A research and use integrated teaching system and method based on human and intelligent engine collaborative design

Abstract

The application discloses a research and use integrated teaching system and method based on human and intelligent engine collaborative design, relates to the intelligent teaching technical field, and constructs the research and use integrated teaching system based on human and intelligent engine collaborative design through understanding education semantics, collaborative AI capability and data asset integration. The system is composed of a closed loop architecture of a teaching application development environment module, an intelligent service platform and a teaching application execution environment module. When an application needs to be created, the teaching application design intention and the teaching context are deeply fused and semantically understood through the system, a structured application requirement description is obtained, and application arrangement is performed, full-link automation and intelligentization from teaching creativity to teaching practice are realized, and the teaching application can be efficiently and accurately created as dynamic, personalized, interactive, specific to specific teaching targets, adaptive to different student cognitive levels and integrated with school local characteristics.

Description

Technical Field

[0001] This application relates to the field of intelligent teaching technology, and more specifically, to a research and application integrated teaching system and method based on the collaborative design of humans and intelligent engines. Background Technology

[0002] With the deepening development of educational informatization and the implementation of the core competency education concept, traditional and standardized digital educational resources can no longer meet the needs of personalized, inquiry-based, and interdisciplinary teaching. Modern teaching practice urgently requires the rapid generation of interactive teaching applications that align with specific teaching objectives, adapt to different students' cognitive levels, and incorporate local school characteristics.

[0003] Existing methods for creating interactive teaching applications generally rely on two paths. The first path involves customization by professional software development teams, which is characterized by long development cycles, high costs, and high technical barriers, and educators cannot directly participate. The second path involves educators using general-purpose multimedia courseware tools (such as PPT) or simple online activity templates to create them. However, such tools have weak interactivity and low intelligence levels, making it difficult to support complex and dynamic subject knowledge exploration and personalized learning path guidance.

[0004] Interactive teaching applications created through the two paths described above are difficult to seamlessly integrate with and dynamically utilize the school's existing data assets (such as student profiles, school-based resource databases, subject knowledge point systems, etc.), and cannot achieve large-scale teaching objectives, personalized content, interactive application teaching that caters to different students' cognitive levels and integrates the school's local characteristics.

[0005] Therefore, how to efficiently and accurately create personalized, interactive teaching applications that align with specific teaching objectives, adapt to different students' cognitive levels, and incorporate local school characteristics is a problem that this application urgently needs to solve. Summary of the Invention

[0006] In view of this, this application discloses a research-application integrated teaching system and method based on human-intelligent engine collaborative design. It aims to construct an intelligent research-application integrated teaching system that can understand educational semantics, collaborate with AI capabilities, and integrate data assets. When applications need to be created, the research-application integrated teaching system can realize the automation and intelligence of the entire chain from teaching ideas to teaching practice. It can efficiently and accurately create personalized, interactive, interactive teaching applications that are tailored to specific teaching objectives, adapt to different students' cognitive levels, and incorporate the local characteristics of the school.

[0007] To achieve the above objectives, the disclosed technical solution is as follows:

[0008] The first aspect of this application discloses a research and application integrated teaching system based on the collaborative design of humans and intelligent engines. The system includes a teaching application development environment module, an intelligent service platform, and a teaching application execution environment module. The intelligent service platform includes at least a message integration engine and an application orchestration engine.

[0009] The teaching application development environment module is used to obtain teacher configuration information through a teaching-specific AI component; wherein, the teaching-specific AI component is obtained by domain modeling; the teacher configuration information includes at least the teaching application design intent; the message integration engine is used to obtain the teaching context and perform deep fusion and semantic understanding of the teaching context and the teaching application design intent to obtain a structured application requirement description;

[0010] The application orchestration engine is used to orchestrate the application requirements description to obtain an application assembly blueprint.

[0011] The teaching application execution environment module is used to call the AI ​​service through the unified scheduling entry of internal and external services to load the application assembly blueprint and deploy it to the teaching application execution environment module for application execution and closed-loop optimization.

[0012] The second aspect of this application discloses a research-application integrated teaching method based on human-intelligent engine collaborative design. This method is applied to the research-application integrated teaching system based on human-intelligent engine collaborative design described in any of the first aspects above. The system includes a teaching application development environment module, an intelligent service platform, and a teaching application execution environment module. The intelligent service platform includes at least a message integration engine and an application orchestration engine. The method includes:

[0013] The teaching application development environment module calls the teaching-specific AI component to obtain teacher configuration information; wherein, the teaching-specific AI component is obtained by domain modeling; the teacher configuration information includes at least the teaching application design intent; the message integration engine is used to obtain the teaching context, and deeply integrate and semantically understand the teaching context and the teaching application design intent to obtain a structured application requirement description;

[0014] The application requirements description is orchestrated using the application orchestration engine to obtain an application assembly blueprint.

[0015] The AI ​​service is invoked through the unified scheduling entry point of the teaching application execution environment module and internal and external services to load the application assembly blueprint and deploy it to the teaching application execution environment module for application execution and closed-loop optimization.

[0016] As can be seen from the above technical solution, this application discloses a research-application integrated teaching system and method based on human-intelligence engine collaborative design. This solution constructs a research-application integrated teaching system based on human-intelligence engine collaborative design that can understand educational semantics, collaborate with AI capabilities, and integrate data assets. The system consists of a closed-loop architecture comprising a teaching application development environment, an intelligent service platform, and a teaching application execution environment. When an application needs to be created, the research-application integrated teaching system based on human-intelligence engine collaborative design deeply integrates and semantically understands the design intent of the teaching application and the teaching context, obtaining a structured application requirement description. This description is then automatically arranged into applications, achieving full-link automation and intelligence from teaching ideas to teaching practice. This enables educators to efficiently and accurately create dynamic, personalized, interactive, and school-specific interactive teaching applications that align with specific teaching objectives, adapt to different students' cognitive levels, and incorporate local school characteristics. Attached Figure Description

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

[0018] Figure 1 This is a schematic diagram of the architecture of a research and application integrated teaching system based on collaborative design between humans and intelligent engines, as disclosed in an embodiment of this application.

[0019] Figure 2 This is a schematic diagram of the core architecture and workflow of the research and application integrated teaching system based on human-intelligent engine collaborative design disclosed in an embodiment of this application;

[0020] Figure 3 This is a schematic diagram of the deployment architecture of the research and application integrated teaching system based on the collaborative design of humans and intelligent engines disclosed in the embodiments of this application;

[0021] Figure 4 This is a schematic diagram of the workflow of a research-application integrated teaching system based on collaborative design between humans and intelligent engines, as disclosed in an embodiment of this application.

[0022] Figure 5 This is an example diagram of the teaching application development environment disclosed in the embodiments of this application;

[0023] Figure 6 This is an example preview page diagram disclosed in an embodiment of this application;

[0024] Figure 7 This is another example preview page diagram disclosed in an embodiment of this application;

[0025] Figure 8 This is a detailed flowchart illustrating the application design and application generation process disclosed in the embodiments of this application;

[0026] Figure 9 This is a flowchart illustrating a research-application integrated teaching method based on collaborative design between humans and intelligent engines, as disclosed in an embodiment of this application. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] In this application, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0029] refer to Figure 1 As shown in the embodiment of this application, a research and application integrated teaching system based on human-intelligent engine collaborative design is disclosed. The system includes a teaching application development environment module, an intelligent service platform, and a teaching application execution environment module; wherein, the intelligent service platform includes at least a message integration engine and an application orchestration engine.

[0030] The workflow of the research-application integrated teaching system, specifically designed based on the collaborative design of humans and intelligent engines, is as follows:

[0031] The teaching application development environment module receives user configurations based on teaching-specific AI components and triggers subsequent processes. Specifically, the teaching application development environment module obtains teacher configuration information through teaching-specific AI components; these components are derived from domain modeling; and the teacher configuration information includes at least the teaching application design intent.

[0032] The intelligent service platform serves as the brain of a research-application integrated teaching system designed collaboratively by humans and an intelligent engine. It includes a message integration engine, an application orchestration engine, a teaching AI service layer, and a teaching data layer.

[0033] The message integration engine is used to obtain the teaching context and deeply integrate and semantically understand the teaching context and the design intent of the teaching application to obtain a structured description of application requirements. Specifically, the message integration engine is used to call the knowledge point graph engine to establish cross-disciplinary knowledge relationships; call the content generation engine to generate customized content that conforms to the theme and has teaching focus annotations through cross-disciplinary knowledge relationships; and integrate the customized content that conforms to the theme and has teaching focus annotations to obtain a structured description of application requirements.

