Intelligent international education management method and system

Abstract

The application provides an intelligent international education management method and system, comprising: obtaining term interpretation text content and corresponding involved knowledge point list in each language version teaching resource, identifying interpretation difference degree and semantic deviation degree of the term in different teaching systems, and obtaining preliminary evaluation result of meaning deviation of the term between different languages; obtaining term application instances and context expression methods in student homework, extracting actual application scene of the term in the homework text and matching with the involved knowledge point list, combining the preliminary evaluation result of the meaning deviation, and determining the semantic deviation distance; according to the interpretation version of the divergence classification result and the unified semantic benchmark, dynamically updating the reference relationship of the term in the teaching resources between the chapters and the hierarchical structure between the knowledge points in the teaching materials, and obtaining the aligned knowledge framework.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to an intelligent international education management method and system. Background Technology

[0002] Intelligent international education management, as a core driver of global digital transformation in education, achieves core management functions such as unified transnational curriculum standards, monitoring of teaching quality, and mutual recognition of credits through multilingual teaching resources. As transnational educational cooperation projects expand, students from different countries have increasingly higher expectations for localized teaching content. International education management faces the dual challenge of achieving cultural adaptation while maintaining the consistency of knowledge content. In synchronous courses within transnational joint programs, when original English textbooks are translated into multiple languages ​​such as Chinese, German, and French, core terms in humanities and social sciences fields such as economics, law, and philosophy may have shifted meanings due to polysemy and cultural context differences. The same term may refer to different knowledge connotations in different languages. In international credit recognition assessment scenarios, translation differences in the same professional terms in course syllabi submitted by institutions from different countries can lead to disagreements among review experts regarding the course content, affecting the determination of credit equivalence. In international professional qualification certification examinations involving multiple countries, different language versions of the same test question can cause differences in understanding among candidates due to discrepancies in terminology interpretation, undermining the fairness of the assessment. More critically, existing solutions lack the ability to dynamically track the consistency of knowledge content, making it impossible for administrators to promptly identify and correct conceptual shifts arising during language translation, thus weakening the credibility of the entire management system. Take the concept of "value" in economics as an example: in English-language teaching systems, this term emphasizes price manifestation in market exchange, while in German-language systems it focuses more on intrinsic value judgments at the philosophical level, and in Chinese-language systems it distinguishes between the dual connotations of use value and exchange value. This shift in meaning not only affects the understanding of individual concepts but also creates a chain reaction within the knowledge network, leading to fundamental differences in the overall knowledge frameworks constructed by students from different countries, thus hindering the implementation of management functions such as credit recognition and degree conferral. Due to the lack of a precise mechanism for quantifying the degree of meaning shift and an assessment system for the magnitude of differences in knowledge systems, existing international education management technologies cannot establish a dynamic mapping relationship between the two. This lack of mapping prevents the system from adjusting the precision of resource allocation strategies based on actual deviations, resulting in uncontrollable diffusion of differences in knowledge transfer between different language versions of teaching resources. Therefore, how to construct a mapping refinement and adjustment mechanism based on the dynamic correlation between the degree of semantic deviation and the magnitude of differences in knowledge systems, and accurately identify and quantify conceptual deviations in multilingual teaching resources, has become a key issue in improving the level of international education management and achieving consistent control of cross-border teaching quality. Summary of the Invention

[0003] This invention provides an intelligent international education management method, comprising:

[0004] Obtain the terminology definition texts and corresponding knowledge point lists from teaching resources in various language versions, identify the degree of difference and semantic deviation in the interpretation of terms in different teaching systems, and obtain the evaluation results of the semantic shift of terms between different languages;

[0005] Obtain examples of terminology usage and contextual expressions from student assignments, extract application scenarios of terms from assignment texts and match them with the knowledge point list, and determine semantic offset distance by combining the semantic offset evaluation results;

[0006] Based on the semantic offset distance, identify the term definition text types that need adjustment from the meaning offset evaluation results, map the term definition texts with semantic offset distances exceeding the threshold to the standard knowledge connotation description, and obtain the definition version of the unified semantic benchmark.

[0007] By comparing the error types and frequencies of terminology used in student assignments in different language versions, the sources of comprehension discrepancies between students' knowledge systems in different countries are identified, and the discrepancy classification results are obtained.

[0008] Based on the divergence classification results and the interpretation version of the unified semantic benchmark, the reference relationships of terms in the teaching resources between textbook chapters and the hierarchical structure between knowledge points are updated to obtain the aligned knowledge framework.

[0009] The semantic expressions of unified terms are extracted from the aligned knowledge framework and applied to the cultural background descriptions and case analyses of teaching resources in different countries, generating a multilingual knowledge alignment scheme that includes a terminology comparison table, unified definition standards, and cross-cultural teaching suggestions.

[0010] Furthermore, the process of obtaining terminology definition texts and corresponding knowledge point lists from teaching resources in various language versions, identifying the degree of difference and semantic deviation in the definitions of terms in different teaching systems, and obtaining the evaluation results of the semantic shift of terms between different languages ​​includes:

[0011] The terminology definition texts are extracted from teaching resources in various languages. The chapter position information and professional field classification labels of the terms in the textbooks are obtained. The semantic distance values ​​of the definition texts between different language versions are calculated to obtain the quantitative results of the degree of difference in terminology definitions.

[0012] Based on the quantitative results of the degree of difference in the interpretation, the conceptual hierarchy of terms in different teaching systems is identified, and it is determined whether the terms belong to the basic concept layer, the application method layer, or the theoretical principle layer, thus forming a hierarchical classification of the knowledge points of the terms.

[0013] Using the hierarchical classification of knowledge points, cultural context factors and application scenario descriptions of terms in teaching cases are extracted, semantic feature vectors of terms are constructed, the distance between the semantic feature vectors and standard definition vectors is calculated, the degree of semantic deviation is obtained, and the evaluation results of the meaning shift of terms in different languages ​​are generated.

[0014] Furthermore, the process of obtaining examples of terminology usage and contextual expressions in student assignments, extracting application scenarios of terms from the assignment text and matching them with the knowledge point list, and determining the semantic offset distance based on the semantic offset evaluation results, includes:

[0015] Extract examples of terminology usage from student assignment texts, identify the location and frequency of terminology in the problem-solving process, obtain the context of the vocabulary units before and after the terminology, and extract the application context of the terminology.

[0016] Based on the co-occurrence relationship between the term and other concepts in the application context, the semantic association strength value between the term and adjacent concepts is calculated to obtain the context association degree value;

[0017] Using the aforementioned contextual relevance value, the application scenario types of terms in different question types are identified. The differences between the application scenario types and the standard application scenarios defined in the textbook are compared. The number of editing operations between the student's answer step sequence and the standard problem-solving step sequence is calculated to obtain a quantitative value of the degree of comprehension deviation. Combined with the meaning shift evaluation results, the semantic shift distance is determined.

[0018] Furthermore, based on the semantic offset distance, the step of identifying the term definition text types that need adjustment from the meaning offset evaluation results, and mapping term definition texts with semantic offset distances exceeding a threshold to standard knowledge connotation descriptions, to obtain a definition version with a unified semantic benchmark, includes:

[0019] Based on the semantic offset distance, extract the text type markers of the terms' definitions, identify basic concept types, application method types, or theoretical principle types, retrieve the standardized descriptions of the terms in standard documents, and obtain the core semantic units and logical relation words.

[0020] Using the core semantic units and logical relation words, the semantic components in the original interpretation text are identified, expressions with a deviation less than the threshold are retained, and semantic components with a deviation exceeding the threshold are replaced to generate a interpretation version with a unified semantic benchmark.

[0021] Furthermore, the standardized description of the search terms in standard documents, obtaining core semantic units and logical relation words, includes:

[0022] Extract standardized descriptions of terms from standard documents and identify core semantic units, logical relation words, and qualifying modifiers in the standardized descriptions.

[0023] Based on the core semantic units and logical relation words, a standard semantic expression structure for terms is constructed, the correspondence between terms and standard structures is recorded, a cross-language term mapping table is formed, and a unified code consisting of subject codes and concept levels is assigned.

