Accounting teaching method based on multi-dimensional task model

Abstract

The invention relates to the field of accounting teaching, in particular to an accounting teaching method based on a multi-dimensional task model, and the method comprises the steps: inputting a to-be-teaching task into a teaching platform, and constructing a task matrix; based on the task matrix, generating a corresponding ability evaluation question bank through an intelligent evaluation model; the students perform operation answering on the teaching platform based on the ability evaluation question bank; the intelligent evaluation model generates a primary ability evaluation score of the student according to the answer condition, and obtains the primary ability evaluation score of the student and an operation log through a data interface; the dynamic course configuration model dynamically adjusts a task matrix based on the initial ability evaluation score and the operation log; corresponding virtual roles are generated according to the task matrix, the virtual roles execute a dynamic game algorithm to perform dynamic game with students to realize accounting teaching, the teaching tasks can be dynamically adjusted according to strong and weak subjects of the students, teaching according to materials is realized, and the teaching quality is improved.

Description

Technical Field

[0001] The present invention relates to the field of accounting teaching, in particular to an accounting teaching method based on a multidimensional task model. Background Art

[0002] In today's education landscape, accounting, a highly practical and comprehensive discipline, requires high-quality teaching to cultivate accounting professionals who meet society's needs. Traditional accounting teaching methods primarily rely on classroom lectures, where teachers impart accounting expertise through theoretical explanations and case studies.

[0003] In traditional offline classrooms, teachers face a large student population, making it difficult to flexibly adjust the difficulty and progress of multi-dimensional task modules based on each student's knowledge base, learning ability, and interests. For example, existing tasks may lack challenge for students with a strong foundation, while students with weaker foundations may gradually lose confidence and motivation due to the excessive difficulty of the tasks. Offline teaching cannot dynamically adjust teaching tasks based on students' strengths and weaknesses in certain subjects, thus failing to achieve individualized teaching.

[0004] In offline teaching, after students complete a task, it often takes time for teachers to grade their work and provide feedback. This feedback often takes the form of focused lectures, making it difficult to provide timely guidance on individual student issues. Furthermore, it's difficult for teachers to quickly collect real-time data on students' performance during tasks, making it difficult to optimize and adjust teaching content and task design in a timely manner. This reduces the degree to which teaching meets students' actual needs, hindering teaching effectiveness.

[0005] To this end, this application proposes an accounting teaching method based on a multidimensional task model. Summary of the Invention

[0006] To solve the above problems, the present invention provides an accounting teaching method based on a multidimensional task model. Through this application, teaching tasks can be dynamically adjusted according to students' strong and weak subjects, so as to achieve teaching in accordance with their aptitude and improve teaching quality.

[0007] To achieve the above object, the technical solution adopted by the present invention is:

[0008] An accounting teaching method based on a multidimensional task model, including:

[0009] Obtain students' initial ability assessment scores and operation logs through the data interface;

[0010] The dynamic course configuration model dynamically adjusts the task matrix based on the initial competency assessment results and operation logs;

[0011] A corresponding virtual character is generated according to the task matrix, and the virtual character executes a dynamic game algorithm to play a dynamic game with students to realize accounting teaching.

[0012] Preferably, before obtaining the student's initial ability assessment score through the data interface, the method further includes:

[0013] Input the tasks to be taught into the teaching platform and construct a task matrix;

[0014] Based on the task matrix, the corresponding ability assessment question bank is matched through the intelligent assessment model;

[0015] Students answer questions based on the ability assessment question bank on the teaching platform;

[0016] The intelligent assessment model generates students' initial ability assessment scores based on their answers.

[0017] Preferably, the dynamic course configuration model dynamically adjusts the task matrix based on the student's initial ability assessment results, including:

[0018] The intelligent assessment model obtains the student's initial ability assessment results and generates an ability profile based on the initial ability assessment results;

[0019] The dynamic course configuration model calls the weight dynamic allocation model to dynamically adjust the weight of the teaching task based on the capability profile;

[0020] Dynamically adjust the task matrix according to the weight of the teaching tasks.

[0021] As a preferred method, the calculation formula of the weight dynamic allocation model is:

[0022]

[0023] Where W D,i，j is the weight of the j-th teaching task in the i-th adjustment, ability i is the comprehensive ability score after the i-th adjustment, W D,0，j is the initial weight of the i-th teaching task;

[0024] α is the degree of tool dependence, C K，i，j The task score of the j-th teaching task after the i-th adjustment, U J is the task urgency, R is the available resource score, γ is the time decay factor, β is the correct answer rate, C 0，j Score the preset task of the jth teaching task, W D,0，j is the initial weight of the i-th teaching task, and n is the number of teaching tasks.