[0034] The application orchestration engine receives the requirement specifications, coordinates the teaching AI service layer and data resources, and generates an executable application assembly blueprint. Specifically, the application orchestration engine is used to orchestrate the application requirement descriptions to obtain the application assembly blueprint. The application orchestration engine is specifically used to acquire pre-stored digital teaching resources (such as the historical timeline of the Forbidden City, 3D models of dougong brackets, etc.); to perform intelligent scheduling of external resources, acquiring and integrating the required multimedia materials, documents, and map data from authorized and open external data sources; and to generate a structured application assembly blueprint based on the digital teaching resources, multimedia materials, documents, and map data.

[0035] The teaching application execution environment module is used to load and run application assembly blueprints, providing students with a learning experience and recording learning data feedback to the platform, forming a closed loop of "research and application integration". Specifically, the teaching application execution environment module uses a unified scheduling entry point for internal and external services to call AI services to load application assembly blueprints and deploy them to the teaching application execution environment module for application execution and closed-loop optimization.

[0036] It should be noted that an application assembly blueprint refers to the executable definition of a complete application.

[0037] Application Assembly Blueprint: A JSON file that defines "how the application is assembled," including:

[0038] Define the module's type, functional interface, and configuration specifications (such as a timeline component that supports bidirectional sliding and can embed multimedia markers).

[0039] Define the application's view structure, module layout, and navigation logic (e.g., the "Story Reader" module needs to provide the onWordClick event and consume the query service of the "Smart Assistant" module).

[0040] Define data contracts, event protocols, and dependency declarations between modules (e.g., using 'tab navigation' as the main framework, containing three top-level views (Tabs)).

[0041] Define the logical binding relationship with the backend core services (such as the “Learning Analysis Service” subscribing to the output of the “Learning Recorder” and generating reports for the “Dashboard Module” to render).

[0042] The following is an example of applying assembly blueprints:

[0043] {

[0044] "version": "1.0.0",

[0045] "app_metadata": {

[0046] "app_id": "cultural_comparison_app_v1.2",

[0047] "name": "Palace Dialogues: Forbidden City & Versailles",

[0048] "description": "An interactive bilingual application for middleschool students to explore and compare the architectural and cultural aspects of the Forbidden City and the Palace of Versailles, with embedded vocabulary learning.",

[0049] "target_platform": "web",

[0050] "created_at": "2025-09-12T10:00:00Z"

[0051] },

[0052] "modules": {

[0053] "bilingual_comparison_story": {

[0054] "type": "InteractiveBilingualStoryViewer",

[0055] "dependencies": ["global_i18n_service", "smart_assistant_client"],

[0056] "config": {

[0057] "story_title": {

[0058] "en": "A Tale of Two Palaces: Power, Beauty, and Belief",

[0059] "zh": "A Tale of Two Cities: Power, Aesthetics, and Faith"

[0060] },

[0061] "story_content": {

[0062] "en": "[This is the AI-generated English story text, wheretarget words like <highlight data-word="architecture"> architecture< / highlight> , <highlight data-word="heritage"> heritage< / highlight> are annotated.

[0063] "zh": "[This is the AI-generated Chinese story text, corresponding to the English content.]"

[0064] },

[0065] "target_vocabulary": [

[0066] {

[0067] "word": "architecture",

[0068] "lemma": "architecture"

[0069] "pos": "n."

[0070] "difficulty": "B1"

[0071] "explanations": {

[0072] "en": "the art or practice of designing and constructingbuildings.",

[0073] "zh": "Architectural design or style; architecture"

[0074] },

[0075] "example_sentences": [

[0076] {

[0077] "en": "The **architecture** of the Forbidden City reflects the absolute power of the emperor.",

[0078] "zh": "The architectural style of the Forbidden City reflects the absolute power of the emperor."

[0079] }

[0080] ],

[0081] "common_errors": ["Confusion between 'architecture' and 'structure' (the latter focuses more on physical structure)"]

[0082] },

[0083] {

[0084] "word": "heritage"

[0085] "lemma": "heritage"

[0086] "pos": "n."

[0087] "difficulty": "B2"

[0088] "explanations": {

[0089] "en": "features belonging to the culture of a particular society, such as traditions, languages, or buildings, that still exist from the past and have historical importance.",

[0090] "zh": "Heritage (referring to the culture, traditions, architecture, etc., preserved from history)"

[0091] },

[0092] "example_sentences": [

[0093] {

[0094] "en": "Both palaces are part of our world cultural **heritage**.",

[0095] "zh": "Both palaces are part of the World Cultural Heritage."

[0096] }

[0097] ],

[0098] "common_errors": ["Incorrectly used as a countable noun (Correct: our heritage; Incorrect: a heritage)"]

[0099] }

[0100] ],

[0101] "comparison_data": {

[0102] "dimensions": ["Architectural Philosophy", "Garden Layout", "Court Function"],

[0103] "points": [

[0104] {

[0105] "dimension": "Architectural Philosophy",

[0106] "forbidden_city": "Emphasizes symmetry, hierarchy, and cosmological symbolism (eg, the Hall of Supreme Harmony representing the center of the universe).",

[0107] "versailles": "Demonstrates axiality, grandeur, and the celebration of monarchical power and artistic achievement (eg, the Hall of Mirrors)."

[0108] } ]

[0110] },

[0111] "multimedia_assets": {

[0112] "images": [

[0113] {

[0114] "id": "img_fc_plan"

[0115] "url": "https: / / assets.yourschool.com / forbidden_city_plan.jpg",

[0116] "caption_en": "Layout of the Forbidden City, showingaxial symmetry.",

[0117] "caption_zh": "A floor plan of the Forbidden City, showing the symmetrical layout along the central axis."

[0118] },

[0119] {

[0120] "id": "img_vs_garden"

[0121] "url": "https: / / assets.yourschool.com / versailles_garden_aerial.jpg",

[0122] "caption_en": "Aerial view of the Gardens ofVersailles, showingcasing geometric patterns.",

[0123] "caption_zh": "Aerial view of the Versailles Gardens, showcasing its geometric garden layout."

[0124] } ]

[0126] },

[0127] "ui": {

[0128] "default_language": "zh",

[0129] "supported_languages": ["zh", "en"],

[0130] "reading_tools": {

[0131] "text_to_speech": true,

[0132] "line_focus": false

[0133] }

[0134] }

[0135] },

[0136] "annotations": {

[0137] "design_intent": "Through a coherent, narrative approach, architectural contrasts are integrated with cultural context, allowing students to naturally encounter target vocabulary while reading. Clicking on highlighted words triggers in-depth analysis, rather than interrupting reading."

[0138] "Learning Objective": "Students will be able to understand the core differences in the architectural concepts of the two palaces and accurately understand and use the five target vocabulary words, including 'architecture' and 'heritage,' in context."

[0139] }

[0140] },

[0141] "vocabulary_flashcard": {

[0142] "type": "ContextualVocabularyDrill",

[0143] "dependencies": ["bilingual_comparison_story"],

[0144] "config": {

[0145] "activation_trigger": "word_click",

[0146] "display_mode": "modal_overlay",

[0147] "drill_types": ["definition", "example", "common_error"]

[0148] }

[0149] },

[0150] "smart_assistant_interface": {

[0151] "type": "FloatingAssistantWidget",

[0152] "dependencies": ["assistant_service"],

[0153] "config": {

[0154] "assistant_id": "ASSISTANT_XIAOTANG_V2",

[0155] "default_persona": "Cultural Comparison Tutor",

[0156] "invocation_methods": ["click", "voice_keyword"],

[0157] "position": "bottom_right"

[0158] }

[0159] }

[0160] },

[0161] "routing_logic": [

[0162] {

[0163] "id": "rl_001",

[0164] "description":"When a highlighted word is clicked in the story,show its detailed flashcard and notify the assistant for contextual help.",

[0165] "from": "bilingual_comparison_story.word_clicked",

[0166] "to":["vocabulary_flashcard.show","smart_assistant_interface.set_context"],

[0167] "data_mapping":{

[0168] "clicked_word":"event.word",

[0169] "sentence_context":"event.sentence",

[0170] "story_paragraph_id":"event.paragraph_id"

[0171] }

[0172] },

[0173] {

[0174] "id": "rl_002",

[0175] "description": "When the language toggle is switched, reloadthe story content and UI labels in the new language.",

[0176] "from": "global_ui_controller.language_changed",

[0177] "to": "bilingual_comparison_story.update_content_language",

[0178] "data_mapping": {

[0179] "new_language": "event.language"

[0180] }

[0181] },

[0182] {

[0183] "id": "rl_003",

[0184] "description": "If the user asks the assistant about acomparison point, highlight the relevant section in the story viewer.",

[0185] "from": "smart_assistant_interface.user_asked_about_comparison",

[0186] "to": "bilingual_comparison_story.highlight_section",

[0187] "data_mapping": {

[0188] "comparison_dimension": "event.dimension"

[0189] }

[0190] }

[0191] ],

[0192] "data_sources":{

[0193] "internal":{

[0194] "vocabulary_list":"ref: / / vocabulary / junior_high_common_errors",

[0195] "school_history":"ref: / / archive / SCH_ARCH_1958_001"

[0196] },

[0197] "external": {

[0198] "versailles_official_text":"https: / / en.chateauversailles.fr / discover / history"

[0199] }

[0200] },

[0201] "global_config": {

[0202] "internationalization": {

[0203] "default_locale": "zh-CN",

[0204] "supported_locales": ["zh-CN", "en-US"]

[0205] },

[0206] "theme": {

[0207] "primary_color":"#2c3e50",

[0208] "font_family": "'Noto Serif SC',serif"

[0209] }

[0210] }

[0211] }

[0212] Through the above architecture, this solution achieves full-link automation and intelligence from teaching ideas to teaching practice, effectively solving the gap between "educational ideas and technological implementation".

