcription, through the joint modeling of semantic analysis, rule engine and knowledge graph, the automatic conversion from unstructured text to structured lesson plan is realized. It supports template adaptation and editable export, which greatly reduces the preparation time of teachers and improves the efficiency and standardization of lesson plan generation. (4) Refined and dynamic knowledge point modeling: Construct a three-level knowledge point system and a three-dimensional knowledge point coordinate system to achieve precise hierarchical modeling and positioning of knowledge points. At the same time, dynamically adjust the knowledge graph association relationship in combination with students' learning behavior to provide precise knowledge point support for personalized recommendations. (5) Personalized and contextualized resource recommendations: By analyzing students' learning behavior data and knowledge mastery, and combining time sequence and environmental context, the learning path is dynamically reconstructed and personalized resources are pushed, which solves the problem of traditional single recommendation and improves students' learning effect; (6) Strong system scalability and adaptability: The platform adopts a modular structure design, with each functional module being independent and interconnected, making it easy to expand new functional modules. It also supports local and cloud platform collaboration, multi-hardware access, and multi-format resource compatibility, and can adapt to the teaching needs of all stages and subjects in basic education, higher education, and vocational training. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is an overall architecture diagram of the intelligent computing service platform for digital teachers according to the present invention; Figure 2 This is a schematic diagram of the three-level knowledge point system of this invention; Figure 3 This is an overall flowchart of the teaching resource processing method of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The core of this invention is to provide an intelligent computing service platform and teaching resource processing method for digital teachers, which integrates functions such as intelligent resource management, classroom content collection and analysis, intelligent lesson plan generation, knowledge point modeling and personalized recommendation. Through artificial intelligence, knowledge graph and big data analysis technologies, it solves many pain points in educational informatization.

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] Example 1 like Figure 1 As shown, the intelligent computing service platform for digital teachers provided by this invention includes a resource library management module, a classroom content collection and analysis module, an intelligent lesson plan generation module, a knowledge graph construction module, and a personalized recommendation module. These modules work together to provide intelligent support for the entire teaching process.

[0023] 1. Resource Library Management Module The intelligent tagging unit automatically labels uploaded resources with knowledge points, such as automatically labeling math courseware with knowledge points such as "Pythagorean theorem" and "linear function"; the resource association unit realizes cross-database association based on knowledge graph, such as associating course resources for "linear function" with corresponding lesson plans and exercise resources.

[0024] 1.1 Resource Repository Structure This module constructs a multi-dimensional resource system, including a course resource library, a courseware resource library, a lesson plan library, an equipment and experiment resource model library, and a job knowledge base. Course Resource Library: Stores course content, textbooks, reference books, exercise sets, etc. for various subjects, and supports resource updates and version management.

[0025] Courseware Resource Library: Stores various multimedia teaching materials such as PPTs, videos, and audios, supporting multi-format uploading and categorized management.

[0026] Lesson plan database: Stores teaching designs, teaching methods, evaluation criteria, etc. for each subject.

[0027] Equipment and Experiment Resource Model Library: Constructs a virtual teaching environment to assist in theoretical teaching and practical training.

[0028] Job Knowledge Base: Integrates core knowledge points from various disciplines to support intelligent question answering.

[0029] 1.2 Feature Highlights Intelligent resource classification and management: It adopts artificial intelligence for resource tagging management, and supports fast search, resource association and version management.

[0030] Automatic knowledge point annotation: When uploading resources, the system automatically identifies and annotates knowledge points, constructs a knowledge graph, and facilitates association and recommendation.

[0031] Sharing and Collaboration: Supports resource sharing among faculty members and the publication and management of public resource repositories.

[0032] Dynamic knowledge graph relationship adjustment: Learning behavior drives graph evolution: Based on behavioral data such as students' answer accuracy, number of times knowledge points are repeatedly learned, and time spent on questions, the edge weights of nodes (knowledge points) in the knowledge graph are dynamically adjusted.

[0033] Weaken irrelevant paths: For knowledge points that have been mastered (e.g., accuracy rate > 90%), reduce the recommended priority of their extended paths to avoid repetitive learning.