[0025] Preferably, after the virtual role teaching model provides accounting teaching to students according to the teaching task, it further comprises:

[0026] Based on the adjustment task matrix, the corresponding ability assessment question bank is generated through the intelligent assessment model;

[0027] Students answer questions based on the ability assessment question bank through secondary operations on the teaching platform;

[0028] The intelligent assessment model generates students' secondary ability assessment scores based on their answers.

[0029] Preferably, the intelligent assessment model performs assessment based on the secondary ability assessment results, including:

[0030] The intelligent assessment model obtains students’ comprehensive ability scores;

[0031] If the comprehensive ability score meets the threshold, multiple teaching task scores are obtained;

[0032] Determine whether the scores of multiple teaching tasks are consistent with the preset task scores;

[0033] If the scores of multiple teaching tasks meet the preset task scores, the teaching will be exited;

[0034] If the scores of multiple teaching tasks do not all meet the preset task scores, the teaching tasks that meet the preset task scores will be removed;

[0035] The dynamic course configuration model dynamically adjusts the task matrix again based on the secondary ability assessment results and operation logs until the assessment results meet the threshold.

[0036] Preferably, when the comprehensive ability score meets the threshold, after obtaining the scores of several teaching tasks, the method further includes:

[0037] If the comprehensive ability score does not meet the threshold, the dynamic course configuration model will dynamically adjust the task matrix again based on the secondary ability assessment results and operation logs until the evaluation results meet the threshold.

[0038] Preferably, if the scores of the multiple teaching tasks do not all meet the preset task scores, the method further includes:

[0039] Obtaining teaching tasks that do not meet the preset task scores;

[0040] Obtaining the difference between the score of the teaching task and the score of the corresponding preset task through a PID controller;

[0041] When the difference is greater than the maximum threshold, the AR model is called to simulate teaching of the teaching task.

[0042] Preferably, the task matrix includes financial accounting, tax coordination, audit practice and cost control.

[0043] Preferably, the capability profile includes identity ID, knowledge domain, real-time weight and tool dependency.

[0044] The beneficial effects of the present invention are:

[0045] 1. This application sets up an initial ability assessment to obtain students' strong and weak subjects. Based on the initial ability assessment results and operation logs, the teaching content and task design are timely optimized and adjusted, so as to achieve teaching according to students' aptitude and improve teaching quality.

[0046] 2. This application can provide timely feedback during teaching by setting up a virtual role teaching model. Virtual role teaching solves the problems of subjective evaluation and slow feedback in accounting teaching through dynamic scene generation and multi-role collaborative game. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A flowchart of an accounting teaching method for a multi-dimensional task model in a specific embodiment of the invention;

[0048] Figure 2 A flow chart is provided for generating the initial ability assessment results in a specific embodiment of the invention;

[0049] Figure 3 Dynamically adjust the task matrix flow chart in a specific embodiment of the invention;

[0050] Figure 4 A flowchart for testing accounting teaching results in a specific embodiment of the invention;

[0051] Figure 5 This is a flowchart of a task matrix that is dynamically adjusted twice based on teaching results in a specific embodiment of the invention. DETAILED DESCRIPTION

[0052] Example 1

[0053] See also Figure 1-Figure 5 As shown, the present invention relates to an accounting teaching method of a multidimensional task model, comprising:

[0054] Input the teaching tasks to be taught into the teaching platform and construct a task matrix. In this embodiment, the task matrix includes financial accounting, tax planning, audit practice and cost control, that is, the number of teaching tasks is 4;

[0055] The task matrix includes the task urgency U j , R for available resources of teaching tasks, C for preset ability of teaching tasks 0,j , and the initial weight of each teaching task W 0,i，j ;

[0056] Based on the task matrix, the corresponding ability assessment question bank is generated through the intelligent assessment model, specifically including:

[0057] Through the core capability requirements in the NLP semantic parsing task (for example, the task description is "training cross-border M&A tax planning"), the semantic features of the core capability requirements are extracted through the BERT model and mapped to the knowledge graph to generate a structured vector;

[0058] Through structured vector matching of accounting ability assessment question bank, the ability assessment question bank that meets the matching value is selected. In the initial state, the initial weights of each teaching task in the ability assessment question bank are consistent. The specific initial weights are: Where n is the number of teaching tasks;

[0059] Students operate and answer questions based on the ability assessment question bank on the teaching platform, and the intelligent assessment model generates students' initial ability assessment scores and operation logs based on their answers.

