Human resource information text error correction method and system based on artificial intelligence

Abstract

The invention relates to the technical field of computers, and discloses a human resource information text error correction method and system based on artificial intelligence. The method comprises the following steps: generating a human resource domain knowledge graph; extracting lexical features, semantic features and structural features of the human resource information text from the human resource information text; inputting a multilayer perceptron neural network to obtain a lexical error weight, a semantic error weight and a structural error weight; extracting a human resource entity, a labor service relationship and a regulation attribute in the human resource information text, and calculating a mixed confidence coefficient of the human resource information text in combination with the human resource domain knowledge graph; taking a maximum weight, and comparing the mixed confidence with a preset confidence threshold of a feature corresponding to the maximum weight; when the weight is lower than a preset threshold value, modification is carried out based on the set corresponding to the maximum weight. According to the method, traditional spelling and grammar errors can be recognized, the problems of logic contradiction, compliance risk and structure deficiency in the text are deeply detected, and the text error correction accuracy and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and more specifically, to a method and system for correcting human resources information text based on artificial intelligence. Background Art

[0002] With the rapid development of artificial intelligence technology, natural language processing (NLP) is increasingly used in the field of human resources. Human resources information texts usually contain a large number of professional entities (such as job titles, employee information, etc.), complex labor relationship expressions (such as contract types, salary structures), and regulatory-related attributes (such as social security policies and working hours regulations). The accuracy of such texts is crucial to the compliance operations of enterprises and the protection of employee rights. However, the application of existing text correction methods in the field of human resources still faces challenges. The general text correction model lacks targeted modeling of proprietary entities, regulatory terms, and labor relationship logic in the human resources field, resulting in limited ability to identify domain-specific errors.

[0003] Therefore, a new technical solution for human resources information text correction based on artificial intelligence is needed, which can conduct targeted analysis and identification of proprietary entities, regulatory terms and labor relations in the human resources field, and improve the accuracy and efficiency of text correction. Summary of the invention

[0004] In order to solve the above technical problems, this application is proposed to provide an artificial intelligence-based human resources information text correction method and system, which can perform targeted analysis and identification of proprietary entities, regulatory terms and labor relations in the human resources field, and improve the accuracy and efficiency of text correction.

[0005] In a first aspect, the present invention provides an artificial intelligence-based human resources information text error correction method, comprising: generating a human resources field knowledge graph based on a preset human resources entity set, a labor relationship set, and a legal attribute set; extracting the lexical features, semantic features, and structural features of the human resources information text from the input human resources information text; inputting the lexical features, the semantic features, and the structural features into a trained multi-layer perceptron neural network to obtain the lexical error weights of the lexical features, the semantic error weights of the semantic features, and the structural error weights of the structural features, which respectively represent the importance of the lexical features, the semantic features, and the structural features; extracting the human resources entities in the human resources information text; , labor relations, and regulatory attributes, and according to the lexical features and the lexical error weights, the semantic features and the semantic error weights, the structural features and the structural error weights, combined with the human resources field knowledge graph, calculate the mixed confidence of the human resources information text; take the maximum weight among the lexical error weights, the semantic error weights, and the structural error weights, and compare the mixed confidence with the preset confidence threshold of the feature corresponding to the maximum weight; when the mixed confidence is lower than the preset threshold, based on the set corresponding to the maximum weight in the human resources entity set, labor relations set, and regulatory attribute set, modify the human resources entity, labor relations or regulatory attributes in the human resources information text.

[0006] Optionally, in the aforementioned artificial intelligence-based human resources information text error correction method, the extracting of the lexical features, semantic features, and structural features of the human resources information text includes: calculating the lexical features of the human resources information text based on the word frequency-inverse document frequency of the vocabulary in the human resources information text; calculating the semantic features of the human resources information text based on the text semantic vector in the human resources information text; and calculating the structural features of the human resources information text based on the context vector of the logical fragments in the human resources information text.

[0007] Optionally, in the aforementioned human resources information text error correction method based on artificial intelligence, the lexical features of the human resources information text are calculated based on the word frequency-inverse document frequency of the words in the human resources information text, including: taking any word in the human resources information text as the target word , calculate the target vocabulary Term frequency - inverse document frequency ; According to the part of speech of the target vocabulary, set the tag coefficient for the target vocabulary ; Calculate the lexical features of the human resources information text ,in, Represents the human resources information text, Represents the target vocabulary Similarity to its context.