[0034] Real-time personalized learning path generation: Dynamic path reconstruction: Traditional systems rely on fixed main paths, while this solution uses reinforcement learning to calculate the optimal path in real time.

[0035] Context-aware recommendation: Dynamically adjust recommended content by combining temporal context (such as pre-class preparation and post-class review) and environmental context (such as device type and study time).

[0036] 2. Classroom Content Acquisition and Analysis Module This module supports real-time acquisition, preview, and layout customization of six video streams: instructor panorama, instructor tracking, student panorama, student close-up, director, and courseware. Users can manually adjust the screen layout through the template library and also support brand customization functions such as adding logos and titles. Hardware access uses a combination of fixed and movable cameras, adapted for practical teaching (such as helicopter equipment support practice) demonstration footage acquisition. Optimized network transmission protocols enable low-latency live streaming of local video streams, and recorded content is simultaneously stored locally and on the cloud platform to ensure data security.

[0037] The speech-to-text unit, based on industry-leading speech recognition algorithms, accurately transcribes classroom lectures into editable word-for-word transcripts with an accuracy rate of over 95%. The real-time subtitle generation unit adds the transcribed text to the video simultaneously, supporting online editing and multilingual translation. The video processing unit can annotate recorded videos with knowledge points, generate video segments of knowledge points, and add intelligent annotations such as text descriptions and arrow indicators. For example, it can annotate key content such as "equipment disassembly step 1" in practical videos.

[0038] 3. Intelligent Lesson Plan Generation Module The intelligent lesson plan generation module includes a text-to-text preprocessing unit, a key teaching node detection unit, a paragraph clustering and merging unit, a structured tagging unit, and a template adaptation and export unit. The text-to-text preprocessing unit cleanses the text, removes redundant content, and breaks it down into sentence-level units. The key teaching node detection unit identifies teaching segment boundaries and categorizes teaching behaviors by recognizing explicit transition words (including detecting sequence words and topic switching words to determine paragraph start points) and implicit semantic analysis (using a pre-trained model to calculate sentence semantic similarity to identify topic abrupt change points). The paragraph clustering and merging unit merges short paragraphs of the same type of teaching behavior (including explanatory, interactive, and activity-based types) and verifies the rationality of the paragraphs through logical coherence checks. The structured tagging unit adds teaching attribute tags (including paragraph type, knowledge point association, and difficulty level) to each paragraph. The template adaptation and export unit automatically fills the processed paragraphs according to a preset lesson plan template, generates a standardized lesson plan, and supports exporting it in an editable format.

[0039] The intelligent lesson plan generation module enables intelligent lesson plan generation and export, automatically segmenting the text into paragraphs to generate structured lesson plans. Lesson plans can be exported as Word documents using different templates for teachers to prepare lessons and use in teaching.

[0040] The specific process for generating and exporting intelligent lesson plans is as follows: (1) Preprocessing and text cleaning Remove redundant content: Delete colloquial pauses (such as "um" and "ah"), repetitive corrections (such as "then and then"), and irrelevant interjections.

[0041] Basic segmentation: The word-for-word manuscript is divided into sentence-level units according to punctuation marks such as periods and question marks, which facilitates subsequent analysis.

[0042] (2) Testing of key teaching nodes Segment boundaries are identified using the following methods: Explicit transition word identification: Detect sequential words such as "first," "next," "finally," and "in conclusion," or topic-switching words such as "now let's move on" and "let's see," and use them as the starting points of paragraphs.

[0043] (3) Implicit semantic analysis Topic shift detection: Sentence semantic similarity is calculated using pre-trained models such as BERT to identify topic shift points (such as a shift from "concept definition" to "example demonstration").

[0044] (4) Classification of teaching behaviors Explanatory paragraphs: These include areas dense with technical terms (such as "definition of trigonometric functions"), formula derivations, or concept explanations.

[0045] Interactive paragraphs: These include questions (such as "Who can answer this question?"), discussion instructions (such as "Group discussion for 3 minutes"), or student responses.

[0046] Activity-based paragraphs: These include instructions such as experiments, exercises, and games (e.g., "Open the textbook to page 12" or "Complete exercise 3 after class").