[0060] Specific intelligent assessment models include:

[0061] Answer result collector, recording options / entries / numerical answers;

[0062] Operation log, tracking operation trajectory (mouse path / number of withdrawals);

[0063] Physiological signal sensor, collecting eye movements (pupil diameter) and brain waves (EEG);

[0064] Timing recorder, records the time taken for each step in milliseconds;

[0065] Tool usage times collector collects tool usage times to determine the tool dependency level. The specific tool dependency level α = actual usage times / quota usage times;

[0066] Obtain students' initial ability assessment scores and operation logs through the data interface;

[0067] The dynamic course configuration model dynamically adjusts the task matrix based on the initial competency assessment results and operation logs, including:

[0068] The intelligent assessment model obtains the student's initial ability assessment results and generates an ability profile based on the initial ability assessment results, specifically including:

[0069] The data collected by the intelligent evaluation model is used as data input, including:

[0070] Assessment report card (e.g., policy application = 72, risk sensitivity = 65, audit rigor = 68);

[0071] Behavioral data (timeout rate = 25%, number of withdrawals = 6);

[0072] Physiological data (cognitive load index CL = 0.82);

[0073] Use Z-Score to standardize the dimensions of input data, eliminate question differences, standardize input data, and correct behavioral data;

[0074] After standardization and correction, the input data is fused in multiple dimensions through a fusion algorithm to generate a capability portrait, which includes identity ID, knowledge domain, real-time weight, tool dependency and other data.

[0075] The dynamic course configuration model is based on the capability portrait, calls the weight dynamic allocation model to dynamically adjust the weight of the teaching task, and dynamically adjusts the task matrix according to the weight of the teaching task.

[0076] Specifically, the calculation formula of the weight dynamic allocation model is:

[0077]

[0078] Where W D,i，j is the weight of the j-th teaching task in the i-th adjustment, ability i is the comprehensive ability score after the i-th adjustment, W D,0，j is the initial weight of the i-th teaching task;

[0079] α is the degree of tool dependence, C K，i，j Score the task of the jth teaching after the i-th adjustment, U J is the task urgency, R is the available resource score, γ is the time decay factor, β is the correct answer rate, C 0，j Preset the ability score for the j-th teaching task, W D,0，j is the initial weight of the i-th teaching task, and n is the number of teaching tasks.

[0080] Based on the adjustment task matrix, the corresponding ability assessment question bank is generated through the intelligent assessment model;

[0081] Students answer questions based on the ability assessment question bank through secondary operations on the teaching platform;

[0082] The intelligent assessment model generates students’ secondary ability assessment scores based on their answers;

[0083] The intelligent assessment model is evaluated based on the results of the secondary ability assessment, including:

[0084] The intelligent assessment model obtains students’ comprehensive ability scores;

[0085] 1. When the comprehensive ability score meets the threshold, obtain scores for several individual teaching tasks;

[0086] If the scores of several individual teaching tasks meet the threshold, the teaching will be withdrawn;

[0087] If the evaluation results do not all meet the threshold, the teaching tasks that meet the threshold are removed. The dynamic course configuration model dynamically adjusts the task matrix based on the teaching tasks that do not meet the threshold, including:

[0088] Input the teaching tasks that do not meet the threshold into the teaching platform, and dynamically adjust the task matrix according to the weight of the teaching tasks;

[0089] Repeat the above steps until all evaluation results meet the threshold.

[0090] 2. When the comprehensive ability score does not meet the threshold, the PID controller is used to obtain the difference between several teaching task scores and their corresponding preset ability scores;

[0091] When the difference is greater than the maximum threshold, the specific maximum threshold is twice the minimum threshold, and the AR model is called to simulate the teaching task;

[0092] The dynamic course configuration model dynamically adjusts the task matrix based on the secondary ability assessment results and operation logs, and repeats the above steps until the assessment results all meet the threshold.

[0093] The above steps ensure that each teaching task of the students meets the threshold, avoiding the situation where the score of a certain teaching task is too high and affects the overall ability score.

[0094] The virtual role teaching model teaches accounting to students according to the task matrix, specifically including:

[0095] 1. AI analyzes student operation data (such as voucher error rate, audit vulnerability identification speed, basic voucher errors, and financial report analysis and decision-making), dynamically adjusts teaching content (such as pushing VAT case studies to weak links), records student operation data through blockchain, matches error types based on knowledge graphs, and triggers virtual character generation.