[0024] Furthermore, by comparing the types and frequencies of errors in the use of terminology in different language versions of student assignments, the sources of comprehension discrepancies between students' knowledge systems from different countries are identified, resulting in a discrepancy classification, including:

[0025] Extract terminology usage error types and error frequencies from student assignments, identify error pattern categories of terms in different language versions, and calculate the correspondence between error pattern categories and conceptual confusion markers;

[0026] Based on the error pattern categories, the analysis reveals that the errors originate from semantic differences in terminology definition texts or differences in the knowledge structure of teaching systems, generating divergence classification results.

[0027] Furthermore, based on the error pattern category, the analysis of whether the errors originate from semantic differences in terminology definition texts or knowledge structure differences in teaching systems, generating divergence classification results, includes:

[0028] Extract the manifestations of terminological misuse and the location of errors in the problem-solving process from the error pattern categories, and obtain data on students' understanding of the terms and the original expressions in the textbook;

[0029] Calculate the matching degree between the understanding perspective data and the original expression method, analyze the correlation coefficient between errors and semantic differences or knowledge structure differences in terminology interpretation texts, and determine whether the discrepancies belong to semantic meaning shifts or cognitive cultural understandings.

[0030] Furthermore, based on the divergence classification results and the interpretation version of the unified semantic benchmark, the reference relationships of terms in the teaching resources between textbook chapters and the hierarchical structure between knowledge points are updated to obtain the aligned knowledge framework, including:

[0031] Based on the aforementioned divergence classification results, identify the list of terminology reference relationships, scan the mutual references of terms in the textbook, and adjust the priority order of terms.

[0032] Based on the interpretation of the unified semantic benchmark, the hierarchical structure of knowledge points is rearranged, a directed acyclic graph between terms is constructed, and an aligned knowledge framework is generated.

[0033] Furthermore, the semantic expression of unified terminology extracted from the aligned knowledge framework is applied to the cultural background explanations and case analyses of teaching resources from different countries, generating a multilingual knowledge alignment scheme that includes a terminology comparison table, unified definition standards, and cross-cultural teaching suggestions, including:

[0034] Extract the semantic expression of unified terms from the aligned knowledge framework, obtain the corresponding forms of the semantic expression in each language version, and establish a correspondence table between terms and local cases;

[0035] Compile a terminology lookup table containing the original terms, standard definitions, and cultural annotations, apply it to case analysis, and generate a multilingual knowledge alignment scheme.

[0036] On the other hand, the present invention also provides an intelligent international education management system, the system comprising:

[0037] The terminology definition difference identification module is used to obtain the terminology definition text content and corresponding knowledge point list in teaching resources of various language versions, identify the degree of difference in the definition of terms and the degree of semantic deviation of terms in different teaching systems, and obtain the evaluation results of the meaning shift of terms between different languages.

[0038] The semantic offset distance determination module is used to obtain examples of term usage and contextual expressions in student assignments, extract the application scenarios of terms in the assignment text and match them with the knowledge point list, and determine the semantic offset distance in combination with the semantic offset evaluation results.

[0039] The unified semantic benchmark generation module is used to identify the types of term definition texts that need to be adjusted from the meaning offset evaluation results based on the semantic offset distance, and map the term definition texts with semantic offset distances exceeding the threshold to the standard knowledge connotation descriptions to obtain the definition version of the unified semantic benchmark.

[0040] The Understanding Disagreement Assessment module is used to identify the sources of understanding disagreements between students' knowledge systems from different countries by comparing the types and frequencies of errors in the use of terminology in different language versions of student assignments, and to obtain the disagreement classification results.

[0041] The knowledge framework alignment module is used to update the reference relationships between terms in the teaching resources and the hierarchical structure between knowledge points in the textbook chapters based on the divergence classification results and the interpretation version of the unified semantic benchmark, so as to obtain the aligned knowledge framework.

[0042] A multilingual knowledge alignment scheme generation module is used to extract the semantic expression of unified terms from the aligned knowledge framework, apply it to the cultural background explanation and case analysis of teaching resources from different countries, and generate a multilingual knowledge alignment scheme that includes a terminology comparison table, unified definition standards, and cross-cultural teaching suggestions. The technical solution provided by this invention can include the following beneficial effects:

[0043] This invention discloses an intelligent international education management method and system. It extracts terminology definitions and knowledge point lists from multilingual teaching resources, assesses differences in definitions and semantic shifts between different languages, integrates terminology application examples and contextual matching in student assignments, calculates semantic shift distances, and filters out texts requiring adjustment to map to a unified semantic benchmark, forming standardized definition versions. Furthermore, it compares the types and frequencies of errors in student assignments from different countries, traces the source of discrepancies to understand whether they stem from semantic shifts, and categorizes the discrepancies. Based on this, it dynamically updates textbook analysis, user requests (which appear to be about handling terminology issues in multilingual teaching resources, involving definition differences, student assignment analysis, and knowledge framework alignment), citation relationships, and knowledge levels, constructs an aligned knowledge framework, and ultimately generates a multilingual alignment scheme containing a terminology comparison table, definition standards, and cross-cultural suggestions. This achieves semantic consistency of terminology and global unification of teaching resources, improving the accuracy of cross-cultural education and the efficiency of student knowledge transfer. Attached Figure Description

[0044] Figure 1 This is a flowchart of an intelligent international education management method according to the present invention.

[0045] Figure 2 This is a schematic diagram of the structure of an intelligent international education management system according to the present invention. Detailed Implementation

[0046] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0047] like Figure 1-2 This embodiment of an intelligent international education management method and system may specifically include:

[0048] Step S101: Obtain the terminology definition text content and the corresponding knowledge point list from the teaching resources of each language version, identify the degree of difference in the definition of terms and the degree of semantic deviation in different teaching systems, and obtain the preliminary assessment results of the meaning shift of terms in different languages.

[0049] The terminology definition texts are obtained from teaching resources in various languages. The chapter position information and professional domain classification tags of the terms in the textbooks are extracted. The semantic distance between the definition texts in different language versions is calculated using cosine similarity to obtain a quantitative result of the degree of difference in terminology definitions. Based on this quantitative result, the conceptual hierarchy of terms in different teaching system categories is identified, determining whether a term belongs to the basic concept layer, application method layer, or theoretical principle layer, thus forming a knowledge point hierarchical classification of the term. Using this knowledge point hierarchical classification, the cultural context factors and application scenario descriptions of the terms in actual teaching cases are obtained, constructing semantic feature vectors for the terms. The Euclidean distance between the semantic feature vectors and the standard definition vectors in different cultural backgrounds is calculated to obtain a semantic deviation value. If the semantic deviation value exceeds a first preset threshold, the term is marked as a high-deviation term; if the value is between the first and second preset thresholds, it is classified as a medium-deviation term. Based on the classification tags of high-biased and medium-biased terms, subject-verb-object grammatical structures and modifiers are extracted from the explanatory texts of high-biased terms, and keyword phrases and semantic relationships are extracted from the explanatory texts of medium-biased terms. The differences in the correspondence of the grammatical structures and keyword phrases in different language versions are compared to determine the preliminary assessment results of the semantic shift of terms in various languages.

[0050] For example, in one implementation, when obtaining terminology definitions from teaching resources in various languages, for the economic term "supply and demand equilibrium," the English version emphasizes quantity balance under the market price mechanism, the German version focuses on the social coordination of production distribution, and the Chinese version includes the dual meaning of planned regulation and market regulation. Through word segmentation and word vectorization, the definitions in each language are converted into high-dimensional vector representations. Specifically, firstly, language-specific word segmentation tools (such as NLTK for English and jieba for Chinese) are applied to the definitions in each language to segment them, removing stop words to obtain a core vocabulary set. Then, a pre-trained multilingual word vector model (such as MultilingualBERT or XLM-RoBERTa) is used to map each word to a 300-dimensional vector representation, and a weighted average (TF-IDF weights) is used to generate the vector representation of the entire definition text. For specialized terms, the system introduces domain-specific word vectors for weighted enhancement to improve representation accuracy. Finally, the cosine similarity formula cos(θ) = A·B / (|A||B|) is used to calculate the semantic distance value, where A and B are the definition vectors for different language versions. To handle differences in expression habits between languages, language-specific correction coefficients are introduced, such as a coefficient of 1.05 for English-German comparison and a coefficient of 1.12 for English-Chinese comparison, to compensate for the natural distance caused by differences in language structure. This optimized similarity calculation method can more accurately reflect the actual semantic distance between term definitions in different languages.