[0047] (5) Segment clustering and merging Short paragraph merging: If multiple consecutive short sentences belong to the same teaching behavior (such as explaining a definition in succession), they are merged into a complete paragraph.

[0048] (6) Logical coherence check Verify the rationality of a paragraph by using logical connectors such as cause and effect, chronological order, etc. (e.g., "therefore" or "next").

[0049] Fragmented content (such as isolated questions or blackboard writing actions) can be added to adjacent paragraphs based on context.

[0050] (7) Structured tagging Add teaching attribute tags to each paragraph: Paragraph types: introduction, explanation, question, exercise, summary, etc.

[0051] Knowledge point association: Mapping to subject knowledge points (such as "Pythagorean theorem") through keyword matching or knowledge graphs.

[0052] Difficulty level: The difficulty is marked based on the complexity of terminology and sentence structure (such as the density of long and difficult sentences).

[0053] (8) Manual calibration and template adaptation Teacher intervention interface: Allows manual adjustment of segmentation results (such as splitting / merging paragraphs, modifying tags).

[0054] Template mapping: Automatically fills the segmented results according to the preset lesson plan template (such as "teaching objectives → teaching process → homework assignment") to generate standardized lesson plans.

[0055] Through joint modeling using semantic analysis, rule engines, and knowledge graphs, the system can automatically convert unstructured verbatim manuscripts into structured lesson plans, balancing efficiency and accuracy while retaining the flexibility of manual calibration, ultimately generating high-quality lesson plans that conform to teaching logic.

[0056] (9) Production of independent video resources It supports instructors uploading course videos independently, automatically generating intro and outro videos to enhance the professionalism of the videos. It also automatically segments and marks key knowledge points in the videos for easy student review.

[0057] (10) Synchronous jump and content association By clicking on text or video content, users can jump to the corresponding teaching segment in real time.

[0058] 4. Knowledge Graph Construction Module The knowledge graph construction module builds a three-level knowledge point system, such as Figure 2 As shown, the three-level knowledge point system includes first-level subject modules, second-level unit themes, and third-level specific ability points. Each level has standardized quantity ranges, description requirements, and correlation dimensions. The knowledge graph construction module is also equipped with a knowledge point coordinate system, which constructs the three-dimensional coordinates of knowledge points through horizontal logical position, vertical difficulty level, and depth of three-dimensional application, thereby realizing the location of knowledge points. The specific specifications of the three-level knowledge point system are shown in Table 1.

[0059] Table 1. Attribute Specifications for Each Level of the Three-Level Knowledge Point System

[0060] Simultaneously construct a three-dimensional coordinate system for knowledge points:

[0061] in, L x For horizontal logical positions, D pre The number of prerequisite knowledge points. L y Difficulty levels are arranged vertically. S diff Difficulty level (1-5), L z For 3D application depth, C app To address the number of application scenarios, this enables precise location and association of knowledge points.

[0062] 5. Personalized Recommendation Module This module collects behavioral data such as students' answer accuracy, learning time for knowledge points, and number of repeated learning sessions to analyze students' mastery of knowledge points. It adjusts the recommendation strategy by combining temporal context (basic knowledge points are pushed for pre-class preview, and extended knowledge points are pushed for post-class review) and environmental context (short video clips are pushed for mobile devices, and complete course resources are pushed for computer devices). Through reinforcement learning, it calculates the optimal learning path in real time and realizes the dynamic reconstruction of the path. For example, for students who have not mastered the "properties of linear function graphs", it prioritizes pushing explanation videos and exercise resources for this knowledge point, and associates them with review resources for the prerequisite knowledge point "variables and functions".

[0063] Specifically, the personalized recommendation module includes a learning behavior analysis unit, a context awareness unit, and a learning path generation unit. The learning behavior analysis unit collects and analyzes students' learning behavior data to determine their mastery of knowledge points. The context awareness unit adjusts resource recommendation strategies by combining temporal context and environmental context. The temporal context includes pre-class preparation and post-class review, and the environmental context includes device type and learning time period. The learning path generation unit calculates the optimal learning path in real time through reinforcement learning to achieve dynamic reconstruction of the learning path.