[0096] 2. Role 2. Virtual Expert Dialogue System

[0097] Technical implementation:

[0098] Multi-role AI tutor:

[0099] Financial AI: Resolving differences in standards (such as the new revenue standard);

[0100] Tax AI: simulate tax audit questions and answers (e.g., “Please explain the rationality of transfer pricing”);

[0101] Audit AI: Question the authenticity of the voucher (e.g., “Why is there no logistics order matching this revenue?”).

[0102] Cost AI: requires students to reduce costs while ensuring quality (e.g., “questioning omissions and requesting a new proposal”)

[0103] Using an identity switching model, we call the corresponding NLP knowledge base and dialogue strategy based on the task type (finance / tax / audit / cost).

[0104] 3. Real-time risk simulation coach

[0105] Technical implementation:

[0106] After students complete their tax planning proposals, AI simulates the consequences of different audit intensities (probability of back taxes, multiples of fines);

[0107] Generate a risk heat map (e.g., aggressive tax avoidance scheme → red high-risk warning).

[0108] By accessing the real data of the enterprise into the database, the risk exposure value is calculated by combining the policy library and historical audit cases.

[0109] Example 2

[0110] Based on the above-mentioned embodiment 1, in the accounting teaching system, the teaching content, equipment and virtual role simulation of financial accounting, tax planning, auditing practice and cost control include the following:

[0111] 1. Teaching content and equipment configuration

[0112] 1. Financial Accounting

[0113] Teaching content:

[0114] Core: Application of accounting standards (such as IFRS / GAAP), accounting entries, and preparation of financial statements (balance sheet / income statement / cash flow statement).

[0115] Difficulties: Consolidated financial statements, accounting for financial instruments, and revenue recognition (five-step method).

[0116] Teaching equipment:

[0117] Virtual enterprise simulation software: Unity 3D builds dynamic enterprise scenarios and generates transaction data streams in real time.

[0118] Smart voucher generator: OCR scans invoices to automatically generate electronic vouchers (such as UFIDA U8 interface).

[0119] Interactive reporting platform: Tableau / Power BI visually analyzes financial indicator anomalies.

[0120] Game algorithm selection: rule engine + state machine, mandatory verification of accounting standards (such as revenue recognition time / asset impairment).

[0121] AI Financial Director:

[0122] Responsibilities: Review reports prepared by trainees and make adjustment suggestions (e.g. "accrued expenses not fully accrued")

[0123] Example conversation:

[0124] Student: "How does the subsidiary's dividend distribution affect the consolidated financial statements?"

[0125] AI CFO: "Investment income needs to be offset. Refer to Article 33 of the Consolidation Standards and generate the offsetting entry: Debit: Investment income Credit: Long-term equity investment"

[0126] Virtual Bank Manager:

[0127] Simulate the loan approval process and require students to submit a financial feasibility report.

[0128] 2. Tax planning

[0129] Teaching content:

[0130] Core: interpretation of tax laws and policies (VAT / income tax), application for tax incentives, and cross-border tax arrangements.

[0131] Difficulties: transfer pricing and BEPS (base erosion and profit shifting) rules.

[0132] Teaching equipment:

[0133] Tax simulation platform: Automated simulation of tax declaration process (such as electronic tax bureau API connection).

[0134] Policy database: Integrates the State Administration of Taxation's regulatory database and AI pushes policy change warnings (NLP keyword monitoring).

[0135] Tax sandbox system: Virtually adjust corporate structure / transaction model and calculate tax burden changes in real time.

[0136] Game algorithm selection: constrained optimization + Monte Carlo simulation, calculate the optimal tax saving path within the tax law framework, and simulate the audit probability.

[0137] Virtual Characters:

[0138] AI Tax Inspector:

[0139] Responsibilities: Question students' planning plans and inquire about tax law provisions

[0140] Example conversation:

[0141] Student: "My group transfers profits to low-tax subsidiaries."

[0142] The AI ​​auditor stated: "This arrangement violates BEPS Actions 8-10. Please explain the economic substance and rationality. Otherwise, you will be required to pay additional taxes and penalty interest in accordance with the Special Tax Adjustment Measures!"

[0143] International Tax Consultants:

[0144] Provide guidance on using tax treaties to avoid double taxation.

[0145] 3. Audit Practice

[0146] Teaching content:

[0147] Core: Execution of audit procedures (confirmation / inventory), risk assessment, internal control testing, and audit report writing.

[0148] Difficulties: Big data auditing and blockchain evidence verification.