[0051] Specifically, when constructing semantic feature vectors, five dimensions of features are extracted from actual teaching cases of terminology: the core elements of the concept definition, the specific description of the application scenario, the frequency of citation of relevant theories, the focus of case analysis, and the examination angle of the after-class exercises. By applying 0-1 normalization to each dimension, a 5-dimensional vector space is constructed. Principal component analysis (PCA) is used to extract the first three principal components, explaining more than 85% of the total variance, forming the dimensionality-reduced semantic feature vector. To handle differences in expression across languages, the system also introduces a language-specific correction matrix to linearly transform the feature vectors, eliminating systematic biases caused by differences in language structure.

[0052] For example, the term "market failure" in US textbooks emphasizes externalities and information asymmetry, in European textbooks it highlights the problem of public goods provision, and in Asian textbooks it focuses on the necessity of government intervention. Principal component analysis is used to extract the main components of these features, construct a standardized semantic feature vector, and calculate the Euclidean distance between this vector and the preset standard interpretation vector.

[0053] In one possible implementation, when extracting grammatical structure from the explanatory text of high-skew terms, the basic components of subject, predicate, and object, as well as modifying and limiting components such as attributive, adverbial, and complement, are identified. Taking the legal term "due process" as an example, the English version's subject-predicate structure emphasizes the fairness and transparency of the procedure, while the Chinese version includes the dual requirements of substantive justice and procedural justice, and the German version focuses more on the legality review of administrative actions. Through dependency parsing, the core semantic relations in each language version are extracted, including verb-object relations, subject-predicate relations, and modification relations, and the degree of correspondence of these relations between different languages ​​is calculated.

[0054] It should be noted that when comparing the differences in grammatical structure and keyword phrases, a semantic dependency graph of term definitions is constructed, where nodes are semantic units and edges represent the strength of dependency relations. The input is the definition sentence, which is used to generate the graph through dependency parsing. A subgraph isomorphism algorithm is used to calculate the structural similarity of the graphs in different language versions. The input is two graphs, and the output is the similarity value. When the similarity is below a threshold, it is marked as a term that needs to be aligned, forming a preliminary assessment result of semantic shift.

[0055] Step S102: Obtain examples of terminology usage and contextual expressions in student assignments, extract the actual application scenarios of terms in the assignment text and match them with the list of relevant knowledge points, and determine the semantic offset distance by combining the preliminary evaluation results of semantic offset.

[0056] This study extracts examples of terminology usage from student assignments, identifies the location and frequency of terminology in the problem-solving process, and obtains the contextual content of a preset number of vocabulary units before and after each term using a text window method to extract its specific application context. Based on the co-occurrence relationship between the term and other concepts in the application context, the semantic association strength value between the term and adjacent concepts is calculated to obtain the term's contextual association degree value. Using this contextual association degree value, the actual application scenario types of the term in economic calculation questions, case analysis questions, and essay questions are identified, and the differences between these actual application scenario types and the standard application scenarios defined in the textbook are compared. By calculating the minimum number of editing operations between the student's answer sequence and the standard problem-solving sequence, including the cumulative number of insertion, deletion, and replacement operations, a quantitative value of the degree of comprehension deviation is obtained. Based on this quantitative value of the degree of comprehension deviation, the error pattern categories of terminology use in the assignments are identified, including concept substitution errors, logical order errors, and semantic scope errors, and a correspondence between error pattern categories and concept confusion markers is constructed. If the frequency of a certain error pattern category exceeds a preset threshold, the term is marked as a term with deviation. The semantic offset distance is determined by combining the semantic distance value in the preliminary assessment of semantic offset with the term tags that have deviations, using a weighted summation method.

[0057] For example, in one implementation, when extracting instances of terminology usage from student assignment text, an adaptive sliding window mechanism is used to traverse the entire assignment text. Unlike traditional fixed windows, the adaptive window dynamically adjusts its size according to the syntactic structure: for simple sentence structures, the window size is set to 5 words before and after; for complex sentence structures, the window expands to the entire subordinate clause; for paragraph-level logical structures, the window covers the entire argument unit.

[0058] Specifically, when the target term "market equilibrium" is identified as appearing in a student's economics assignment, 10 lexical units before and after the term are extracted as context windows to record the co-occurrence of the term with related concepts such as "supply curve," "demand curve," and "price mechanism." By calculating the point mutual information value between the term and each co-occurring concept, the semantic association strength is quantified, where a higher association strength value indicates that the student's understanding of the term in context is more accurate.

[0059] It should be noted that the calculation of the number of editing operations uses an improved Levenshtein distance algorithm, optimized for the characteristics of the teaching domain. Specifically, the sequence of student answer steps is represented as an ordered list S1, and the sequence of standard solution steps is represented as an ordered list S2. A dynamic programming matrix D of (m+1)×(n+1) is constructed, where m and n are the lengths of S1 and S2, respectively. Initialize D[0,0]=0, D[i,0]=i, D[0,j]=j, representing the number of operations to convert an empty sequence into the target sequence. For the case where i>0, j>0, D[i,j]=min(D[i-1,j]+1,D[i,j-1]+1,D[i-1,j-1]+cost), where cost is the cost of the replacement operation. When S1[i]=S2[j], cost=0; otherwise, cost=the replacement weight. Unlike the standard Levenshtein algorithm, this system assigns differentiated weights to different types of editing operations: replacement operations involving conceptual errors have a weight of 2.0, indicating a serious deviation in understanding; operations involving reversing the order of steps have a weight of 1.5, indicating a misunderstanding of the process; and insertion or deletion operations involving missing or redundant steps have a weight of 1.0, indicating an incomplete solution process. By backtracking to the optimal path, the system can accurately identify the specific types and locations of errors in the student's problem-solving process, providing detailed evidence for subsequent analysis of comprehension deviations.

[0060] Understandably, this multi-dimensional offset distance quantification method can accurately identify the differences in understanding that arise from different language versions of teaching resources in actual teaching applications, providing a quantitative basis for subsequent knowledge alignment and teaching resource optimization.

[0061] Step S103: Based on the semantic offset distance, identify the term definition text types that need adjustment from the preliminary assessment results of semantic offset, and map the term definition texts with semantic offset distances exceeding a preset threshold to the standard knowledge connotation description to obtain the definition version of the unified semantic benchmark.

[0062] Based on the semantic offset distance values, the terminology's explanatory text type tags are extracted from the preliminary assessment results of semantic offset. These tags include basic concept type, application method type, and theoretical principle type. The semantic offset distance corresponding to these explanatory text type tags is compared with a preset threshold. If the semantic offset distance exceeds the threshold, the standardized description of the term in international educational standard documents is retrieved, and core semantic units, logical relation words, and limiting modifiers are extracted to obtain the standard semantic expression structure of the term. Using this standard semantic expression structure as a template, the semantic components in the original explanatory text corresponding to the standard structure are identified, including three types of elements: subject concept, attribute features, and relation description. Based on the degree of semantic component offset, the original expressions with offset less than the secondary threshold are retained, while semantic components with offset exceeding the secondary threshold are replaced with standard expressions. The retained and replaced semantic components are recombined and arranged in the order of subject concept, attribute features, and relation description to form semantically aligned explanatory content. Based on the semantically aligned explanatory content, a cross-language terminology mapping table is established to record the correspondence between the original explanatory text, adjustment process tags, and the final explanatory version. Assign a unified code consisting of a subject code, concept level, and version number to each adjusted term, and mark the definition as the baseline version to obtain the unified semantic baseline definition version.

[0063] For example, in one implementation, when classifying terms based on semantic offset distance values, the system first reads the type label of each term from the preliminary semantic offset assessment results. Basic conceptual terms, such as fundamental economic concepts like "supply" and "demand," have an offset threshold set at 0.3; applied methodological terms, such as statistical methods like "regression analysis" and "hypothesis testing," have a threshold set at 0.5; and theoretical principle-based terms, such as higher-order theories like "Pareto optimality" and "Nash equilibrium," have a threshold set at 0.7. When the semantic offset distance of a term exceeds the threshold for its corresponding type, the system automatically triggers a standardization process.