[0064] The platform provided by this invention can achieve: Real-time speech-to-text transcription and synchronized subtitle generation: Based on speech recognition and natural language processing technologies, it achieves high-precision speech-to-text transcription and synchronized subtitle generation.

[0065] Intelligent lesson plan generation: By automatically extracting the core content of the verbatim transcript and combining it with lesson plan templates, a complete lesson plan is automatically generated, reducing teachers' preparation time.

[0066] Knowledge point graph construction and association: The knowledge points in the teaching resources are structured to form a visual knowledge graph, which makes it easier for students to grasp the internal connections between the knowledge points.

[0067] Intelligent resource recommendation: Based on students' learning behavior, mastery of knowledge points, and learning preferences, personalized learning resources are recommended in real time to improve learning outcomes.

[0068] Automated resource structuring: Utilizing OCR recognition and AI processing technologies, uploaded teaching resources are automatically structurated, supporting rapid searching and matching.

[0069] The platform and method of this invention can be adapted to various application scenarios such as basic education, higher education, and vocational training: primary and secondary school teachers can use the platform to prepare lessons, generate lesson plans, and record classroom sessions; university teachers can manage teaching resources, analyze student learning data, and provide personalized teaching; vocational education institutions can use it to record practical courses, provide employee skills training, and conduct precise review of knowledge points.

[0070] In summary, this invention, through a modular and integrated intelligent computing service platform, realizes intelligent management of teaching resources, intelligent collection and analysis of classroom content, automated generation of lesson plans, and personalized recommendation of learning resources. This significantly improves the teaching efficiency of digital teachers, provides precise learning support for students, and promotes the development of educational informatization towards intelligence and personalization.

[0071] Example 2 like Figure 3 As shown, the present invention also provides a teaching resource processing method for digital teachers, applied to the aforementioned intelligent computing service platform for digital teachers, comprising the following steps: 1) Structured management of teaching resources: Upload various types of teaching resources to the resource library management module to complete automatic knowledge point annotation and cross-library association of resources, and dynamically adjust the node association relationship of the knowledge graph based on student learning behavior data; 2) Intelligent acquisition of classroom content: Real-time acquisition and live streaming of multi-view video streams are achieved through multiple cameras, and classroom speech transcription and real-time video subtitle generation are completed simultaneously. The recorded content adopts a dual storage mode of local and cloud platform. 3) Automatic generation of structured lesson plans: The text is preprocessed and cleaned from the transcribed speech, and the teaching content is segmented by detecting key teaching nodes. After the segments are clustered and merged and labeled with structured tags, standardized lesson plans are generated according to preset templates. 4) Knowledge Point Video Slicing and Annotation: Annotate the knowledge points in the recorded classroom videos, generate knowledge point video slices and add intelligent annotations to enable quick location of knowledge points; 5) Personalized learning resource recommendation: Collect and analyze students' learning behavior data, combine the relationships in the knowledge graph to determine students' mastery of knowledge points, generate personalized learning paths based on contextual information, and push suitable teaching resources.

[0072] The key teaching node detection mentioned in step 3) includes explicit transition word identification and implicit semantic analysis. The explicit transition word identification determines the paragraph start point by detecting sequence words and topic switching words. The implicit semantic analysis calculates sentence semantic similarity through a pre-trained model to identify topic change points and classifies explanatory, interactive, and activity-based teaching behaviors.

[0073] The personalized learning resource recommendation mentioned in step 5) also includes: recommending resources that weaken the extension path of knowledge points that students have mastered, and recommending resources that strengthen the prerequisite and application relationships of knowledge points that students have not mastered; at the same time, the form and difficulty of the recommended content are dynamically adjusted according to the temporal and environmental context.

[0074] In summary, the intelligent computing service platform and teaching resource processing method for digital teachers provided by this invention achieve the following technical improvements: (1) Performance advantages Efficient resource management: Enables automatic classification, association, and recommendation of teaching resources, improving instructors' resource management efficiency.

[0075] Real-time transcription and synchronized subtitles: Provides real-time transcription and synchronized subtitles for all grade levels and subjects, and supports multilingual translation.