[0149] Teaching equipment:

[0150] AI working paper platform: Automatically extract financial data anomalies (Benford's law detects fraud).

[0151] Blockchain audit tracking system: Hyperledger records audit trails and tamper-proof operation logs.

[0152] AR remote inventory tool: Hololens 2 scans physical objects in the warehouse and automatically compares them with book inventory.

[0153] Game algorithm selection: Bayesian network + adversarial reinforcement learning, identifying signs of fraud, and dynamically generating audit evidence chain challenges.

[0154] Virtual Characters:

[0155] AI Audit Project Manager:

[0156] Responsibilities: Assign audit tasks and review the quality of working papers.

[0157] Example conversation:

[0158] Student: “The audited company refused to provide a list of suppliers.”

[0159] AI Project Manager: "According to Auditing Standard No. 1151, the audit scope should be considered limited, and a qualified opinion report should be considered."

[0160] Virtual Enterprise Finance:

[0161] Deliberately set up fraud traps (such as fictitious suppliers) to test trainees' professional skepticism.

[0162] 4. Cost Control

[0163] Teaching content:

[0164] Core: Activity-based costing (ABC), standard cost setting, cost-volume-profit analysis (CVP).

[0165] Difficulties: supply chain cost optimization and lean production accounting.

[0166] Teaching equipment:

[0167] IoT cost collection terminal: RFID tags collect production line material consumption data in real time.

[0168] Digital Twin Simulated Factory: Anylogic simulates production processes and dynamically optimizes cost drivers.

[0169] BI early warning dashboard: Power BI monitors cost deviations and triggers threshold alerts (SMS / email).

[0170] Game algorithm selection: Nash equilibrium + genetic algorithm, balance the quality / cost / delivery time triangle contradiction, and optimize resource allocation.

[0171] Virtual Characters:

[0172] AI Production Director:

[0173] Responsibilities: Require students to reduce costs while ensuring quality.

[0174] Example conversation:

[0175] Student: "We suggest cutting down on the quality inspection process to save 10% of the cost."

[0176] AI Production Director: "The lack of quality inspection will lead to an 8% increase in customer returns and increased overall losses. Please re-propose!"

[0177] Supply chain managers: Collaboratively optimize purchasing batches and warehousing costs.

[0178] Example 3

[0179] Based on the method described in Example 1, the differences between the scores of several teaching tasks and their corresponding preset ability scores are obtained by using a PID controller. When the differences are greater than a maximum threshold, the AR model is called to simulate teaching for the teaching task, specifically including:

[0180] 1. Hardware Configuration and Basic Deployment

[0181] Equipment selection:

[0182] AR glasses: Microsoft HoloLens 2 (gesture recognition + voice commands);

[0183] Mobile terminal: iPad Pro (LiDAR scanning spatial positioning);

[0184] Auxiliary equipment: Bluetooth bill scanner (OCR recognition), IoT seal (digital signature verification);

[0185] Environmental calibration: Paste AR positioning markers (QR codes or April Tags) in the classroom / training room to establish a spatial coordinate system:

[0186] 2. AR scene construction process

[0187] Scenario 1: Bill processing and voucher generation

[0188] Integration of virtual and real: Students scan physical invoices → AR glasses superimpose 3D highlight annotations (price and tax separation, invoice code verification rules);

[0189] Voucher generation: Drag the virtual invoice to the accounting voucher → automatically generate accounting entries:

[0190] Compliance verification: The AR system automatically flags anomalies (e.g., "serial numbered invoices suspected to be fraudulent") and pushes the terms of the "Invoice Management Measures";

[0191] Scenario 2: Inventory Count Audit

[0192] AR scanning warehouse:

[0193] Scan the physical shelf → overlay inventory labels (item name / batch / book quantity) on the virtual interface;

[0194] Virtual-to-real comparison: students count the physical quantity → input the actual number via voice → AR system calculates the discrepancy rate (>5% triggers a red alert);

[0195] Audit evidence collection: Gesture recording of inventory video → automatic generation of blockchain evidence (timestamp + geographic coordinates).

[0196] Scenario 3: Consolidated Financial Statements Practice

[0197] Holographic Data Visualization:

[0198] Expand the virtual group structure chart with a gesture → Click on a subsidiary to trigger a pop-up window with financial data (income / liabilities / equity);

[0199] Offsetting entry training: Drag and drop the parent company's "Long-term Equity Investment" and the subsidiary's "Owner's Equity" to collide - automatically generate offsetting entries:

[0200] Related-party transaction perspective:

[0201] AR glasses can display the internal transaction flow (such as commodity flow / capital flow) through perspective, and mark abnormal pricing in red.