[0213] The research-application integrated teaching system, designed collaboratively by humans and an intelligent engine, is based on a microservices architecture. Its components communicate via RESTful APIs. The specific technical architecture of this system is as follows: Figure 2 As shown.

[0214] Figure 2 The research and application integrated teaching system based on the collaborative design of humans and intelligent engines also includes a teaching application development environment module and a teaching application execution environment module.

[0215] 1. Teaching Application Development Environment Module:

[0216] The teaching application development environment module is positioned as an "innovation workshop" for teachers and students, providing a low-code development experience.

[0217] The core modules of the teaching application development environment module include the application designer, preview debugger, and release manager.

[0218] The application designer provides a graphical interface and supports teachers in designing instructional applications by selecting instruction-specific AI components, configuring properties, and editing workflows.

[0219] The preview debugger is used to preview and debug application effects in real time.

[0220] The release manager is used to release debugged applications with one click and manage versions and permissions.

[0221] Core modules of the instructional application execution environment:

[0222] The teaching application execution environment module includes an application portal, application runner, learning recorder, and intelligent assistant integration.

[0223] The application portal provides a personalized list of applications, learning progress tracking and navigation; it also enables teachers to understand students' learning progress and conduct AI-supported lesson preparation and teaching research activities.

[0224] The application runner's core rendering engine dynamically generates and runs personalized learning content or teacher preparation and research content based on the application assembly blueprint and user profile.

[0225] The learning recorder is used to record students' interactions, answers, and learning outcomes throughout the entire process.

[0226] The smart assistant is seamlessly integrated, and includes multiple wake-up methods and types, supporting real-time guidance via text or voice.

[0227] Intelligent Service Platform:

[0228] The intelligent service platform comprises an intelligent engine layer, an educational AI service layer, and an educational data layer. The intelligent engine layer includes an API gateway. This API gateway serves as a unified scheduling entry point for calling local AI services and generative and basic AI services on the public cloud. The message integration engine, application orchestration engine, and API gateway are all located within the intelligent engine layer. The educational AI service layer and educational data layer provide the engine with AI capabilities and knowledge support.

[0229] Intelligent Engine Layer:

[0230] The intelligent engine layer is positioned as the system's "intelligent brain" and "information hub," connecting the R&D and execution environments.

[0231] Core modules of the intelligent engine layer:

[0232] The message integration engine deeply integrates and semantically understands the user's design intent for teaching applications in the teaching application development environment (such as interdisciplinary fun learning about the Forbidden City) with the rich context of the data hub (such as student profiles, subject knowledge points, and school information), generating a structured application requirement description that is complete and rich in context.

[0233] Application Orchestration Engine: Receives application requirement descriptions from the development environment and automatically generates a complete, technology-neutral application assembly specification called an Application Assembly Blueprint by integrating and scheduling domain-specific AI models (such as content generation and knowledge point reasoning) and multimodal resources. API Gateway: Serves as the unified scheduling entry point for internal and external services, calling both local AI services and securely accessing generative and basic AI services on the public cloud.

[0234] AI service layer for teaching:

[0235] The teaching AI service layer is positioned as follows: it encapsulates a series of intelligent services specifically for educational scenarios.

[0236] The teaching AI service layer includes a natural language processing engine, an information search engine, an intelligent assistant service, a content generation engine, a knowledge graph engine, a learning analysis engine, and a lightweight reasoning service.

[0237] The natural language processing engine is used to understand the teacher's design requirements, parse the natural language description based on the teacher's design requirements and transform it into a structured application configuration, as well as handle the interaction between students and the application;

[0238] The intelligent assistant service is used to provide at least dialogue management, image and voice dispatch support for assistant functions in the execution environment.

[0239] The content generation engine is used to generate an initial content framework, sample materials, interactive designs, and personalized content adaptations for teaching applications based on the teaching themes and objectives set by the teachers.

[0240] Information search engines, knowledge graph engines, and learning analysis engines are used to build knowledge networks, analyze learning data, and recommend learning resources, and are the core of achieving personalization and intelligence.

[0241] The knowledge graph engine provides a semantic framework and reasoning foundation.

[0242] Lightweight inference services are used to handle local computation tasks that meet preset requirements (i.e., high real-time requirements).

[0243] Teaching data layer:

[0244] Positioning of the teaching data layer: the foundation of all data on the platform.

[0245] The core database of the teaching data layer includes at least a teaching-specific AI component library, a core viewpoint library for each subject, a learning behavior record library, an assistant resource library, a dialogue history library, a teaching application definition library, a teaching content template library, a student profile library, and a learning information library;

[0246] The AI ​​component library for teaching is used to store reusable AI teaching function units; it is the foundation for low-code development.

[0247] The core viewpoint database for each discipline is used to store knowledge points for each discipline.

[0248] Both the learning behavior record database and the student profile database are used to record students' personality traits and learning process data.

[0249] Both the assistant database and the conversation history database are used to support personalized interactions with the intelligent assistant.

[0250] The instructional application definition library is used to store application assembly blueprints.

[0251] The teaching content template library is used to store organizational templates for content from various subjects.

[0252] The school information database is used to store school-based information; school-based information includes at least the school icon, school history, and best practices.

[0253] This solution, based on a research-application integrated teaching system designed collaboratively by humans and intelligent engines, adopts a secure public-private hybrid architecture, specifically as follows: Figure 3 As shown. The core design principle of the public-private hybrid architecture is to retain sensitive core educational data on the school's local network while utilizing the AI ​​capabilities of the public cloud on demand. This design ensures data privacy and compliance while enjoying the benefits of cutting-edge AI technology. Specifically:

[0254] 1. Campus local network:

[0255] This layer deploys all the platform's core business logic and sensitive data, isolated from the public network to ensure data sovereignty.

[0256] Client-side: Teachers and students access the server through a browser.

[0257] Server-side: Deployed on the school's internal server, containing the system's core business logic and data.

[0258] 2. Public Cloud:

[0259] This layer, acting as a powerful and scalable AI capability provider, does not store any core educational data. This system integrates various generative AI services provided by the public cloud via APIs, including but not limited to:

[0260] Large language model service: used for teaching content generation and intelligent tutoring dialogue;

[0261] Multimodal content generation services: including text-to-image and text-to-video capabilities;

[0262] Code intelligence service: Supports the development and execution of programming teaching applications.

[0263] In addition, basic AI capabilities (including but not limited to computer vision, speech recognition, and traditional natural language processing) are prioritized for implementation via public cloud APIs to ensure performance and reduce costs.

[0264] Meanwhile, for specific scenarios with extremely high data privacy requirements or extremely low latency requirements, the system supports the local deployment of specific, lightweight basic AI models on the campus server, forming a hybrid AI capability scheduling mode. This allows users to enjoy the powerful capabilities of the cloud while meeting the special compliance and experience requirements of the education scenario.

[0265] The workflow diagram of the research-application integrated teaching system based on human-intelligent engine collaborative design is as follows: Figure 4 As shown. Figure 4 This embodies the integrated research and application process, from instructional application design, generation, and implementation to closed-loop optimization. All participating modules collaborated to achieve intelligent generation and continuous optimization of instructional applications. Figure 4 It includes the application design phase, application generation phase, application execution phase, and closed-loop optimization phase.

[0266] I. Application Design Phase:

[0267] This stage is led by teachers, who guide the transformation from initial ideas to precise design schemes.