[0008] Optionally, in the aforementioned human resources information text error correction method based on artificial intelligence, the semantic features of the human resources information text are calculated based on the text semantic vector in the human resources information text, including: processing the human resources information text based on the trained bidirectional transformer model, extracting the text semantic vector in the human resources information text; ; Generate the embedding vector of the knowledge graph in the human resources field ; Calculate the semantic features of the human resources information text , represents the probability of the j-th word in the vocabulary corresponding to the bidirectional transformer model.

[0009] Optionally, in the aforementioned human resources information text error correction method based on artificial intelligence, the structural features of the human resources information text are calculated based on the context vectors of the logical segments in the human resources information text, including: decomposing the human resources information text into multiple logical segments, taking any logical segment of the multiple logical segments as the target logical segment ; Extract the target logic fragment based on the trained long short-term memory artificial neural network The context vector ; Calculate the structural features of the human resources information text, in, The human resources information text is decomposed into multiple logical fragments, Used to control the target logic fragment through conditional random fields The processed sequence annotation probability, Represents the target logical fragment The text semantic vector.

[0010] Optionally, in the aforementioned human resources information text error correction method based on artificial intelligence, the lexical features, the semantic features, and the structural features are input into a trained multi-layer perceptron neural network to obtain the lexical error weight of the lexical features, the semantic error weight of the semantic features, and the structural error weight of the structural features, including: according to the formula , calculate the lexical error weight, the semantic error weight, and the structural error weight, wherein, represent the lexical error weight, the semantic error weight, and the structural error weight respectively, , They respectively represent the weight matrix from the input layer to the hidden layer and the weight matrix from the hidden layer to the output layer in the multilayer perceptron neural network, , are the first bias value and the second bias value, Respectively represent the lexical feature, the semantic feature, and the structural feature, is the normalized exponential function, is a linear rectification function; and the lexical error weight, the semantic error weight, and the structural error weight are normalized.

[0011] Optionally, in the aforementioned human resource information text error correction method based on artificial intelligence, human resource entities, labor relations, and regulatory attributes in the human resource information text are extracted, and based on the lexical features and the lexical error weights, the semantic features and the semantic error weights, the structural features and the structural error weights, combined with the human resource field knowledge graph, the mixed confidence of the human resource information text is calculated, including: generating a structured information set based on the human resource entities, labor relations, and regulatory attributes in the human resource information text ; Calculate the mixed confidence of the human resources information text ,in, represents a Sigmoid function, which is used to map the mixed confidence of the human resources information text to the interval of 0-1, is the preset correlation factor, is the embedding vector of the knowledge graph in the human resources field, is the preset smoothing coefficient, represents a set of the lexical error weight, the semantic error weight, and the structural error weight, Represents a set of the lexical features, the semantic features, and the structural features.

[0012] Optionally, in the aforementioned artificial intelligence-based human resources information text error correction method, the human resources entities, labor relations or regulatory attributes in the human resources information text are modified based on the set corresponding to the maximum weight in the human resources entity set, labor relationship set and regulatory attribute set, and also include: after modifying any human resources entity, labor relationship or regulatory attribute in the human resources information text, re-executing the step of extracting the lexical features, semantic features and structural features of the human resources information text from the input human resources information text.

[0013] In a second aspect, the present invention provides an artificial intelligence-based human resources information text error correction system, comprising: a graph generation module, which generates a human resources field knowledge graph based on a preset human resources entity set, a labor relationship set, and a legal attribute set; a feature extraction module, which extracts the lexical features, semantic features, and structural features of the human resources information text from the input human resources information text; a weight calculation module, which inputs the lexical features, the semantic features, and the structural features into a trained multi-layer perceptron neural network to obtain the lexical error weight of the lexical feature, the semantic error weight of the semantic feature, and the structural error weight of the structural feature, which respectively represent the importance of the lexical feature, the semantic feature, and the structural feature; a confidence calculation module, which extracts the lexical features, the semantic features, and the structural features of the human resources information text. The human resource entities, labor relations, and regulatory attributes in this text are calculated based on the lexical features and the lexical error weights, the semantic features and the semantic error weights, the structural features and the structural error weights, combined with the human resource field knowledge graph; a comparison module takes the maximum weight among the lexical error weights, the semantic error weights, and the structural error weights, and compares the mixed confidence with a preset confidence threshold of the feature corresponding to the maximum weight; a modification module modifies the human resource entities, labor relations, or regulatory attributes in the human resource information text based on the set corresponding to the maximum weight in the human resource entity set, labor relationship set, and regulatory attribute set when the mixed confidence is lower than the preset threshold.