[0202] Example 4

[0203] Based on the above embodiments, a dynamic game model is introduced in this embodiment. Through dynamic games between students and virtual characters, students can deeply understand accounting principles and internal control logic in simulated business decision conflicts. The specific game settings include:

[0204] 1. Clarify teaching objectives and accounting topics

[0205] Core objectives:

[0206] Based on the ability assessment results, targeted improvement of weak knowledge points is carried out to improve the overall ability score.

[0207] Understand the application and conflicts of accounting standards (e.g., revenue recognition, asset impairment) in real-world decision-making.

[0208] Understand the role of cost-benefit analysis in accounting choices (e.g., depreciation methods, inventory valuation).

[0209] Understand the importance of internal control and the risk points of breach.

[0210] Experience professional ethical dilemmas (such as earnings management, fraud pressure) and their consequences.

[0211] Understand the impact of financial reporting on stakeholder (investors, creditors) decisions.

[0212] Cultivate financial analysis and risk judgment capabilities.

[0213] Select a topic: Choose a topic that is highly competitive and likely to cause conflict:

[0214] Revenue recognition game: the temptation and risks of early / delayed revenue recognition.

[0215] Asset valuation and impairment game: the motivations and consequences of overestimating / underestimating asset values.

[0216] The game between cost allocation and expense capitalization: the choice of aggressive / conservative accounting policies.

[0217] Internal control game: the temptation of employees / management to break internal control and the pressure of audit / regulation.

[0218] Budget preparation and assessment game: budget relaxation, performance embellishment and conflict with assessment goals.

[0219] M&A valuation game: the game between buyers and sellers over the target company's financial data.

[0220] Audit game: the game between auditors and management on audit evidence and adjusting entries.

[0221] 2. Designing Specific Game Situations and Roles

[0222] In financial accounting, there are preset scenarios: disputes over the timing of revenue recognition and asset impairment judgments.

[0223] Scenario: Students play the role of a new CFO at a tech company. Near the end of the quarter, the sales team secures a major deal, but there's uncertainty about product delivery and customer acceptance. The sales director (a fictional character) strongly advocates for early revenue recognition to achieve quarterly targets and ensure a stable team bonus and stock price.

[0224] The board of directors (another fictional character) expects performance targets to be met;

[0225] Student role: CFO - Goals include compliance reporting, maintaining company reputation, and long-term growth, but faces short-term performance pressures and peer relations.

[0226] Fictional Persona 1: Sales Director – Clear goals: Maximize quarterly revenue and meet bonus targets. Behavior Patterns: Pressure, promises, emphasizes customer relationships, and downplays risks. May provide "reasonable" justifications (e.g., verbal promises).

[0227] Virtual Role 2 (Optional): Auditor - Intervene in later stages, question the basis for revenue recognition, and request adjustments. Behavior: Professional, adhere to standards, obtain evidence, and exert pressure for adjustments.

[0228] Virtual Role 3 (optional): Chairman of the Board / Investor - Pays attention to the quarterly report results, expresses satisfaction if "targets are met" and expresses disappointment if "targets are not met" (affects the student CFO's "reputation value").

[0229] III. Defining Game Rules and Core Mechanics

[0230] Round-based structure: rounds are divided according to key decision points (e.g., receipt of sales report - sales director lobbying - CFO decision - auditor review - financial report release - market reaction).

[0231] Student Decision Points:

[0232] Is revenue recognized in advance? (Yes / No);

[0233] How to ask the sales team for additional evidence? (Strong request / Weak request / No request)

[0234] How to respond to the auditor's inquiries? (Provide complete evidence / partial evidence / make an ambiguous response / refuse to make adjustments);

[0235] (Optional) Are risks disclosed proactively? (Detailed / Brief / Not disclosed in the notes to the financial statements)

[0236] Virtual character behavior logic:

[0237] Sales Director:

[0238] If the CFO refuses to confirm in advance: increase lobbying efforts (affect the “pressure value”), possibly providing fabricated / fuzzy evidence (students need to identify risks).

[0239] If the CFO demands strong evidence: they may cooperate (reduce risk) or shirk responsibility / fabricate information (increase subsequent risks).

[0240] Auditor:

[0241] Calculate an “audit risk level” based on the quality of evidence provided by the CFO and revenue recognition decisions.

[0242] High risk level: requires journal entry adjustments, expands the audit scope, and may trigger a regulatory investigation (with serious consequences).