[0268] 1. Creating Applications: Teachers create new applications in the teaching application development environment and express their initial teaching intentions by dragging and dropping and configuring teaching-specific AI components.

[0269] 2. Generate Description (Intelligent Design Enhancement): Teachers click the "Generate Description" button to trigger the system's intelligent design assistance:

[0270] The development environment submits teacher configuration information to the intelligent engine layer;

[0271] The message integration engine starts, coordinating and querying contextual information such as subject knowledge points and student profiles in the teaching data layer;

[0272] The system calls the teaching AI service layer to perform semantic understanding and requirements analysis;

[0273] The engine integrates information from multiple sources to generate structured functional and content requirements specifications. This step marks the shift from "human design" to "intelligent collaborative design," representing a deepening and refinement of the design phase.

[0274] Examples of specific functional and content requirements are as follows:

[0275] Dialogue between Eastern and Western Palaces:

[0276] Educational Objectives: By comparing the Forbidden City and the Palace of Versailles, students will understand the differences between East and West in architectural philosophy, garden art, and court culture, and enhance their cultural appreciation and bilingual expression skills.

[0277] Content elements:

[0278] Cultural Comparison Matrix: Comparing the Forbidden City and the Palace of Versailles across three dimensions: architectural style, garden layout, and palace function (reference standard text ID: TEXT_VERSAILLES_INTRO).

[0279] Thematic Bilingual Reading: Generates bilingual (Chinese and English) explanatory texts that focus on the dimension of comparison.

[0280] Target vocabulary reinforcement: Deeply integrate commonly confused junior high school vocabulary: ["architecture", "structure", "civilization", "style", "heritage", "layout"], and use the school-based differentiation template (template ID: TEMPLATE_VOCAB_JUNIOR_01).

[0281] Key knowledge points: Chinese and Western architectural aesthetics, garden design philosophy, and the social structure of the imperial court.

[0282] Interaction requirements: Support side-by-side comparison views on the same dimension, click on new words to bring up the smart assistant, and one-click switching between Chinese and English content display.

[0283] School-based features: Using a learning path for commonly misspelled vocabulary in junior high school English, customized by the school.

[0284] II. Application Generation Phase:

[0285] In this stage, the system automatically completes the technical implementation based on the confirmed design scheme.

[0286] 3. Generate Application: After confirming the requirements and specifications, the teacher clicks the "Generate Application" button.

[0287] The development environment submits the requirements specifications to the intelligent engine layer;

[0288] The application orchestration engine takes over the process and coordinates the generation of specific content by the teaching AI service layer;

[0289] The engine performs technical orchestration and assembles an executable application blueprint.

[0290] The application assembly blueprint is saved to the teaching application definition library, and the preview result is returned to the teacher for confirmation.

[0291] III. Application Execution Phase:

[0292] 4. Release and Application: After teacher confirmation, the application is released. The application assembly blueprint is stored in the teaching application definition library and registered and configured in the teaching application execution environment.

[0293] 5. Learning Interaction: When students access the application, the execution environment dynamically calls the teaching AI service layer to provide a personalized learning experience.

[0294] IV. Closed-loop optimization stage:

[0295] 6. Data Feedback: The execution environment records learning behavior data to the teaching data layer;

[0296] 7. System Evolution: Learning data is used to optimize student profiles and learning analysis, providing more accurate data support for subsequent applications and forming a complete closed loop of continuous optimization.

[0297] To facilitate understanding of the process of developing instructional applications, combined with Figure 5 Let's illustrate with examples. Figure 5 Example diagram of a development environment for teaching applications. Figure 5 Taking the development of the "Palace Museum Fun Learning" app as an example, this paper illustrates the design and development process of the app.

[0298] I. AI-Dedicated Teaching Components:

[0299] Figure 5 In the middle section, to the left of the "Requirements Description" is the "AI Teaching Components" area. The AI ​​Teaching Components Library is the core innovation of this solution platform. It provides teachers with a series of drag-and-drop, configurable teaching function modules, encapsulating complex AI technologies into functional units specifically for educational scenarios, enabling teachers to quickly build personalized intelligent teaching applications without programming.

[0300] In this solution, based on an understanding of educational scenarios, teaching-specific components are categorized and encapsulated, as described below:

[0301] 1. Mathematics and Science Exploration: Laboratories that cultivate logical thinking and scientific inquiry abilities, as shown in Table 1.

[0302] Table 1

[0303] AI components Support for teachers' creative innovation Application scenarios Concept Visualization Transforming abstract knowledge such as formulas and theorems into perceptible interactive models enables teachers to design "visualized" science courses. Teachers designed an application of the "buoyancy principle" that allows students to observe water level changes and calculate buoyancy in real time by dragging different objects. Visualizing the Principle This allows teachers to demonstrate dynamic scientific processes (such as photosynthesis) or invisible mechanisms (such as electromagnetic fields), breaking through the static limitations of traditional teaching. In the "Fun Tour of the Forbidden City" app, students can "disassemble" the dougong brackets themselves to understand their mechanical and earthquake-resistant principles. Virtual operation of experiments Provide a safe, low-cost, and reproducible experimental environment that allows teachers to design inquiry activities that include high-risk, high-cost, or idealized conditions. Teachers create a "cosmic velocity simulator" where students can adjust the mass and radius of celestial bodies to intuitively understand the principles of rocket launch. Modeling and Analysis Empower teachers to design data-driven inquiry tasks and cultivate students' data analysis skills and scientific empirical spirit. Teachers designed a "Urban Temperature Analysis" project, in which students used components to analyze real meteorological data and explore the "heat island effect".

[0304] 2. Morality and Civic Literacy: An "immersion field" for shaping values ​​and civic awareness, as shown in Table 2.

[0305] Table 2

[0306] AI components Support for teachers' creative innovation Application scenarios Contextualized Moral Experience Teachers can create realistic moral dilemmas or social event simulations to guide students in forming value judgments through "hands-on experience". Teachers designed an "Integrity Store" application where students make shopping decisions in an unattended virtual store, and the system records and guides reflection. Role-playing and empathy Support teachers in designing role-playing activities that allow students to think about issues from different perspectives, thereby cultivating empathy and multiple perspectives. In the "Community Dispute Mediation" application, students take on the roles of residents, property management staff, and members of the owners' committee, experiencing the process of negotiation and compromise. Social Issues Exploration Provide teachers with tools to build exploration spaces for complex local or global issues (such as environmental protection and equity), connecting learning with society. Teachers created the "Exploring Local Intangible Cultural Heritage" project around the theme of "Loving Hometown," guiding students to research, record, and digitally protect local culture. Behavior tracking and feedback This helps teachers embed the recording and positive guidance of students' behavioral choices into the learning journey they design, thereby achieving process-oriented assessment. In the "class self-governance" application, students' collaboration, contributions, and other behaviors are recorded and converted into "citizen points" to incentivize their social responsibility.

[0307] 3. Language and Cultural Exchange: “Dialogue Bridges” to enhance communication skills and cross-cultural understanding, as shown in Table 3.

[0308] Table 3

[0309] AI components Support for teachers' creative innovation Application Examples Personalized language tutoring This allows teachers to provide adaptive language practice and real-time feedback for students of different levels, achieving true differentiated instruction. Embed a smart assistant in English reading apps to provide vocabulary and grammar tips for students who encounter difficulties. Authentic Context Creation It empowers teachers to simulate real-life language use scenarios (such as asking for directions, negotiating, and giving speeches), going beyond the mechanical drills in textbooks. Teachers created a "virtual international summer camp" application where students need to complete a series of collaborative tasks in English with AI-generated international partners. Cultural Comparison and Understanding Provide tools that allow teachers to juxtapose customs and ideas from different cultures and design activities to guide students to understand and respect cultural diversity. In the "Fun Tour of the Forbidden City" activity, students learn about the similarities and differences between the Forbidden City and the Palace of Versailles in terms of architecture, culture, and etiquette. Critical thinking and expression training It supports teachers in designing complex discussion topics, analyzes students' argumentation logic and expression structure through assessment and review components, and provides suggestions for improvement. Teachers organized online debates on the theme of "advantages and disadvantages of technology," and the system intelligently analyzed the students' arguments, evidence, and rebuttals.

[0310] 4. Arts and Health Category: "Creative Workshops" that inspire creativity and nourish the mind and body, as shown in Table 4.