[0076] (2) High reliability Modular design: The system adopts a modular architecture, which can be quickly expanded as needs change, ensuring system stability and reliability.

[0077] Data security and privacy protection: Strict access control is implemented for uploaded teaching resources to ensure data security and user privacy.

[0078] (3) Good user experience Intelligent operation: Provides simple and easy-to-use micro-lesson recording tools and intelligent lesson plan generation functions to enhance teachers' teaching experience.

[0079] Personalized learning: Based on students' mastery of knowledge points, learning resources and classroom recordings are intelligently pushed to improve students' learning outcomes.

[0080] (4) Application scenarios Basic Education: Primary and secondary school teachers use the platform for lesson preparation, lesson plan generation, and resource management.

[0081] Higher Education: University teachers use the platform to manage teaching resources and analyze student learning data to achieve personalized teaching.

[0082] Vocational training: Vocational education institutions use the platform to provide online course management, learning resource recommendations, and student skills assessments.

[0083] The intelligent computing service platform provided by this invention solves the challenges of resource management and personalized teaching in the current process of educational informatization through functions such as resource library management, intelligent lesson plan generation, knowledge graph construction, and personalized resource recommendation. The platform improves teaching efficiency, resource management capabilities, and student learning outcomes through technological innovation, providing strong technical support for the digital transformation of the education industry.

[0084] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. An intelligent computing service platform for digital teachers, characterized in that, It includes a resource library management module, a classroom content collection and analysis module, an intelligent lesson plan generation module, a knowledge graph construction module, and a personalized recommendation module; The resource library management module is used for structured storage, intelligent classification and dynamic management of various types of teaching resources, to build a multi-dimensional resource library system and realize cross-library association of resources; The classroom content acquisition and analysis module is used to realize the real-time acquisition, live streaming and storage of multi-screen classroom content, and to complete the accurate transcription of classroom audio content and the real-time generation of video subtitles; The intelligent lesson plan generation module is used to automatically generate structured and standardized lesson plans based on the verbatim transcript of classroom speech transcription, and supports the editing and export of lesson plans; The knowledge graph construction module is used to automatically label and hierarchically model knowledge points in teaching resources, construct a dynamically evolving knowledge graph, and realize the association and tracing of knowledge points. The personalized recommendation module is used to generate personalized learning paths in real time based on students' learning behavior data and knowledge mastery, combined with the relationships in the knowledge graph, and push suitable teaching resources.

2. The intelligent computing service platform for digital teachers according to claim 1, characterized in that, The resource library management module includes a course resource library, a courseware resource library, a lesson plan library, an equipment and experimental resource model library, and a job knowledge library. Each resource library supports uploading resources in multiple formats, updating resources, version management, and cross-library sharing. The resource library management module is also equipped with an intelligent tagging unit and a resource association unit. The intelligent tagging unit uses artificial intelligence algorithms to automatically tag knowledge points of uploaded teaching resources, and the resource association unit uses knowledge graphs to achieve the association and matching of teaching resources in different resource libraries.

3. The intelligent computing service platform for digital teachers according to claim 2, characterized in that, The knowledge graph of the resource library management module supports dynamic evolution driven by learning behavior. Specifically, it adjusts the edge weights between knowledge point nodes in the knowledge graph based on at least one learning behavior data, such as the student's answer accuracy, the number of times the knowledge point is repeatedly studied, and the time spent on the question. For knowledge points that students have mastered, it lowers the recommendation priority of their extended paths and weakens irrelevant learning paths.

4. The intelligent computing service platform for digital teachers according to claim 1, characterized in that, The classroom content acquisition and analysis module supports real-time acquisition, preview, and layout customization of at least six video streams. The video streams include panoramic views of the instructor, instructor tracking, panoramic views of students, close-ups of students, director footage, and courseware images. The classroom content acquisition and analysis module is equipped with a hardware access unit and a storage unit. The hardware access unit adopts an architecture combining fixed and movable cameras to achieve low-latency live streaming of local video streams. The storage unit supports dual modes of local storage and synchronous storage on the cloud platform for recorded content.