[0243] Market / Board of Directors:

[0244] Based on the final financial report results (whether the standards are met, whether there are any adjustments, and disclosure transparency), give positive / neutral / negative reactions (affecting "stock price" and "reputation value").

[0245] Key state variables:

[0246] Compliance Risk: This measures the likelihood and severity of a violation of accounting standards. This is influenced by the confirmation decision, the strength of evidence, and the level of disclosure.

[0247] "Performance Pressure Value": reflects short-term pressure from sales, the board of directors, and the market, affecting the difficulty of decision-making.

[0248] "Reputation Value": A comprehensive reflection of the CFO's professional ethics and the company's market image. It is influenced by the final results, audit opinions, and market reactions.

[0249] “Audit Risk Level”: calculated by the auditor based on evidence and judgment.

[0250] Payoff function (quantitative / descriptive consequences):

[0251] Short-term benefits: Early recognition of revenue may bring "target-achieving bonuses", "short-term stock price increases", and "board of directors' appreciation".

[0252] Long-term risk / loss:

[0253] Audit adjustments lead to financial report restatements -> reputation value plummets, fines, and litigation risks.

[0254] Being investigated by a regulatory agency – resulting in substantial fines, criminal liability, and ruined reputation.

[0255] Investors lose confidence - stock prices fall for a long time and financing becomes difficult.

[0256] Internal control failure exposure - corporate governance rating downgraded.

[0257] Teaching core: Design rules so that compliant but potentially short-term "failed" choices will result in higher long-term comprehensive returns; radical choices may be tempting in the short term, but are extremely risky and can result in huge losses.

[0258] 4. Development of Technology Platform

[0259] Core elements of the interface:

[0260] Simulated financial system interface (simplified general ledger, sales contract, email).

[0261] Key financial indicator dashboard (revenue, profit, stock price, compliance risk value, reputation value).

[0262] A dialogue / email system with virtual characters (reflecting lobbying, inquiry, and pressure).

[0263] Quick link to the "Standards Library" (check the relevant revenue recognition standards ASC 606 / IFRS15 at any time).

[0264] Evidence viewing window (sales contract, delivery note, acceptance note, customer email - some parts may be blurred / missing / suspicious).

[0265] Backend logic:

[0266] Rule engine: strictly embeds the logic of relevant accounting standards (such as the five-step method for revenue recognition).

[0267] Virtual Character AI:

[0268] Sales Director: Trigger different lobbying strategies (emotion, interest, threat) based on goals, student decisions, and "pressure value".

[0269] Auditor: Generates questions and adjustment requirements based on preset risk models (sufficiency of evidence, transaction complexity, historical issues).

[0270] Consequence calculation model: quantify the impact of different decision paths on state variables (risk, reputation, stock price).

[0271] (Advanced) Fraud Detection Logic: If a student chooses an aggressive path and the evidence is questionable, the probability of being subsequently “discovered” by auditors / regulators is increased.

[0272] 5. Integrate teaching feedback and guidance mechanisms

[0273] Instant professional feedback:

[0274] After the decision: clearly show the quantitative impact of the decision on the state variables.

[0275] Virtual character professional response:

[0276] Auditor: "Based on contract clause X and customer email Y, we believe that recognizing revenue before acceptance does not comply with Revenue Recognition Standard Z. Please provide further evidence or make adjustments."

[0277] Sales Director: "Thank you for your support! Team morale is greatly boosted! This is a record of the customer's verbal commitment."

[0278] Example 6

[0279] Based on Example 5, after the virtual character and the student play a dynamic game, a teaching visualization analysis report is generated, which specifically includes:

[0280] After the conversation, an in-depth review report is generated, including:

[0281] 1. Establish multi-dimensional evaluation indicators

[0282] Standard application, accounting standards matching accuracy, policy selection rationality, by obtaining the rule engine verification records in dynamic games;

[0283] Risk control, compliance risk value fluctuations, crisis response speed, by obtaining dynamic dashboard historical data (risk heat map);

[0284] Ethical decision-making, frequency of short-term interest choices, rule-breaking rate, by obtaining game path records;

[0285] Practical efficiency, single-task time consumption, and resource allocation optimization are evaluated by obtaining operation logs.

[0286] 2. Obtaining a Student Ability Profile

[0287] Obtain student ability portraits and calculate indicators based on multi-dimensional evaluation indicators;

[0288] The calculation results are input into the student ability profile, and the student ability curve is predicted using the LSTM model and random forest algorithm;

[0289] Select the visual reporting method, enter the capability profile into the report template, and generate a visual analysis report.