[0311] Table 4

[0312] AI components Support for teachers' creative innovation Application Examples Low-threshold art creation By reducing the technical difficulty, teachers can design integrated art courses that combine music, painting, and design, thereby stimulating the creative potential of each student. Teachers designed an application that combines classical poems with illustrations. After students recited classical poems, they used AI tools to generate artistic paintings and shared their creative ideas. Aesthetic experience and appreciation It provides a wealth of multimedia resources and interactive tools to help teachers build in-depth art appreciation activities and enhance students' aesthetic literacy. In the "Walking Through Masterpieces" app, students can "step into" the "Along the River During the Qingming Festival" scroll, explore the details in the painting, and listen to explanations of its historical context. Health Habit Simulation and Development This allows teachers to create engaging and interactive tasks that guide students to learn about health, such as nutrition, exercise, and mental well-being, through games. Teachers designed the "My Healthy Plate" app, where students can drag and drop food items to create their meals, and the system provides real-time nutritional analysis and suggestions. Emotion Recognition and Management Provide support for teachers to design interactive courses that help students understand emotions and learn stress management. Teachers created an "Emotion Weather Station" app where students record their emotions daily. The system then displays these emotional changes through charts and provides adjustment tips.

[0313] 5. Classification of Technology and Labor Practices: "Workshops for Making Things" from Concept to Practice, as shown in Table 5.

[0314] Table 5

[0315] AI components Support for teachers' creative innovation Application Examples Mastering digital tools This will enable teachers to seamlessly integrate new-era labor skills such as computational thinking, data analysis, and AI applications into their subject teaching. In history class, the teacher designed a project called "Looking at the Silk Road with Data," in which students were required to clean and analyze trade data and visualize the conclusions. Project planning and execution Provide project management tools (such as Gantt charts and task lists) to support teachers in designing long-term, interdisciplinary project-based learning. The faculty organized a project called "Designing a Smart Water-Saving System for the Campus," in which students used components to conduct research, design solutions, assign tasks, and manage progress. From design to creation It connects virtual design with physical creation, enabling teachers to guide students through the entire process from ideation and modeling to 3D printing. In the "Design My Backpack" application, students first create a virtual design, and after optimization, the model file can be exported for 3D printing. Labor Values ​​Experience By simulating labor scenarios such as production and service, teachers can guide students to appreciate the value of labor and the spirit of cooperation and innovation. Teachers create a "class company" application, and students form teams to operate it, taking on roles such as product, marketing, and finance, and experiencing the operation of a company.

[0316] II. Assembly Requirements:

[0317] This solution's research-application integrated teaching system, designed collaboratively by humans and an intelligent engine, allows users to transform ideas into clear, actionable requirements specifications by simply dragging and dropping or selecting AI-dedicated teaching components, thus quickly building structured application requirement descriptions. The operation steps are as follows:

[0318] 1. Adding and configuring components:

[0319] From the "AI Component List" on the left, drag and drop the required components to the central assembly area. For example, to create a "Fun Learning about the Forbidden City" app, users can drag and drop the following components in sequence:

[0320] Cultural understanding and comparison within the category of language and cultural exchange;

[0321] Visualization of principles within the category of Mathematics and Science Exploration;

[0322] Language tutoring under the category of Language and Cultural Exchange;

[0323] Select a component from the AI ​​component list, and the corresponding property configuration panel will pop up on the right. Users can fill in or select relevant options in detail according to their teaching needs.

[0324] For example, configuring a cultural understanding and comparison component:

[0325] Topic: Interdisciplinary and Fun Learning about the Forbidden City

[0326] Core objective: To enable students to learn about the Forbidden City while integrating subject knowledge into their understanding and practice, learning while having fun.

[0327] Target audience: Junior high school students;

[0328] Comparative perspectives: historical changes, Chinese and foreign perspectives;

[0329] Main contents: historical positioning, dialogues with historical figures, architectural features, short English stories, poster design, etc.

[0330] For example, configuring a principle visualization component:

[0331] Subject: Earthquake Resistance of the Forbidden City Architecture;

[0332] Visualization type: 3D interactive model;

[0333] Key knowledge points: Principles of mechanics in dougong (bracket set) structures;

[0334] For example, configuring a language tutoring component:

[0335] Assistant activation method: Click;

[0336] Default assistant avatar: Little Sugar;

[0337] Default tutoring method: text.

[0338] 2. Adjust the component order:

[0339] Users can drag and drop component cards in the required assembly area to adjust their logical order or presentation priority within the application.

[0340] 3. Generate final description:

[0341] Once all components are configured, please click the "Generate Description" button below.

[0342] The platform's AI will intelligently integrate the user's configuration information for all components, generate a complete and structured description of application requirements, and automatically fill it into the final confirmation dialog box on the right.

[0343] Users can review and fine-tune the generated description in this dialog box to ensure that it fully meets their expectations.

[0344] III. Message Integration:

[0345] Figure 5 The right side is the message dialog area. Main functions and features:

[0346] 1. Automatically generate and present a complete description of the requirements specification:

[0347] Triggering mechanism: After a user configures the various AI components in the requirement assembly area and clicks the "Generate Description" button, the platform will integrate the attributes of all components (such as theme, target, content elements, visualization type, etc.) into a logically coherent and detailed requirement specification description.

[0348] Presentation format: This fully generated requirement specification description will automatically fill in the input boxes in the dialog area, ready to go. This makes the entire requirement conception process transparent to the user, allowing them to see how the AI ​​will understand their needs.

[0349] 2. The interface for final editing and refining:

[0350] After the complete requirements specification is automatically generated, the user has the final right to review and modify it. This includes:

[0351] Fine-tuning the wording: making the instructions more precise or conforming to the user's specific language habits;

[0352] Additional details: New ideas that come to mind can be added directly to the description of the requirements specification;

[0353] Simplify or complicate: Adjust the level of detail in the description according to the user's needs. This step is a crucial final polishing step for the user's application requirements.

[0354] 3. The central hub for direct interaction with the AI ​​assistant:

[0355] Once the requirement specification is ready, users can click "↑" to formally submit the requirement to the system's core AI.

[0356] Afterward, the chat area will transform into an interactive chat interface, where users can:

[0357] Requires explanation: For example, asking "Why was this method chosen to visualize the dougong (bracket set) structure?";

[0358] Propose alternative solutions: For example, "In the comparison between the Forbidden City and the Palace of Versailles, the dimension of garden style should be replaced with the dimension of layout concept";

[0359] Expand upon this: For example, “Add two test questions to each section on easily confused words.”

[0360] 4. Multimedia Input Requirements: Beyond the Limitations of Text:

[0361] Attachment Upload: This is a powerful supplement to the requirement description. Users can click the upload button to provide relevant files as references to the AI, including:

[0362] Image: For example:

[0363] A photograph of the Hall of Supreme Harmony in the Forbidden City was provided, with the instruction: "I want the main visual and color style to be referenced from this image."

[0364] Documents: For example:

[0365] A Word document containing short English stories: Its content can be directly imported into the "Language Tutoring" or "Cultural Understanding" modules.

[0366] IV. Preview and Interactive Testing:

[0367] 1. When a user clicks the preview page, the platform will render a high-fidelity application interaction prototype in real time based on the generated application assembly blueprint. An example of a specific preview page is shown below. Figure 6 and Figure 7 As shown.

[0368] 2. In this mode, users can perform actions such as clicking, dragging, and inputting as if using a real application, including experiencing dialogue with the virtual assistant, operating the 3D bracket model, and accessing core functions such as word recognition.

[0369] 3. If any adjustments are needed during the preview, users can switch back to the "Requirements Description" page to modify the component requirements or directly edit the description text.

[0370] V. Code Generation and Viewing:

[0371] 1. Once the user is satisfied with the preview, they can click the "Code" option page.

[0372] 2. The platform of the research and application integrated teaching system, which is designed based on the collaboration between humans and intelligent engines, will automatically generate clear and readable front-end (such as HTML, CSS, JavaScript) and necessary back-end integration code according to the final application description.

[0373] 3. This page is primarily intended to provide a starting point for users with further development capabilities to reference, learn, or perform secondary development, allowing users to view and copy the generated code.

[0374] VI. Saving and Publishing:

[0375] 1. At any stage of the process, users can click the "Save" button at the top of the interface to save the current application design to their personal project library.

[0376] 2. Once the application is fully designed and tested, click the "Publish" button. The system will store the application assembly blueprint as a new record in the teaching application definition library. Based on the deployment environment policy, the system will distribute the blueprint and its dependent static resources to one or more resource storage nodes that can be efficiently accessed by the 'application runner'. A record generation entry will also be registered in the portal for end users (students) to access.

[0377] For a detailed workflow of application design and application generation, please refer to [reference needed]. Figure 8 Let's illustrate with examples.