[0064] Specifically, the process of acquiring a standard semantic expression structure involves three levels of information extraction. Core semantic units refer to the essential conceptual elements in a term's definition; for example, the core semantic units of "market failure" include "resource allocation" and "efficiency loss." Logical relation words reflect the causal, conditional, and adversative relationships between concepts, such as "leads to," "presupposes," and "however." Qualifying modifiers define the scope and conditions of a term's application, such as "under perfect competition" and "in the short term." These three types of elements together constitute the standard semantic framework of a term, providing a template for subsequent semantic alignment.

[0065] It should be noted that the identification and replacement of semantic components follows a hierarchical processing principle. Using a standard semantic expression structure as a template, the system identifies semantic components in the original explanatory text that correspond to the standard structure using a deep learning model. Specifically, a BERT-based sequence labeling model is first used to label the explanatory text with semantic roles, identifying three core elements: Subject, Attribute, and Relation. The Subject is typically the subject part of the definition, representing the object being defined; Attributes include adjectives, participles, and noun phrases describing the characteristics of the concept in the definition; Relations are verb phrases and prepositional structures connecting the Subject to other concepts. The system uses an attention-enhanced BiLSTM-CRF model, achieving an accuracy of over 92%. The identified semantic components are aligned with the standard template, and the semantic similarity of each component is calculated to obtain a quantified value of the offset. For components with a deviation of less than 0.3, the original expression is retained to preserve cultural characteristics; for components with a deviation between 0.3 and 0.7, partial adjustments are made, retaining the core meaning while standardizing the expression; for components with a deviation exceeding 0.7, they are completely replaced with standard expressions to ensure consistency of core semantics. This hierarchical semantic component processing strategy ensures the uniformity of the core meaning of terms while respecting the expression habits of different languages.

[0066] For example, while the main concept of "opportunity cost" remains consistent across different language versions, the descriptions of its attributes, such as "explicit vs. implicit," differ, and the emphasis on the relationship description, "difference from accounting costs," varies across versions.

[0067] Preferably, the unified coding adopts a three-segment structure: the subject code adopts a four-digit number from the International Standard Classification of Educational Sciences, such as "0311" for economics; the concept level is represented by two digits, "01" for basic concepts, "02" for applied methods, and "03" for theoretical principles; the version number records the revision history to ensure the traceability of the evolution of the interpretation.

[0068] In one embodiment, the cross-language terminology mapping table is stored in a relational database structure, including original terminology fields, language type fields, adjustment tag fields, standard definition fields, and update timestamp fields, to achieve unified management and dynamic maintenance of terminology definitions in multilingual teaching resources.

[0069] The preliminary assessment results of semantic shift are used to obtain the content of the terminology's definition text in various language versions and the chapter position of the term in the textbook. The professional field classification of the term and the hierarchical affiliation of the term in the curriculum system are collected. The differences in the definition boundaries of terms whose semantic shift distance exceeds the standard range in different teaching systems are analyzed. The definition text of the term is assessed to determine whether it belongs to the category of basic concepts, application methods, or theoretical principles. The knowledge category and teaching module to which the definition text of the term that needs to be uniformly expressed are determined.

[0070] From the preliminary assessment results of semantic shift, the chapter position markers and page indexes of terms in textbooks of various language versions are extracted, and the chapter number, unit theme, and list of related terms in which the term first appears are obtained. The importance score of the term is calculated by statistically analyzing the frequency of its occurrence in different chapters and the number of times it is cited by other terms. If the score is higher than a preset threshold, it is judged as a core concept; otherwise, it is considered an auxiliary concept, thus obtaining the textbook positioning characteristics of the term. Using these textbook positioning characteristics, the subject classification information corresponding to the term is extracted from the teaching syllabus, including codes at three levels: the subject category, the major category, and the course module. Based on the distribution ratio of the term in undergraduate basic courses, professional core courses, and postgraduate advanced courses, the course coverage of the term is calculated, determining the course level attribute of the term in the teaching system. Based on the course level attribute, a set of terms whose semantic shift distance exceeds the standard range is filtered, and the definition text of each term in the term set in different teaching systems is extracted. The number of elements of the conceptual connotation, the number of limiting words of the extension range, and the number of constraints of the applicable conditions in the definition text are compared to calculate the definition boundary difference value. If the difference value exceeds a preset threshold, the conceptual complexity and cross-system difference of the term are assessed. The complexity value is used to determine whether a term belongs to the category of basic concepts, applied methods, or theoretical principles. A complexity below a first threshold is classified as a basic concept, between the first and second thresholds as an applied method, and above the second threshold as a theoretical principle. Combining the classification results with cross-system differences, the knowledge category and teaching module affiliation of the terminology definitions requiring standardized expression are determined.

[0071] For example, in one implementation, when extracting term location information from the preliminary assessment results of meaning shift, the system scans the digital textbooks in each language version and records the complete location data of each term.

[0072] Specifically, the chapter location markers include a three-level index: the first level is the chapter number, the second level is the section number, and the third level is the paragraph number. The page number index records the specific location of terms on the physical page, facilitating quick location by teachers. The list of related terms is constructed by identifying professional terms appearing in the same or adjacent paragraphs, forming a local semantic network of terms.

[0073] It should be noted that the importance score is calculated using a composite index system. Frequency of occurrence is counted based on the number of times the term appears throughout the entire textbook, with 1 point awarded for each occurrence. Cited references are counted based on the number of times the term is mentioned in other term definitions, with 2 points awarded for each citation. Chapter coverage is calculated based on the number of chapters covered by the term, with 3 points awarded for each chapter covered. The three scores are weighted, summed, and normalized to obtain an importance score between 0 and 100. A score exceeding 60 is considered a core concept; between 30 and 60 is an important concept; and below 30 is a secondary concept.

[0074] For example, "supply and demand equilibrium" appears 87 times in economics textbooks, is cited by 32 other terms, covers 12 chapters, and scores 78.5 after calculation, making it a core concept.

[0075] Preferably, the extraction of subject classification information follows the International Standard Classification of Educational Subjects (ISCLC). Subject categories use two-digit codes, such as "03" representing social sciences; professional categories use four-digit codes, such as "0311" representing economics; and course modules use six-digit codes, such as "031101" representing microeconomics. The classification codes corresponding to terms are automatically extracted by parsing the metadata of the teaching syllabus. Course coverage calculation considers the distribution of terms across different course levels: if a term appears in undergraduate foundation courses, its weight is 0.3; if it appears in core professional courses, its weight is 0.5; and if it appears in postgraduate advanced courses, its weight is 0.2. Based on the weighted coverage ratio, the course level attribute of the term is determined, and terms with a coverage ratio higher than 0.7 are marked as cross-level core terms.

[0076] For example, the calculation process for the boundary difference value includes a three-dimensional quantitative analysis. The number of conceptual connotation elements is calculated by extracting noun phrases from the definition text. For instance, the English definition of "market failure" contains three core elements: "inefficiency," "allocation," and "resources," while the German definition contains four elements: "Marktversagen," "Allokation," "Effizienz," and "Ressourcen," with a difference of 1 in the number of elements. The number of denotation qualifiers is counted, including adjectives and qualifiers such as "completely," "partially," and "ideally." The number of conditional clauses and presuppositions in the definition is counted, including applicable constraints. The difference values ​​of the three dimensions are standardized and summed to obtain the comprehensive boundary difference value. When the difference value exceeds a preset threshold of 0.6, the term is marked as a high-difference term requiring focused adjustment.

[0077] In one possible implementation, the assessment of concept complexity comprehensively considers both the structural complexity and cognitive load of the term definition. Structural complexity is derived by calculating the syntactic tree depth and number of branches in the definition text; cognitive load is assessed by statistically analyzing the number of prerequisite knowledge points involved in the definition. The average of the two normalized indicators is taken as the concept complexity value. Two thresholds are set for classification: a first threshold of 0.3 and a second threshold of 0.7. Terms with a complexity value below 0.3 are classified as basic concepts, typically the fundamental building blocks of a discipline, such as "demand," "supply," and "price." Terms between 0.3 and 0.7 are classified as applied methods, including various analytical tools and methodologies, such as "marginal analysis," "elasticity calculation," and "regression models." Terms with a complexity value above 0.7 are classified as theoretical principles, encompassing high-level theoretical frameworks and complex models, such as "general equilibrium theory," "game theory Nash equilibrium," and "information asymmetry theory."