[0290] The visual report templates specifically include: capability radar chart, risk control heat map and ethical decision-making trajectory.

[0291] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering technicians in this field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. An accounting teaching method based on a multidimensional task model, characterized by: include: Obtain students' initial ability assessment scores and operation logs through the data interface; The dynamic course configuration model dynamically adjusts the task matrix based on the initial competency assessment results and operation logs; A corresponding virtual character is generated according to the task matrix, and the virtual character executes a dynamic game algorithm to play a dynamic game with students to realize accounting teaching.

2. The accounting teaching method based on the multidimensional task model according to claim 1 is characterized in that: Before obtaining the student's initial ability assessment results through the data interface, the method further includes: Input the tasks to be taught into the teaching platform and construct a task matrix; Based on the task matrix, the corresponding ability assessment question bank is matched through the intelligent assessment model; Students answer questions based on the ability assessment question bank on the teaching platform; The intelligent assessment model generates students' initial ability assessment scores based on their answers.

3. The accounting teaching method based on the multidimensional task model according to claim 1 is characterized in that: The dynamic course configuration model dynamically adjusts the task matrix based on the student's initial ability assessment results, including: The intelligent assessment model obtains the student's initial ability assessment results and generates an ability profile based on the initial ability assessment results; The dynamic course configuration model calls the weight dynamic allocation model to dynamically adjust the weight of the teaching task based on the capability profile; Dynamically adjust the task matrix according to the weight of the teaching tasks.

4. The accounting teaching method based on the multidimensional task model according to claim 3 is characterized in that: The calculation formula of the weight dynamic allocation model is: Where W D,i，j is the weight of the j-th teaching task in the i-th adjustment, ability i is the comprehensive ability score after the i-th adjustment, W D,0，j is the initial weight of the i-th teaching task; α is the degree of tool dependence, C K，i，j Score the task of the jth teaching after the i-th adjustment, U J is the task urgency, R is the available resource score, γ is the time decay factor, β is the correct answer rate, C 0，j Preset the task score for the j-th teaching task, W D,0，j is the initial weight of the i-th teaching task, and n is the number of teaching tasks.

5. The accounting teaching method based on the multidimensional task model according to claim 3 is characterized in that: After the virtual role teaching model provides accounting teaching to students according to the teaching task, the virtual role teaching model further includes: Obtaining the adjusted task matrix, and generating a corresponding ability assessment question bank based on the task matrix through an intelligent assessment model; Students answer questions based on the ability assessment question bank through secondary operations on the teaching platform; The intelligent assessment model generates students’ secondary ability assessment scores based on their answers; The intelligent assessment model is evaluated based on the results of the secondary ability assessment.

6. The accounting teaching method based on the multidimensional task model according to claim 5 is characterized in that: The intelligent assessment model is evaluated based on the secondary ability assessment results, including: The intelligent assessment model obtains students’ comprehensive ability scores; If the comprehensive ability score meets the threshold, multiple teaching task scores are obtained; Determine whether the scores of multiple teaching tasks are consistent with the preset task scores; If the scores of multiple teaching tasks meet the preset task scores, the teaching will be exited; If the scores of multiple teaching tasks do not all meet the preset task scores, the teaching tasks that meet the preset task scores will be removed; The dynamic course configuration model dynamically adjusts the task matrix again based on the secondary ability assessment results and operation logs until the assessment results meet the threshold.

7. The accounting teaching method based on the multidimensional task model according to claim 6 is characterized in that: After obtaining the student's comprehensive ability score, the intelligent assessment model also includes: If the comprehensive ability score does not meet the threshold, the dynamic course configuration model will dynamically adjust the task matrix again based on the secondary ability assessment results and operation logs until the evaluation results meet the threshold.

8. The accounting teaching method based on the multidimensional task model according to claim 6 is characterized in that: If the scores of multiple teaching tasks do not all meet the preset task scores, the following also applies: Obtaining teaching tasks that do not meet the preset task scores; Obtaining the difference between the score of the teaching task and the score of the corresponding preset task through a PID controller; When the difference is greater than the maximum threshold, the AR model is called to simulate teaching of the teaching task.

9. The accounting teaching method based on the multidimensional task model according to claim 1 is characterized in that: The task matrix includes financial accounting, tax coordination, audit practice and cost control.

10. The accounting teaching method based on the multidimensional task model according to claim 3 is characterized in that: The capability profile includes identity ID, knowledge domain, real-time weight and tool dependency.

Images