[0378] Figure 8 Workflow description (using the "Palace Museum Fun Learning" application as an example):

[0379] Phase 1: Application Design - Requirements Specification Generation

[0380] Objective: To transform teachers' teaching ideas into structured "functional and content requirements specifications";

[0381] 1. User Configuration:

[0382] Teachers select three core components on the application development side:

[0383] Cultural understanding and comparison: The theme is "Interdisciplinary Learning at the Palace Museum," and content elements include historical positioning and dialogues between figures.

[0384] Visualizing the principle: Configure the theme "Bougong Structural Mechanics" and select the 3D interactive model type;

[0385] Language tutoring: Configure the assistant avatar "Little Sugar" and set up text-based tutoring methods.

[0386] 2. Intelligent integration and context building:

[0387] After receiving the configuration, the message integration engine constructs a rich teaching context through internal and external collaboration:

[0388] Internal data query:

[0389] Subject knowledge points: the construction history of the Forbidden City, the structural principles of the dougong (bracket set) system, and commonly confused English vocabulary related to the Forbidden City;

[0390] Student profile: Spatial cognition characteristics, historical knowledge base, and learning behavior preferences of junior high school students;

[0391] School Information Database: Stories of alumni's connection with the Palace Museum, outstanding past study tour practices and case studies, etc.

[0392] Teaching templates: timeline templates, character dialogue templates, poster design templates, etc.;

[0393] Intelligent acquisition of external resources (through a security filtering layer):

[0394] Get the latest exhibition information, high-resolution architectural detail images, etc. from the official website of the Palace Museum;

[0395] Obtain comparative data on Chinese and foreign palaces from open knowledge platforms, etc.

[0396] 3. Generate requirement specifications:

[0397] The following is an example of a structured document that outputs the following content:

[0398] I. Application Overview:

[0399] Application Name: Fun Tour of the Forbidden City, Interdisciplinary Inquiry-Based Learning Application

[0400] Target audience: Junior high school students (with cross-disciplinary learning of history, physics, and English)

[0401] Core learning objectives:

[0402] 1. Through historical narratives and architectural deciphering, students can understand the overall value of the Forbidden City as a cultural heritage and cultivate comprehensive humanistic qualities and national pride.

[0403] 2. Master the basic mechanical principles of the dougong structure and its earthquake resistance mechanism, establish a connection between physical knowledge and real-world applications, and cultivate scientific inquiry and spatial thinking abilities.

[0404] 3. In a bilingual immersion context, enhance the ability to understand and use English vocabulary on cultural and architectural themes, and develop a preliminary awareness of cross-cultural comparison.

[0405] II. Content Module Specifications:

[0406] A blend of East and West:

[0407] Educational Objectives: By comparing the Forbidden City and the Palace of Versailles, students will understand the differences between East and West in architectural philosophy, garden art, and court culture, and enhance their cultural appreciation and bilingual expression skills.

[0408] Content elements:

[0409] Cultural Comparison Matrix: Comparing the Forbidden City and the Palace of Versailles across three dimensions: architectural style, garden layout, and palace function (reference standard text ID: TEXT_VERSAILLES_INTRO).

[0410] Thematic Bilingual Reading: Generates bilingual (Chinese and English) explanatory texts that focus on the dimension of comparison.

[0411] Target vocabulary reinforcement: Deeply integrate commonly confused junior high school vocabulary: ["architecture", "structure", "civilization", "style", "heritage", "layout"], and use the school-based differentiation template (template ID: TEMPLATE_VOCAB_JUNIOR_01).

[0412] Key knowledge points: Chinese and Western architectural aesthetics, garden design philosophy, and the social structure of the imperial court.

[0413] Interaction requirements: Support side-by-side comparison views on the same dimension, click on new words to bring up the smart assistant, and one-click switching between Chinese and English content display.

[0414] School-based features: Using a learning path for commonly misspelled vocabulary in junior high school English, customized by the school.

[0415] Example of a message integration engine implementation:

[0416] The message integration engine can be built as an intelligent agent with knowledge of the teaching domain, using a large learning model (LLM) as its "brain" and the system's teaching data layer as its "long-term memory and toolkit," to complete the transformation from component configuration to a complete specification.

[0417] Phase 1: Planning and Insight

[0418] After the intelligent agent receives the user's component configuration (cultural comparison, principle visualization, language tutoring), its "brain" (LLM) will activate:

[0419] 1. Understanding Intent: This interdisciplinary inquiry project integrates humanistic comparison, scientific deciphering, and language support, and is targeted at junior high school students.

[0420] 2. Planning Task: "To generate a complete specification, I need to solve the following in sequence: define the overall goal, expand each component into an independent module, fill each module with specific teaching content and resources, and finally integrate them into a document."

[0421] Phase Two: Invoking the tool and obtaining the context:

[0422] The agent begins to execute according to plan, calling various tools:

[0423] Use the "Knowledge Graph Engine" tool to: "Search for key historical nodes and figures directly related to the 'History of the Forbidden City,' as well as the physical principles involved in the 'dougong structure.'"

[0424] Use the "Teaching Data Layer Query" tool to query the curriculum standards for the target grade level (junior high school); retrieve records related to the Forbidden City from the school's historical archives; and obtain templates for frequently missed questions in junior high school English.

[0425] Use the "Resource Library Search" tool to: "Find the resource ID and metadata (whether disassembly is supported) of the available 'Dou Gong 3D Interactive Model'; obtain authoritative text and image resource links for 'Versailles Palace'."

[0426] Phase 3: Structure Generation and Verification

[0427] After the agent obtains the structured data returned by all the tools, the LLM core begins to work:

[0428] Integration and Drafting: LLM integrates the scattered data returned by the tool (such as {Event: The Forbidden City was completed in 1420, Person: Zhu Di, Resource ID: model_001}) with the initial teaching objectives, and fills them into a pre-set, extremely detailed "requirements specification template" using professional teaching language.

[0429] Format validation: After the initial draft is generated, the system will invoke a format validation rule (or let the LLM perform a self-check) to ensure that the specification document contains all required chapters, module descriptions are complete, and resource references are formatted correctly.

[0430] Key technical implementation points:

[0431] 1. Intelligent agent architecture: The ReAct (Reasoning + Acting) model or Meta AI's Agent Framework can be adopted to enable LLM to cycle through "thinking-acting-observing".

[0432] 2. Tool encapsulation: The system's "knowledge point graph engine" and "data layer query" are all encapsulated into tools with clear functional descriptions for intelligent agents to call.

[0433] 3. System Prompt Engineering: The core prompt for the agent needs to precisely define its role. For example: "You are a senior instructional designer, responsible for configuring and integrating scattered instructional components into a professional, complete, and immediately executable application requirement specification. You must obtain accurate data by calling tools and strictly follow the output template."

[0434] 4. Output template control: Provide strict, exemplified structured output templates (such as using JSON Schema or an explicit Markdown heading system) to force LLM to organize information in this format and ensure output stability.

[0435] Phase Two: Application Generation - From Specification to Runnable Application

[0436] Objective: To transform requirements specifications into an interactive application preview;

[0437] 4. Technical orchestration and resource coordination:

[0438] The application orchestration engine first parses the requirement specifications, such as identifying modules, extracting resource lists, and clarifying interaction constraints, and then coordinates resources from all parties for scheduling and production.

[0439] Call the content generation engine: for example, generate an English story comparing the Forbidden City and Versailles, embedding easily misspelled words;

[0440] Utilize a knowledge graph engine: for example, to establish connections between historical events and architectural features;

[0441] Retrieve data from the teaching data layer based on IDs, such as 3D models of bracket sets, original texts of school history archives, and vocabulary discrimination templates.

[0442] Intelligent scheduling of external resources: floor plans and architectural detail drawings from the Palace Museum's official website, high-definition floor plans of the Palace of Versailles, etc.;

[0443] And perform adaptation and optimization, such as ensuring that the 3D model format is compatible with WebGL, and converting image materials into versions adapted to different screen resolutions.

[0444] Assembly and generation:

[0445] The application orchestration engine assembles all the aforementioned "semi-finished products" and "parts" according to a predefined blueprint JSON schema: Module definition: Detailed JSON definitions are generated for each module. For example:

[0446] {

[0447] "modules": {

[0448] "history_timeline": {

[0449] "type": "InteractiveTimeline",

[0450] "config": {

[0451] "dataSource": "The generated array of historical nodes and stories",

[0452] "supportLang": ["zh", "en"]

[0453] }

[0454] },

[0455] "palace_comparison": {

[0456] "type": "BilingualComparisonViewer",

[0457] "config": {

[0458] "content": "Bilingual comparative stories generated by AIGC"

[0459] "highlightWords": A list of frequently misspelled words.