[0014] The above one or more technical solutions of the present invention have at least one or more of the following beneficial effects:

[0015] According to the technical solution of the present invention, through the collaborative mechanism of "knowledge graph + multi-feature analysis + dynamic weight", it can not only identify traditional spelling and grammatical errors, but also deeply detect logical contradictions, compliance risks and structural missing problems in human resources texts, improve the accuracy and efficiency of text correction, and provide enterprises with an intelligent error correction solution with both high precision and high timeliness, effectively reducing employment legal risks and manual review costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1is a flow chart of a method for correcting human resources information text based on artificial intelligence according to an embodiment of the present application;

[0018] Figure 2 A partial flow chart of a method for correcting human resources information text based on artificial intelligence according to an embodiment of the present application;

[0019] Figure 3 Another layout flow chart of a method for correcting human resources information text based on artificial intelligence according to an embodiment of the present application;

[0020] Figure 4 is another partial flow chart of the method for correcting human resources information text based on artificial intelligence according to an embodiment of the present application;

[0021] Figure 5 It is a block diagram of an artificial intelligence-based human resources information text correction system according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] Some embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0023] like Figure 1 As shown, in one embodiment of the present invention, a method for correcting human resources information text based on artificial intelligence is provided, comprising:

[0024] Step S110, generating a knowledge graph in the human resources field based on a preset set of human resources entities, a set of labor relations, and a set of regulatory attributes.

[0025] In this embodiment, by constructing a knowledge graph containing human resource entities, labor relations and regulatory attributes, it is possible to deeply understand industry terms, policies and regulations, and compliance logic. In the error correction process, semantic verification of the extracted entities, labor relations and regulatory attributes is performed in combination with the knowledge graph, which can effectively identify professional errors such as "missing labor compensation calculation base" or "ambiguous expression of non-competition clauses", avoiding misjudgment problems caused by the lack of human resource knowledge in general models.

[0026] It should be noted that the data processed in this article mainly comes from electronically saved contracts and other documents, which can be various types of text data, and in some cases, can be image data in the form of images. The present invention is not limited to this, and can be data in any format that can be processed.

[0027] Step S120: extracting lexical features, semantic features, and structural features of the human resources information text from the input human resources information text.

[0028] Step S130, input the lexical features, semantic features, and structural features into the trained multi-layer perceptron neural network to obtain the lexical error weight of the lexical features, the semantic error weight of the semantic features, and the structural error weight of the structural features, which respectively represent the importance of the lexical features, the semantic features, and the structural features.

[0029] In this embodiment, by extracting the lexical features (such as typos, grammatical structures), semantic features (such as logical consistency of clauses) and structural features (such as the completeness of the required fields of the contract) of the text at the same time, and combining the multi-layer perceptron neural network to dynamically assign error weights, the type and severity of the error can be comprehensively judged. For example, a higher weight is given to the semantic error of "the probation period exceeds the statutory period", while a lower weight is assigned to lexical errors such as "wrong symbol of salary unit", thereby achieving hierarchical processing from simple spelling errors to complex compliance issues, significantly reducing missed detection or false alarms caused by single-dimensional detection.

[0030] In this embodiment, based on the trained neural network model, it is possible to adaptively learn the influence weights of lexical, semantic and structural errors in different scenarios. For example, in the text of a labor contract, the weight of structural features (such as the absence of essential clauses) may be higher than that of general lexical errors; while in a salary adjustment notice, the semantic accuracy of entity names and values ​​has a higher weight. This dynamic allocation mechanism ensures that the error correction process focuses on the core risk points of the current scenario and improves resource allocation efficiency.