[0078] Specifically, the calculation of cross-system dissimilarity focuses on the degree of difference in the expression of the same term in different teaching systems. A preliminary dissimilarity is obtained by extracting semantic feature vectors of term definitions from each teaching system and calculating the cosine distance between these vectors. Then, cultural contextual factors are adjusted, such as an adjustment coefficient of 1.2 between the European / American and Asian systems, and 0.8 between different countries within the same cultural sphere, to obtain the final cross-system dissimilarity value.

[0079] Understandably, the determination of knowledge categories and teaching module affiliations employs a dual mapping mechanism. First, the terminology's classification results are mapped to a predefined knowledge category system, including three main categories: concepts and principles, methods and tools, and application cases. Then, the terms are mapped to specific teaching modules based on their position within the curriculum, such as introductory modules, core theory modules, practical application modules, and cutting-edge extension modules. This dual mapping achieves precise positioning of terms within both the knowledge and teaching systems, providing a clear classification basis for subsequent standardization.

[0080] Step S104: By comparing the error types and frequency of terminology used in student assignments in different language versions, the source and severity of the comprehension differences between students' knowledge systems in different countries are identified, and it is assessed whether the differences stem from semantic shifts, thus obtaining the difference classification results.

[0081] From the preliminary assessment results of semantic shift, the chapter position markers of terms in textbooks of various language versions are extracted, including chapter number, section number, and paragraph number. Page indexes and lists of related terms are recorded. Importance scores are calculated by statistically analyzing the frequency of occurrence, number of citations, and breadth of chapter distribution of terms. Terms with scores exceeding a first preset threshold are identified as core concepts, while those below a second preset threshold are identified as auxiliary concepts, thus obtaining the textbook positioning characteristics of the terms. Using these textbook positioning characteristics, the corresponding subject category code, professional category code, and course module code are extracted from the teaching syllabus. Based on the distribution ratio of terms in undergraduate foundation courses, professional core courses, and postgraduate advanced courses, a weighted course coverage is calculated. Terms with coverage exceeding a preset threshold are marked as cross-level core terms, determining the course level attribute of the terms. Based on the course level attribute, a set of terms whose semantic shift distance exceeds the standard range is filtered, and the number of conceptual connotation elements, extensional scope limiting words, and applicable condition constraints of the terms in the definition texts of different teaching systems are extracted. The difference values ​​of these three dimensions are calculated and standardized. When the comprehensive difference value exceeds a preset threshold, the conceptual complexity and cross-system difference of the term are assessed. The terminology category is determined by the complexity value of the concepts. Terms with complexity below a first threshold are categorized as basic concepts; those between the first and second thresholds are categorized as application methods; and those above the second threshold are categorized as theoretical principles. Combining the terminology category with cross-system differences, this is mapped to a predefined knowledge category system and specific teaching modules to determine the knowledge category and teaching module affiliation of the terminology definitions that require unified expression.

[0082] For example, in one implementation, when extracting term location information from the preliminary assessment results of semantic shift, the system first establishes a structured indexing system for the textbook. The chapter location markers employ a three-level coding structure, with the chapter number at the top, identifying the main knowledge unit where the term resides; the section number at the second level, detailing specific knowledge point categories; and the paragraph number precisely pinpointing the exact text location where the term first appears. A list of related terms is constructed by scanning professional vocabulary within a fixed window range before and after the term's appearance, forming a local semantic association network that reflects the knowledge association structure of the term within the textbook.

[0083] It should be noted that the importance score is calculated using a three-dimensional weighted evaluation mechanism. The frequency dimension counts the total number of times the term appears throughout the entire textbook, assigning a base score for each occurrence, reflecting the term's usage density. The citation count dimension counts the number of times the term is cited in other term definitions or explanations, assigning a higher weight score to reflect the term's fundamental status. The chapter distribution breadth dimension calculates the number of independent chapters covered by the term, assigning the highest weight score to reflect the term's knowledge coverage. After normalizing the scores from the three dimensions, a weighted sum is obtained to obtain the comprehensive importance score. When this score exceeds a first preset threshold, the system marks the term as a core concept, indicating its fundamental status in the teaching system; when the score is below a second preset threshold, it is marked as an auxiliary concept, mainly serving a supplementary explanation function; terms between the two thresholds are classified as important concepts, playing a key role in specific knowledge modules.

[0084] Specifically, the extraction of subject classification codes follows the hierarchical system of the International Standard Classification of Educational Sciences (ISCC). Subject category codes use two-digit identifiers, covering broad categories such as natural sciences, social sciences, and humanities; professional category codes extend to four digits, subdividing into specific subject areas, such as economics, management, and law; course module codes further extend to six digits, corresponding to specific course units, such as microeconomics, macroeconomics, and econometrics. The system automatically extracts the complete code chains corresponding to terms by parsing the metadata structure of the teaching syllabus, establishing a mapping relationship between terms and the teaching system.

[0085] Preferably, the calculation of course coverage comprehensively considers the distribution characteristics of terms across different teaching levels. Undergraduate foundation courses are assigned a basic weight value to reflect the popularity of the terms; core professional courses are assigned a medium weight value to reflect the professional importance of the terms; and advanced postgraduate courses are assigned a low weight value to represent the in-depth application of the terms. The comprehensive course coverage index is obtained by summing the products of the proportion of occurrence of the terms in each level of courses and their corresponding weights.

[0086] In one possible implementation, the calculation of the definition boundary difference value involves a quantitative analysis process across three key dimensions. First, the number of conceptual connotation elements is extracted from the definition text using natural language processing techniques, statistically analyzing the differences in the number of conceptual elements across different language versions. Second, the denotation scope is quantified by identifying adjectives, adverbs, and other modifiers in the definition, quantifying the differences in the breadth of the concept's extension. Third, the applicable condition constraints are assessed by extracting conditional clauses, presuppositions, and other restrictive expressions in the definition, evaluating the degree of difference in the concept's applicable scope. The original difference values ​​of the three dimensions are standardized to eliminate the influence of dimensions, and then a weighted summation method is used to calculate the comprehensive definition boundary difference value. When this difference value exceeds a preset tolerance threshold, it indicates a significant comprehension discrepancy between different teaching systems, requiring focused adjustments.

[0087] For example, the assessment of concept complexity is based on two aspects: the structural features of the definition text and the cognitive load. Structural features are comprehensively assessed by analyzing indicators such as the syntactic complexity, nesting level, and number of logical relationships in the definition text; cognitive load is quantified by statistically analyzing factors such as the number of prerequisite knowledge points required to understand the term, the number of related concepts involved, and the complexity of the reasoning steps. The two types of indicators are normalized and then weighted to obtain a comprehensive assessment value of concept complexity. Furthermore, a dual-threshold classification mechanism is used to determine the term category. A first threshold is set to distinguish between basic concepts and applied methods, and a second threshold is set to distinguish between applied methods and theoretical principles. Basic concept terms are usually the fundamental components of a discipline, with relatively simple and clear definitions that do not depend on other complex concepts; applied method terms involve specific operational procedures or analytical tools, requiring certain practical experience to fully understand; theoretical principle terms contain abstract theoretical frameworks or complex logical derivations, requiring a deep disciplinary foundation for accurate grasp.

[0088] Understandably, the assessment of cross-system differences considers not only linguistic differences but also the influence of deeper factors such as cultural background, educational traditions, and academic norms. By constructing a multi-dimensional difference assessment matrix, the system can identify the essential and superficial differences in terminology across different teaching systems, providing a precise basis for subsequent standardization decisions.

[0089] For example, a hierarchical mapping mechanism is used to determine the classification of knowledge categories and teaching modules. First, terms are mapped to primary knowledge categories based on their category attributes, including concepts and principles, methods and tools, and application cases. Then, based on the terminology's position and function within the curriculum, it is further mapped to specific teaching modules, such as introductory modules, core theory modules, practical application modules, and cutting-edge extension modules. This refined classification and attribution mechanism achieves standardized positioning of terms within the international education management system, laying the foundation for unified management of multilingual teaching resources.