[0460] }

[0461] }

[0462] }

[0463] }

[0464] Design the interaction logic (routing logic): Describe the event flow using a declarative language. For example:

[0465] {

[0466] "routing_logic": [

[0467] {

[0468] "from": "palace_comparison.word_click",

[0469] "to": "smart_assistant.query",

[0470] "data_mapping": "clickedWord -> question"

[0471] },

[0472] {

[0473] "from": "smart_assistant.answer_ready",

[0474] "to": "floating_card.show",

[0475] "data_mapping": "answer -> content"

[0476] } ]

[0478] }

[0479] This describes the complete interaction chain of "clicking on a new word -> calling the smart assistant -> displaying the answer in a floating card".

[0480] Inject global configuration: Set the application's default language, theme color, bind the smart assistant service ID (ASSISTANT_XIAOTANG_V2), etc.

[0481] Optimization: The application orchestration engine sends the initial draft (a version that conforms to the JSON Schema but may be rough and incomplete) or its key parts, along with the design goals and constraints, to the large model for "supplementary optimization".

[0482] Output and Delivery: Finally, the engine produces a complete, compliant application assembly blueprint JSON file. This application assembly blueprint can be directly stored in the "Teaching Application Definition Library" (for the application runner to interpret and execute in real time), or packaged into a "Production Instruction Set" and sent to the cloud-based generation engine to be packaged into a standalone application.

[0483] Example of large model integration process:

[0484] Step 1: Constructing precise "blueprint generation instructions" using an orchestration engine. The engine needs to package the results of all the preliminary work (generated stories, acquired images, templates) into a "fill-in-the-blank question" that a multimodal LLM can accurately understand.

[0485] Key actions:

[0486] 1. Prepare a "fill-in-the-blank" template: Create a pre-defined JSON Schema template that conforms to your "Application Assembly Blueprint Specification" and contains a large number of {{placeholder}} that need to be populated by the LLM.

[0487] 2. Inject resources: Inject the pre-processed resource links and content (such as story text, 3D model URLs, and vocabulary discrimination JSON) into the corresponding positions in the template.

[0488] 3. Final Synthesis Instructions: Send the template and detailed fill rules to the LLM via "system prompts".

[0489] A simplified example of the instruction:

[0490] / / Message sent by the orchestration engine to the LLM

[0491] {

[0492] [Task Name] Cultural Comparison Experience: The Forbidden City and the Palace of Versailles

[0493] [Core Teaching Needs]:

[0494] 1. Content: Compare the architectural styles and functions of the Forbidden City and the Palace of Versailles, naturally incorporating and highlighting at least 5 commonly misspelled English words in junior high school.

[0495] 2. Interaction: It must support switching between Chinese and English interfaces and provide an interactive function to identify easily confused words.

[0496] [Resource Input - This is non-changeable, deterministic content]:

[0497] 1. Bilingual story text:

[0498] Title: "A Tale of Two Palaces"

[0499] Main text: "[This is a fully generated bilingual (Chinese and English) story, in which 5 target words have been used..." <highlight>[Tag highlighting]

[0500] 2. Data on commonly confused words:

[0501] `[{"word": "heritage", "explanation": "...", "examples": [...]}, ...]` (5 entries in total)

[0502] 3. Digital resource address:

[0503] 3D model of the Hall of Supreme Harmony: `https: / / cdn.yourschool.com / models / thd_v2.glb`

[0504] 3D model of the Hall of Mirrors at Versailles: `https: / / cdn.yourschool.com / models / versailles_mirrorhall.glb`

[0505] The output should be a JSON object, which must contain the following top-level structure:

[0506] {

[0507] "app_metadata": { ...}, / / Please generate reasonable application name, description, and other metadata.

[0508] "modules": { ...}, / / Defines each module and its configuration

[0509] "routing_logic": [...] / / Defines the interaction and data flow logic between modules.

[0510] }

[0511] [Reasonable Design Guidelines]:

[0512] Please optimize the design in the following aspects and reflect the design in the JSON above:

[0513] 1. **Practicality of Bilingual Switching**: Design the `bilingual_ui_controller` module to enable global language switching and consider the refresh logic of UI elements during switching.

[0514] 2. **Interactive details of vocabulary discrimination**: Design the specific conditions for the `vocabulary_drill` module to be triggered (such as clicking on the highlighted words in the story), and plan the switching or display methods of its displayed content (definitions, example sentences).

[0515] 3. **Natural Inter-Module Linkage:** In `routing_logic`, you can design whether `story_reader` can automatically scroll to the relevant story paragraph when the user views a specific component in `3d_comparison_viewer`, and vice versa.

[0516] 4. **Default and Fault Tolerance**: Set reasonable default values ​​for all configurable parameters (such as initial language and initial model viewpoint).

[0517] Ensure that all designs are **necessary, clear, and technically feasible**, with the goal of ensuring unambiguous interaction and a smooth user experience in the final application.

[0518] Output Requirements:

[0519] Step 1: Directly output a complete and valid JSON application assembly blueprint.

[0520] Step 2: Call LLMAPI and obtain the "blue image segment to be assembled":

[0521] Invocation: The engine sends the above-constructed instructions via an API (such as Google AI Studio or Vertex AI).

[0522] Expected output: The LLM should return a JSON structure that has been preliminarily populated with content (i.e., "blue image segment to be assembled").

[0523] Step 3: Local post-processing, verification, and generation of the final blueprint:

[0524] 1. Parsing and Extraction: Extract the JSON portion from the LLM's response.

[0525] 2. Structure Validation and Error Correction: Use a local validator to check whether the JSON conforms to the Blueprint Schema and automatically correct common errors (such as missing fields and format errors).

[0526] 3. Inject engine control parameters: Inject the necessary metadata of the blueprint (such as unique ID, version number, and binding configuration with the background "Smart Assistant" service).

[0527] 4. Generate the final blueprint: Output a fully compliant final application assembly blueprint file that can be stored in the "Teaching Application Definition Library".

[0528] Phase 3: Application Execution and Closed-Loop Optimization

[0529] Objective: To launch and continuously optimize the application;

[0530] Application publishing and use;

[0531] After the application is released, students can access it through the teaching application execution environment;

[0532] Real-time access to teaching AI services:

[0533] When students read English stories, they can access language tutoring services for vocabulary analysis.

[0534] When students explore alumni stories, the knowledge graph engine is used to display related historical information.

[0535] Data feedback and system evolution:

[0536] The learning logger collects the following data:

[0537] Students' mastery of commonly misspelled English words;

[0538] Distribution of learning difficulties in 3D model manipulation;

[0539] Closed-loop optimization effect:

[0540] Based on data feedback, the subsequently generated applications will:

[0541] Strengthen the English vocabulary learning process, which is a common challenge for students;

[0542] It provides more precise auxiliary explanations to address the difficulties in understanding the dougong (bracket set).

[0543] This solution proposes a complete, end-to-end teaching system based on human-AI engine collaboration, encompassing a teaching application development environment module, an intelligent service platform, a message integration engine, an application orchestration engine, and a teaching application execution environment module. This system achieves full-chain automation from teaching ideation to teaching practice, from expressing teaching intent and intelligent generation to operational feedback. Through collaborative querying of multi-source data (subject knowledge, student profiles, school information, and teaching templates), the system transforms simple component configurations by teachers into structured requirement specifications rich in contextual information, enabling the understanding of educational semantics and the intelligent construction of teaching context. Furthermore, this solution provides a dedicated AI component library for education (such as "Cultural Understanding and Comparison" and "Principle Visualization") using a low-code development paradigm specific to the education field. Designed using educational semantics rather than technical components, it significantly lowers the barrier to entry and supports the automatic integration and personalized adaptation of school-based content. This solution clarifies the four-stage workflow of "design → generation → execution → optimization"; establishes a closed-loop mechanism that drives the continuous evolution of the system through feedback of learning behavior data; and achieves adaptive optimization capabilities that become "smarter with use".

[0544] This solution's research-application integrated teaching system, designed collaboratively by humans and intelligent engines, enables frontline teachers to independently realize their teaching ideas, eliminating reliance on technical personnel, significantly lowering the technical threshold, shortening the application development cycle from weeks to minutes, quickly validating teaching ideas, and improving the efficiency of teaching innovation; data-driven personalized design enhances teaching effectiveness and strengthens teaching relevance.