5. The intelligent computing service platform for digital teachers according to claim 4, characterized in that, The classroom content acquisition and analysis module also includes a speech-to-text unit, a real-time subtitle generation unit, and a video processing unit; the speech-to-text unit uses speech recognition technology to transcribe classroom lectures into editable verbatim transcripts. The real-time subtitle generation unit synchronously adds the transcribed text to the recorded / live video and supports online subtitle editing; The video processing unit supports marking knowledge points, slicing knowledge point videos, and intelligent annotations on classroom recordings. The intelligent annotations include text descriptions and arrow indicators.

6. The intelligent computing service platform for digital teachers according to claim 1, characterized in that, The intelligent lesson plan generation module includes a verbatim script preprocessing unit, a key teaching node detection unit, a paragraph clustering and merging unit, a structured tagging unit, and a template adaptation and export unit. The verbatim script preprocessing unit cleans the verbatim script, removes redundant content, and breaks it down into sentence-level units. The key teaching node detection unit identifies teaching segment boundaries through explicit transition word recognition and implicit semantic analysis, and classifies teaching behaviors. The paragraph clustering and merging unit merges short paragraphs of the same type of teaching behavior and verifies the rationality of the paragraphs through logical coherence checks. The structured tagging unit adds teaching attribute tags to each paragraph; the template adaptation and export unit automatically fills the processed paragraphs according to the preset lesson plan template, generates standardized lesson plans, and supports exporting them in an editable format.

7. The intelligent computing service platform for digital teachers according to claim 6, characterized in that, The explicit transition word identification includes detecting sequence words and topic switching words to determine the paragraph start point; the implicit semantic analysis uses a pre-trained model to calculate sentence semantic similarity to identify topic change points; the teaching behaviors include explanatory, interactive, and activity-based; and the teaching attribute tags include paragraph type, knowledge point association, and difficulty level.

8. The intelligent computing service platform for digital teachers according to claim 1, characterized in that, The knowledge graph construction module constructs a three-level knowledge point system, which includes a first-level subject module, a second-level unit theme, and a third-level specific ability point. Each level has a standardized quantity range, description requirements, and correlation dimensions. The knowledge graph construction module is also equipped with a knowledge point coordinate system, which constructs the three-dimensional coordinates of knowledge points through horizontal logical position, vertical difficulty level, and three-dimensional application depth to achieve knowledge point positioning.

9. The intelligent computing service platform for digital teachers according to claim 1, characterized in that, The personalized recommendation module includes a learning behavior analysis unit, a context awareness unit, and a learning path generation unit. The learning behavior analysis unit collects and analyzes students' learning behavior data to determine their mastery of knowledge points. The context awareness unit adjusts the resource recommendation strategy by combining temporal context and environmental context. The temporal context includes pre-class preparation and post-class review, and the environmental context includes device type and learning time period. The learning path generation unit calculates the optimal learning path in real time through reinforcement learning, thereby realizing the dynamic reconstruction of the learning path.

10. A method for processing teaching resources for digital teachers, applied to the intelligent computing service platform for digital teachers as described in any one of claims 1-9, characterized in that, Includes the following steps: Structured management of teaching resources: Upload various types of teaching resources to the resource library management module, complete automatic knowledge point annotation and cross-library association of resources, and dynamically adjust the node association relationship of the knowledge graph based on student learning behavior data; Intelligent acquisition of classroom content: Real-time acquisition and live streaming of multi-view video streams are achieved through multiple cameras, and classroom speech transcription and real-time video subtitle generation are completed simultaneously. The recorded content adopts a dual storage mode of local and cloud platform. Automatic generation of structured lesson plans: The text is preprocessed and cleaned from the speech-to-text transcript. The teaching content is segmented by detecting key teaching nodes. After the segments are clustered and merged and labeled with structured tags, standardized lesson plans are generated according to preset templates. Knowledge point video slicing and annotation: Mark knowledge points in classroom recording videos, generate knowledge point video slices and add intelligent annotations to enable quick location of knowledge points; Personalized learning resource recommendation: Collect and analyze students' learning behavior data, combine the relationships in the knowledge graph to determine students' mastery of knowledge points, generate personalized learning paths based on contextual information, and push suitable teaching resources.

Images