[0090] By identifying the specific manifestations of terminology misuse and the incorrect application of terms in the problem-solving process from the types of errors in student assignments, we can collect information on the perspectives of students from different countries regarding the understanding of terms and the original expressions of terms in their local textbooks. We can analyze whether the forms of error stem from semantic differences in the terminology definition texts or from differences in the knowledge structure of the teaching system. We can assess whether the students' comprehension deviations are caused by shifts in terminology meaning or by differences in cultural background cognition, and determine whether the discrepancies belong to semantic shifts or cultural understanding at the cognitive level.

[0091] This study extracts specific manifestations of terminology misuse from student assignments, categorizing them into three types: concept substitution errors, logical reasoning errors, and scope of application errors. It records the specific location number and context of each error during problem-solving. The study then obtains the problem-solving process before and after the error location, identifying the student's reasoning path and conceptual association methods when using the terminology, thus generating a feature dataset of terminology misuse. Data on the understanding of the same terminology from students in different countries is collected, extracting key elements of terminology explanations from student answers and comparing these key elements with the original expressions in local textbooks. The matching degree between student understanding elements and standard definition elements is calculated, recording the differences in definition structure, exemplification methods, and application scenarios of the terminology in textbooks from different countries, forming a cross-national understanding difference record. Based on the feature dataset of terminology misuse and the cross-national understanding difference record, the Pearson correlation coefficient is calculated. If the correlation coefficient exceeds a preset threshold, the error is determined to stem from semantic differences in the terminology's definition text; if the correlation coefficient is below the threshold, structured information such as prerequisite knowledge requirements, conceptual hierarchy relationships, and reasoning logic chains within the teaching system is extracted, determining that the error stems from differences in knowledge structure. Based on the results of the error root cause determination, and combined with the characteristics of the educational traditions, thinking patterns, and expression habits of the student's region, the causes of the comprehension deviation are assessed. If the correlation between the deviation and the shift in the meaning of terminology exceeds a preset threshold, it is classified as a semantic shift; if the correlation between the deviation and the cultural cognitive pattern exceeds a preset threshold, it is classified as a cognitive cultural understanding, thus determining the final category of the disagreement.

[0092] For example, in one implementation, when extracting specific manifestations of terminology misuse from student assignments, the system establishes a three-tiered error classification system. Concept substitution errors manifest as students incorrectly replacing one term with another similar but different term, such as misusing "marginal cost" as "average cost," or confusing "opportunity cost" with "sunk cost." Logical reasoning errors occur when students use correct terminology but violate the inherent logical relationships between terms during reasoning, such as ignoring the role of price regulation mechanisms when analyzing "supply and demand equilibrium." Scope errors occur when students apply terms specific to certain conditions to inapplicable scenarios, such as incorrectly applying the characteristics of a "perfectly competitive market" to the analysis of a "monopoly market." The system scans the assignment text, recording the line number, paragraph position, and the context of the five sentences preceding and following each error, constructing an error location index.

[0093] It should be noted that the construction process of the terminology misuse feature dataset involves multi-dimensional information extraction. The identification of reasoning paths is achieved by tracking the complete thought process of students from asking questions to drawing conclusions. The system analyzes the way terms are used in each reasoning process to determine whether the terms are used as premises, reasoning basis, or conclusion support. The extraction of concept association methods focuses on the logical connections between terms and other concepts, including causal relationships, parallel relationships, and progressive relationships, forming a semantic network structure of terminology usage.

[0094] Specifically, the formation of a record of cross-border understanding differences requires comparative analysis of differences across multiple dimensions. Differences in definition structure are reflected in the way textbooks from different countries define the same term. Some countries prefer deductive definitions, deriving from general principles to specific concepts; others use inductive definitions, summarizing from specific cases to abstract concepts. Differences in exemplification methods are reflected in the cultural background of the selected cases. US textbooks tend to use cases of corporate competition, European textbooks cite more social welfare cases, and Asian textbooks emphasize government regulation cases. Differences in application scenarios are manifested in the actual application areas of the terminology; the same economic term may be used in different scenarios such as market analysis, policy making, or academic research in different countries.

[0095] Preferably, the Pearson correlation coefficient is calculated using standardized data. The indicators in the terminology misuse feature dataset are converted into standard scores to eliminate the influence of dimensions; the degree of difference in cross-national understanding records is quantified into numerical variables. The correlation coefficient r is obtained by calculating the ratio of the product of the covariance and the standard deviation of the two sets of variables, with a value ranging from -1 to 1. When the absolute value of r is greater than 0.7, it indicates a strong correlation, and the system's judgment error mainly stems from semantic differences in terminology interpretation; when the absolute value of r is less than 0.3, it indicates a weak correlation, requiring further analysis of the impact of differences in knowledge structure.

[0096] In one possible implementation, the analysis of knowledge structure differences encompasses three core elements. First, prerequisite knowledge requirements refer to the set of basic concepts needed to understand a term. Different teaching systems have different prerequisite knowledge requirements; for example, understanding "general equilibrium" in some systems requires prior knowledge of "partial equilibrium," while in others it may be directly introduced from the concept of "market clearing." Second, the hierarchical relationship of concepts reflects the position and importance of a term within the knowledge system; the same term may occupy different levels in different systems. Third, the logical reasoning chain embodies the derivation process from basic assumptions to the final conclusion; the length and complexity of the logical chain vary significantly across different teaching systems. The system extracts this structured information to construct a knowledge structure feature vector, which is used to determine the underlying causes of errors.

[0097] For example, the quantification of cultural background cognitive factors involves three dimensions: educational tradition, thinking patterns, and expression habits. The educational tradition dimension is quantified by analyzing the region's teaching method preferences, such as emphasis on memorization versus comprehension, and individual thinking versus group discussion. The thinking pattern dimension assesses students' cognitive tendencies, including the proportion of analytical, holistic, and intuitive thinking. The expression habits dimension examines students' language organization in answering questions, such as a preference for direct statements versus indirect expression, and an emphasis on logical deduction versus experiential induction. After normalizing the feature values ​​of the three dimensions, a comprehensive feature vector of cultural background cognition is formed. Furthermore, the assessment of the causes of comprehension bias employs a dual-judgment mechanism. First, the correlation between comprehension bias and terminological shift is calculated by comparing the similarity between students' error patterns and known semantic shift patterns. Then, the correlation between comprehension bias and cultural cognitive patterns is calculated by analyzing the consistency between students' thinking patterns and their cultural background characteristics. When the semantic shift relevance exceeds the first preset threshold and is higher than the cultural cognition relevance, the disagreement is classified as semantic shift at the semantic level, indicating that the problem mainly stems from differences in language translation and terminology definition; when the cultural cognition relevance exceeds the second preset threshold and is higher than the semantic shift relevance, the disagreement is classified as cultural understanding at the cognitive level, indicating that the root of the problem lies in the differences in thinking patterns under different cultural backgrounds.

[0098] Understandably, this classification mechanism can accurately identify the sources of misunderstanding in international education, providing a basis for developing targeted teaching interventions and achieving accuracy and consistency in knowledge transfer in cross-cultural teaching environments.

[0099] Step S105: Based on the divergence classification results and the interpretation version of the unified semantic benchmark, dynamically update the reference relationships of terms in the teaching resources between textbook chapters and the hierarchical structure between knowledge points to obtain the aligned knowledge framework.

[0100] Based on the divergence classification results, sets of terms belonging to the semantic meaning deviation category and the cognitive cultural understanding category are identified, and unified semantic benchmark interpretation versions corresponding to these term sets are obtained. By comparing the chapter position index of terms in the original teaching resources with the standard position requirements in the unified interpretation version, position deviation values ​​are calculated to determine the list of reference relationships that need to be adjusted. Using the reference relationship list, the mutual reference of terms in each chapter of the textbook is scanned to identify the frequency of terms being cited when defining other concepts, the number of times they are used in case descriptions, and the number of times they are used as premises in reasoning. According to the term dependency relationships specified in the unified semantic benchmark interpretation version, the precedence and succession order between terms is adjusted, and the hierarchical structure of knowledge points is rearranged. Through the adjusted hierarchical structure, the semantic distance and logical dependency strength between each term node are calculated. If the hierarchical position change of a term exceeds a preset threshold, the reference path of its downstream related terms and the connection relationship of its upstream supporting terms are modified. By integrating the updated reference relationships and the adjusted hierarchical structure, a directed acyclic graph between terms is constructed to obtain the aligned knowledge framework.