[0545] In this embodiment, a research-application integrated teaching system based on human-intelligence engine collaborative design is constructed, capable of understanding educational semantics, collaborating with AI capabilities, and integrating data assets. This system comprises a closed-loop architecture consisting of a teaching application development environment, an intelligent service platform, and a teaching application execution environment. When an application needs to be created, the system deeply integrates and semantically understands the design intent and teaching context of the application, resulting in a structured application requirement description. This description is then automatically arranged into applications, achieving full-link automation and intelligence from teaching ideas to teaching practice. This enables educators to efficiently and accurately create dynamic, personalized, interactive teaching applications that align with specific teaching objectives, adapt to different students' cognitive levels, and incorporate local school characteristics.

[0546] refer to Figure 9 The illustration shows a research-application integrated teaching method based on human-intelligent engine collaborative design, disclosed in this application. This research-application integrated teaching method based on human-intelligent engine collaborative design is applied to the above embodiment. Figure 1 The aforementioned research-application integrated teaching system based on human-intelligent engine collaborative design includes a teaching application development environment module, an intelligent service platform, and a teaching application execution environment module; wherein, the intelligent service platform includes at least a message integration engine and an application orchestration engine; the research-application integrated teaching method based on human-intelligent engine collaborative design mainly includes the following steps:

[0547] S901: Obtain teacher configuration information by calling the teaching-specific AI component through the teaching application development environment module; wherein, the teaching-specific AI component is obtained by domain modeling; the teacher configuration information includes at least the teaching application design intent;

[0548] S902: The message integration engine is used to obtain the teaching context, and the teaching context and the teaching application design intent are deeply integrated and semantically understood to obtain a structured description of application requirements.

[0549] S903: The application requirements description is orchestrated through the application orchestration engine to obtain the application assembly blueprint.

[0550] S904: The teaching application execution environment module is used to call AI services through the unified scheduling entry point of internal and external services to load application assembly blueprints and deploy them to the teaching application execution environment module for application execution and closed-loop optimization.

[0551] The execution process and execution principle of S901-S904 are the same as those in the above embodiments. Figure 1 The execution process and principles of the publicly available teaching system based on the collaborative design of humans and intelligent engines are consistent and can be used as a reference, so they will not be elaborated here.

[0552] In this embodiment, the solution constructs a research-application integrated teaching system based on human-intelligence engine collaborative design, capable of understanding educational semantics, collaborating with AI capabilities, and integrating data assets. This system comprises a closed-loop architecture consisting of a teaching application development environment, an intelligent service platform, and a teaching application execution environment. When an application needs to be created, the system deeply integrates and semantically understands the design intent and teaching context of the application, resulting in a structured application requirement description. This description is then automatically arranged into applications, achieving full-link automation and intelligence from teaching ideas to teaching practice. This enables educators to efficiently and accurately create dynamic, personalized, interactive teaching applications that align with specific teaching objectives, adapt to different students' cognitive levels, and incorporate local school characteristics.

[0553] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0554] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0555] The steps in the methods of the various embodiments of this application can be adjusted, combined, or deleted according to actual needs.

[0556] Finally, it should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0557] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0558] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.< / highlight>

Claims

1. A research-application integrated teaching system based on collaborative design between humans and intelligent engines, characterized in that, The system includes a teaching application development environment module, an intelligent service platform, and a teaching application execution environment module; wherein, the intelligent service platform includes at least a message integration engine and an application orchestration engine; The teaching application development environment module is used to obtain teacher configuration information through a teaching-specific AI component; wherein, the teaching-specific AI component is obtained through domain modeling; and the teacher configuration information includes at least the teaching application design intent. The message integration engine is used to obtain the teaching context and to deeply integrate and semantically understand the teaching context and the teaching application design intent to obtain a structured description of application requirements. The application orchestration engine is used to orchestrate the application requirements description to obtain an application assembly blueprint. The teaching application execution environment module is used to call the AI ​​service to load the application assembly blueprint through the unified scheduling entry point of internal and external services, and deploy it to the teaching application execution environment module for application execution and closed-loop optimization.

2. The system according to claim 1, characterized in that, The message integration engine is specifically used to call the knowledge point graph engine to establish interdisciplinary knowledge relationships; and to call the content generation engine to generate customized content that conforms to the theme and has teaching focus annotations through the interdisciplinary knowledge relationships. The customized content that conforms to the theme and has teaching focus markings is integrated to obtain a structured description of application requirements.

3. The system according to claim 1, characterized in that, The application orchestration engine is specifically used to acquire pre-stored digital teaching resources; perform intelligent scheduling of external resources to acquire and integrate the required multimedia materials, documents and map data from authorized and open external data sources; A structured application assembly blueprint is generated based on the digital teaching resources, multimedia materials, documents, and map data.

4. The system according to claim 1, characterized in that, The AI ​​component specifically designed for teaching is used for editing operations to obtain the user's configuration requirements.

5. The system according to claim 1, characterized in that, The intelligent service platform also includes a teaching AI service layer; the teaching AI service layer includes a natural language processing engine, an information search engine, an intelligent assistant service, a content generation engine, a knowledge point graph engine, a learning analysis engine, and a reasoning service; The natural language processing engine is used to parse natural language descriptions according to design requirements and transform them into structured application configurations, as well as to handle student interactions with the application. The intelligent assistant service is used to provide at least dialogue management, image and voice dispatch functions for the assistant function in the execution environment; The content generation engine is used to generate an initial content framework, sample materials, interactive designs, and personalized content adaptations for teaching applications based on teaching themes and objectives. The knowledge point graph engine is used to provide a semantic framework and reasoning foundation; The information search engine and the learning analysis engine are used to recommend learning resources and analyze learning data. The inference service is used to process local computing tasks with preset requirements.

6. The system according to claim 1, characterized in that, The teaching data layer includes at least a teaching-specific AI component library, a core viewpoint library for each subject, a learning behavior record library, an assistant resource library, a dialogue history library, a teaching application definition library, a teaching content template library, a student profile library, and a school information library. The dedicated AI component library for teaching is used to store reusable teaching function units; The core viewpoint databases for each discipline are used to store knowledge points for each discipline. Both the learning behavior record database and the student profile database are used to record student personality traits and learning process data. Both the assistant database and the dialogue history database are used to support personalized interaction of the intelligent assistant. The instructional application definition library is used to store the application assembly blueprints; The teaching content template library is used to store organizational templates for content from various subjects; The school information database is used to store school-based information; the school-based information includes at least the school icon, school history, and best practices.

7. The system according to claim 1, characterized in that, The teaching application development environment module includes an application designer, a preview debugger, and a release manager; The application designer provides a graphical interface and allows users to design instructional applications by selecting instructional AI components, configuring properties, and editing workflows. The preview debugger is used to preview and debug application effects in real time; The release manager is used to release debugged applications to the platform with one click and manage versions and permissions.

8. The system according to claim 1, characterized in that, The teaching application execution environment module includes an application portal, application runner, learning recorder, and intelligent assistant integration; The application portal is used to provide a personalized application list, learning progress tracking and navigation, student learning status, and to enable AI-supported lesson preparation and teaching research activities. The application runner is used to dynamically generate and run personalized learning content and / or teaching and research content based on the application definition and user profile. The learning recorder is used to record students' interactive behavior, answering questions, and learning outcomes throughout the entire process. The intelligent assistant integrates functions to provide at least multiple wake-up methods, multiple types, dialogue management, image, text scheduling, and voice scheduling.

9. The system according to claim 1, characterized in that, The intelligent service platform also includes an intelligent engine layer; the intelligent engine layer is equipped with an API gateway; the API gateway serves as a unified scheduling entry point for calling AI services and for calling generative AI and basic AI services on the public cloud.

10. A research-application integrated teaching method based on collaborative design between humans and intelligent engines, characterized in that... The method is applied to the research-application integrated teaching system based on human-intelligent engine collaborative design as described in any one of claims 1 to 9, wherein the system includes a teaching application development environment module, an intelligent service platform, and a teaching application execution environment module; wherein the intelligent service platform includes at least a message integration engine and an application orchestration engine; the method includes: The teaching application development environment module calls a dedicated teaching AI component to obtain teacher configuration information; wherein, the dedicated teaching AI component is obtained through domain modeling; and the teacher configuration information includes at least the teaching application design intent. The teaching context is obtained through the message integration engine, and the teaching context and the teaching application design intent are deeply integrated and semantically understood to obtain a structured application requirement description. The application requirements description is orchestrated using the application orchestration engine to obtain an application assembly blueprint. The AI ​​service is invoked through the unified scheduling entry point of the teaching application execution environment module and internal and external services to load the application assembly blueprint and deploy it to the teaching application execution environment module for application execution and closed-loop optimization.

Images