[0101] For example, in one implementation, when identifying a term set based on the divergence classification results, the system first obtains labeled semantic-level meaning-shifting terms and cognitive-level cultural understanding terms from the preprocessing. For each term, the system reads its corresponding unified semantic benchmark definition, which includes the term's standard definition, recommended usage location, and dependency description. By comparing the differences between the actual chapter number and page range of the term in the original textbook and the standard location requirements, a location deviation value is calculated. When the deviation value exceeds a preset threshold, the term is added to the list to be adjusted.

[0102] It should be noted that the scanning of terminology citations covers three dimensions. Definition citations refer to a term being directly cited when defining other concepts, such as "marginal cost" being cited when defining "marginal revenue"; case citations refer to the use of a term in actual case analysis, demonstrating its application value; and reasoning citations refer to the appearance of a term as a premise or intermediate step in logical deduction, reflecting its supporting role in the knowledge system. The system statistically analyzes the frequency of each citation type to form a terminology citation feature vector.

[0103] Specifically, the adjustment of the hierarchical structure follows the principle of knowledge dependency. Based on the inter-term dependencies specified in the unified semantic benchmark interpretation version, the system identifies which terms must be taught before other terms and which terms can be learned in parallel.

[0104] For example, "demand elasticity" must be taught after "demand curve," while "supply elasticity" and "demand elasticity" can be learned in parallel. The system rearranges the presentation order of knowledge points based on these constraints to ensure the logical nature of the learning path.

[0105] Preferably, in the construction of the directed acyclic graph (DAG), each term serves as a node in the graph, and directed edges between nodes represent knowledge dependencies. Semantic distance is calculated by determining the similarity between the definition texts of two terms; a smaller distance indicates closer conceptual similarity. Logical dependency strength reflects the necessity of one term for understanding another, and its value is determined by statistically analyzing citation patterns in textbooks. When the hierarchical position of a term changes beyond a threshold, the system automatically adjusts the weights of all edges connected to it to maintain the consistency of the knowledge structure.

[0106] In one embodiment, by integrating the updated reference relationships and adjusted hierarchical structure, the system employs a high-level graph algorithm to construct a directed acyclic graph (DAG) between terms, forming a rigorous knowledge dependency framework. The construction process first represents each term as a node in the graph, with node attributes including term ID, complexity value, importance score, and hierarchical affiliation. Then, directed edges are created based on the dependencies between terms, with edge attributes including dependency type (necessary prerequisite / recommended prerequisite / related reference), dependency strength (a real number between 0 and 1), and conversion difficulty (representing the cognitive span from the source node to the target node). The system uses a modified Kahn algorithm for topological sorting to verify the acyclicity of the graph structure: first, all nodes with an in-degree of 0 are added to a queue; then, nodes are removed from the queue one by one, and the in-degree of their adjacent nodes is reduced, repeating this process until the queue is empty. If edges still exist in the final graph, it indicates a circular dependency. The system automatically identifies the minimum feedback edge set, i.e., the minimum set of edges that need to be removed to break the cycle. The edge removal priority is based on a comprehensive score of dependency strength and term importance, preserving key dependencies. The constructed DAG is visualized using a hierarchical layout algorithm (such as the improved Sugiyama algorithm), with nodes arranged vertically according to hierarchy and edges colored according to dependency strength, providing an intuitive reference for planning the teaching sequence. The system also calculates the critical path of the graph, i.e., the longest path from basic concepts to advanced theories, representing the essential path to mastering the complete knowledge system. This knowledge framework, constructed in this way, clearly demonstrates the logical relationships between terms, providing structured guidance for curriculum design and teaching arrangements in international education management.

[0107] Step S106: Obtain the semantic expression of unified terminology from the aligned knowledge framework, apply it to the cultural background description and case analysis sections of teaching resources from different countries, and generate a multilingual knowledge alignment scheme that includes a terminology comparison table, unified definition standards, and cross-cultural teaching suggestions.

[0108] The semantic expressions of unified terminology are extracted from the aligned knowledge framework, including standard definition texts, core conceptual elements, and logical relationship descriptions. The corresponding forms of these semantic expressions in various language versions are obtained. Based on the cultural background characteristics of teaching resources in different countries, a correspondence table between terminology and local cases is established, resulting in a multilingual terminology mapping relationship. Using this multilingual terminology mapping relationship and in conjunction with the requirements for unified interpretation standards, a terminology comparison table is compiled, containing three fields: the original terminology, the standard interpretation, and cultural annotations. This terminology comparison table is applied to the case analysis section, adding cross-cultural teaching suggestions to form a multilingual knowledge alignment scheme that includes a terminology comparison table, unified interpretation standards, and cross-cultural teaching suggestions.

[0109] For example, in one implementation, when extracting unified terminology from the aligned knowledge framework, the system reads the standardized semantic information of each term node. The standard definition text contains the standardized expression of the term, the core concept elements list the key components of the term, and the logical relationship description explains how the term relates to other concepts.

[0110] It should be noted that the terminology reference table adopts a multi-column structure. The original language column records the original expressions in each language, such as "opportunity cost," "opportunity cost," and "Opportunitatskosten"; the standard definition column provides a unified definition of the concept; and the cultural annotation column explains the differences in understanding in different cultural contexts, such as the United States emphasizing the value of individual choices, while Japan focuses on the balance of collective interests.

[0111] Preferably, cross-cultural teaching recommendations are tailored to the cognitive characteristics of students from different countries. For European students who are accustomed to deductive thinking, it is recommended to introduce specific concepts from a theoretical framework; for Asian students who prefer inductive learning, it is recommended to summarize abstract principles through case studies.

[0112] In one embodiment, the multilingual knowledge alignment scheme is output in the form of an electronic document, which includes three parts: a terminology comparison table, a standardized text of definitions, and a cross-cultural teaching guide, providing a standardized reference for textbook compilation and curriculum design in international education management.

[0113] This invention provides an intelligent international education management system, mainly comprising: a terminology definition difference identification module, used to acquire terminology definition text content and corresponding knowledge point lists from teaching resources in various language versions, identify the degree of definitional differences and semantic deviation of terms in different teaching systems, and obtain preliminary assessment results of the semantic shift of terms between different languages; a semantic shift distance determination module, used to acquire examples of terminology usage and contextual expressions in student assignments, extract the actual application scenarios of terms in the assignment text and match them with the relevant knowledge point lists, and determine the semantic shift distance based on the preliminary assessment results of semantic shift; and a unified semantic benchmark generation module, used to identify the types of terminology definition texts that need adjustment from the preliminary assessment results of semantic shift based on the semantic shift distance, and map terminology definition texts with semantic shift distances exceeding a preset threshold to standard knowledge connotation descriptions. The system obtains a unified semantic benchmark version; a disagreement assessment module is used to identify the sources and severity of disagreements in the knowledge systems of students from different countries by comparing the types and frequencies of errors in the use of terms in student assignments in different language versions, assessing whether the disagreements stem from semantic shifts, and obtaining a disagreement classification result; a knowledge framework alignment module is used to dynamically update the reference relationships between terms in textbook chapters and the hierarchical structure between knowledge points in teaching resources based on the disagreement classification result and the unified semantic benchmark version, obtaining an aligned knowledge framework; a multilingual knowledge alignment scheme generation module is used to obtain the semantic expression of unified terms from the aligned knowledge framework, apply it to the cultural background description and case analysis sections of teaching resources from different countries, and generate a multilingual knowledge alignment scheme including a terminology comparison table, unified definition specifications, and cross-cultural teaching suggestions. The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. The present invention has been described in detail with reference to preferred embodiments. Those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An intelligent international education management method, characterized in that, include: Obtain the terminology definition texts and corresponding knowledge point lists from teaching resources in various language versions, identify the degree of difference and semantic deviation in the interpretation of terms in different teaching systems, and obtain the evaluation results of the semantic shift of terms between different languages; Obtain examples of terminology usage and contextual expressions from student assignments, extract application scenarios of terms from assignment texts and match them with the knowledge point list, and determine semantic offset distance by combining the semantic offset evaluation results; Based on the semantic offset distance, identify the term definition text types that need adjustment from the meaning offset evaluation results, map the term definition texts with semantic offset distances exceeding the threshold to the standard knowledge connotation description, and obtain the definition version of the unified semantic benchmark. By comparing the types and frequencies of errors in the use of terminology in different language versions in student assignments, the sources of misunderstanding between students' knowledge systems in different countries are identified, and the results of the misunderstanding classification are obtained. Based on the divergence classification results and the interpretation version of the unified semantic benchmark, the reference relationships of terms in the teaching resources between textbook chapters and the hierarchical structure between knowledge points are updated to obtain the aligned knowledge framework. The semantic expressions of unified terms are extracted from the aligned knowledge framework and applied to the cultural background descriptions and case analyses of teaching resources in different countries, generating a multilingual knowledge alignment scheme that includes a terminology comparison table, unified definition standards, and cross-cultural teaching suggestions.

2. The intelligent international education management method according to claim 1, characterized in that, The process involves acquiring terminology definitions and corresponding knowledge point lists from teaching resources in various language versions, identifying the degree of difference and semantic deviation in terminology definitions across different teaching systems, and obtaining assessment results of the semantic shift of terms across different languages, including: The terminology definition texts are extracted from teaching resources in various languages. The chapter position information and professional field classification labels of the terms in the textbooks are obtained. The semantic distance values ​​of the definition texts between different language versions are calculated to obtain the quantitative results of the degree of difference in terminology definitions. Based on the quantitative results of the degree of difference in the interpretation, the conceptual hierarchy of terms in different teaching systems is identified, and it is determined whether the terms belong to the basic concept layer, the application method layer, or the theoretical principle layer, thus forming a hierarchical classification of the knowledge points of the terms. Using the hierarchical classification of knowledge points, cultural context factors and application scenario descriptions of terms in teaching cases are extracted, semantic feature vectors of terms are constructed, the distance between the semantic feature vectors and standard definition vectors is calculated, the degree of semantic deviation is obtained, and the evaluation results of the meaning shift of terms in different languages ​​are generated.

3. The intelligent international education management method according to claim 1, characterized in that, The process of obtaining examples of terminology usage and contextual expressions in student assignments, extracting application scenarios of terms from the assignment text and matching them with the knowledge point list, and combining the semantic offset evaluation results to determine the semantic offset distance includes: Extract examples of terminology usage from student assignment texts, identify the location and frequency of terminology in the problem-solving process, obtain the context of the vocabulary units before and after the terminology, and extract the application context of the terminology. Based on the co-occurrence relationship between the term and other concepts in the application context, the semantic association strength value between the term and adjacent concepts is calculated to obtain the context association degree value; Using the aforementioned contextual relevance value, the application scenario types of terms in different question types are identified. The differences between the application scenario types and the standard application scenarios defined in the textbook are compared. The number of editing operations between the student's answer step sequence and the standard problem-solving step sequence is calculated to obtain a quantitative value of the degree of comprehension deviation. Combined with the meaning shift evaluation results, the semantic shift distance is determined.

4. The intelligent international education management method according to claim 1, characterized in that, The step involves identifying the types of terminology definitions that need adjustment from the semantic offset evaluation results based on the semantic offset distance, mapping terminology definitions with semantic offset distances exceeding a threshold to standard knowledge connotation descriptions, and obtaining a definition version with a unified semantic benchmark, including: Based on the semantic offset distance, extract the text type markers of the terms' definitions, identify basic concept types, application method types, or theoretical principle types, retrieve the standardized descriptions of the terms in standard documents, and obtain the core semantic units and logical relation words. Using the core semantic units and logical relation words, the semantic components in the original interpretation text are identified, expressions with a deviation less than the threshold are retained, and semantic components with a deviation exceeding the threshold are replaced to generate a interpretation version with a unified semantic benchmark.

5. The intelligent international education management method according to claim 4, characterized in that, The standardized descriptions of the search terms in standard documents are used to obtain core semantic units and logical relation words, including: Extract standardized descriptions of terms from standard documents and identify core semantic units, logical relation words, and qualifying modifiers in the standardized descriptions. Based on the core semantic units and logical relation words, a standard semantic expression structure for terms is constructed, the correspondence between terms and standard structures is recorded, a cross-language term mapping table is formed, and a unified code consisting of subject codes and concept levels is assigned.

6. The intelligent international education management method according to claim 1, characterized in that, The method identifies the sources of misunderstanding between students' knowledge systems from different countries by comparing the types and frequencies of errors in the use of terminology in different language versions in student assignments, and obtains the results of the misunderstanding classification, including: Extract terminology usage error types and error frequencies from student assignments, identify error pattern categories of terms in different language versions, and calculate the correspondence between error pattern categories and conceptual confusion markers; Based on the error pattern categories, the analysis reveals that the errors originate from semantic differences in terminology definition texts or differences in the knowledge structure of teaching systems, generating divergence classification results.

7. The intelligent international education management method according to claim 6, characterized in that, The step involves analyzing the source of errors based on the error pattern categories, identifying semantic differences in terminology definitions or knowledge structure differences in teaching systems, and generating divergence classification results, including: Extract the manifestations of terminological misuse and the location of errors in the problem-solving process from the error pattern categories, and obtain data on students' understanding of the terms and the original expressions in the textbook; Calculate the matching degree between the understanding perspective data and the original expression method, analyze the correlation coefficient between errors and semantic differences or knowledge structure differences in terminology interpretation texts, and determine whether the discrepancies belong to semantic meaning shifts or cognitive cultural understandings.

8. The intelligent international education management method according to claim 1, characterized in that, The process involves updating the terminology reference relationships between textbook chapters and the hierarchical structure between knowledge points in the teaching resources based on the divergence classification results and the interpretation versions of the unified semantic benchmark, to obtain an aligned knowledge framework, including: Based on the aforementioned divergence classification results, identify the list of terminology reference relationships, scan the mutual references of terms in the textbook, and adjust the priority order of terms. Based on the interpretation of the unified semantic benchmark, the hierarchical structure of knowledge points is rearranged, a directed acyclic graph between terms is constructed, and an aligned knowledge framework is generated.

9. The intelligent international education management method according to claim 1, characterized in that, The semantic expression of unified terminology extracted from the aligned knowledge framework is applied to the cultural background explanations and case analyses of teaching resources from different countries, generating a multilingual knowledge alignment scheme that includes a terminology comparison table, unified definition standards, and cross-cultural teaching suggestions, including: Extract the semantic expression of unified terms from the aligned knowledge framework, obtain the corresponding forms of the semantic expression in each language version, and establish a correspondence table between terms and local cases; Compile a terminology lookup table containing the original terms, standard definitions, and cultural annotations, apply it to case analysis, and generate a multilingual knowledge alignment scheme.

10. An intelligent international education management system, characterized in that, The system includes: The terminology definition difference identification module is used to obtain the terminology definition text content and corresponding knowledge point list in teaching resources of various language versions, identify the degree of difference in the definition of terms and the degree of semantic deviation of terms in different teaching systems, and obtain the evaluation results of the meaning shift of terms between different languages. The semantic offset distance determination module is used to obtain examples of term usage and contextual expressions in student assignments, extract the application scenarios of terms in the assignment text and match them with the knowledge point list, and determine the semantic offset distance in combination with the semantic offset evaluation results. The unified semantic benchmark generation module is used to identify the types of term definition texts that need to be adjusted from the meaning offset evaluation results based on the semantic offset distance, and map the term definition texts with semantic offset distances exceeding the threshold to the standard knowledge connotation descriptions to obtain the definition version of the unified semantic benchmark. The Understanding Disagreement Assessment module is used to identify the sources of understanding disagreements between students' knowledge systems from different countries by comparing the types and frequencies of errors in the use of terminology in different language versions of student assignments, and to obtain the disagreement classification results. The knowledge framework alignment module is used to update the reference relationships between terms in the teaching resources and the hierarchical structure between knowledge points in the textbook chapters based on the divergence classification results and the interpretation version of the unified semantic benchmark, so as to obtain the aligned knowledge framework. The multilingual knowledge alignment scheme generation module is used to extract the semantic expression of unified terms from the aligned knowledge framework, apply it to the cultural background description and case analysis of teaching resources in different countries, and generate a multilingual knowledge alignment scheme that includes a terminology comparison table, unified definition standards, and cross-cultural teaching suggestions.

Images