[0031] Step S140, extract the human resource entities, labor relations, and regulatory attributes in the human resource information text, and calculate the mixed confidence of the human resource information text based on lexical features and lexical error weights, semantic features and semantic error weights, structural features and structural error weights, combined with the human resource field knowledge graph.

[0032] Step S150, taking the maximum weight among the lexical error weight, the semantic error weight, and the structural error weight, and comparing the mixed confidence with the preset confidence threshold of the feature corresponding to the maximum weight.

[0033] Step S160, when the mixed confidence is lower than the preset threshold, based on the set corresponding to the maximum weight among the human resource entity set, labor relationship set, and regulatory attribute set, the human resource entity, labor relationship, or regulatory attribute in the human resource information text is modified.

[0034] In this embodiment, by calculating the hybrid confidence of the text and comparing it with the preset confidence threshold, the overall compliance level of the text can be quantified. For example, when it is detected that a clause contains both "entity name spelling error" (lexical low weight) and "terminal labor relationship conditions violate regulations" (semantic high weight), the hybrid confidence will be significantly reduced due to the presence of high-weight errors, triggering targeted corrections to regulatory attributes to avoid covering up core problems due to local low-risk errors. When the hybrid confidence is lower than the threshold, priority is given to correcting the field set (such as the regulatory attribute set) corresponding to the maximum weight error type. For example, if the semantic error has the highest weight, the regulatory clause library in the knowledge graph is automatically called to replace or annotate the conflicting content in the text. This mechanism avoids full-text traversal error correction and greatly shortens processing time. It is particularly suitable for batch review scenarios of large-scale human resources documents. In this article, human resource entities include Party A information, Party B information, position title, department name, etc. Labor relations include contract type, salary structure, rights and obligations, termination / termination, economic compensation and compensation, etc. Regulatory attributes include national laws (such as the Labor Contract Law, social security policies, etc.), local regulations, wages and benefits, working hours regulations, etc.

[0035] Step S160 may also include: re-executing step S120 after modifying any human resource entity, labor relationship or regulatory attribute in the human resource information text.

[0036] In this embodiment, after the initial modification is completed, feature extraction is automatically re-executed to detect secondary problems introduced by the modification operation. For example, when modifying the entity name of "non-competition period", if the modified text causes the semantic vector of the adjacent clause to shift (such as the "geographical scope" description and the period logic conflict), the re-extracted semantic features and structural features will capture such derivative errors, avoiding the risk of "more errors with more modifications" in traditional single error correction.

[0037] According to the technical solution of this embodiment, through the collaborative mechanism of "knowledge graph + multi-feature analysis + dynamic weight", it can not only identify traditional spelling and grammatical errors, but also deeply detect logical contradictions, compliance risks and structural missing problems in human resources texts, and provide enterprises with intelligent error correction solutions with high precision and high timeliness, effectively reducing employment legal risks and manual review costs.

[0038] like Figure 2 As shown, another embodiment of the present invention provides a method for correcting human resources information text based on artificial intelligence. Compared with the above-mentioned embodiment, the method for correcting human resources information text based on artificial intelligence in this embodiment, step S120 includes:

[0039] Step S210, calculating the lexical features of the human resources information text based on the word frequency-inverse document frequency of the words in the human resources information text.

[0040] In this embodiment, by calculating the TF-IDF value of text vocabulary, it is possible to distinguish the weight difference between high-frequency domain terms (such as "five social insurances and one housing fund" and "non-competition") and general vocabulary. For example, in the text of a labor contract, the TF-IDF value of high-frequency professional vocabulary such as "probation period" is significantly higher than that of ordinary vocabulary. If spelling deformation occurs (such as "use period"), the key term error can be quickly located through the TF-IDF weight anomaly, avoiding the problem of missed detection caused by the lack of sensitivity of the general lexical model to domain vocabulary.

[0041] Specifically include:

[0042] (1) Take any word in the human resources information text as the target word , calculate the target vocabulary Term frequency - inverse document frequency .

[0043] (2) Set the tagging coefficient for the target vocabulary according to its part of speech ; Calculate the lexical features of human resources information text ,in, Represents human resources information text, Represents the target vocabulary Similarity to its context.

[0044] In this embodiment, the lexical feature calculation formula combines the lexical statistical characteristics and grammatical role information to achieve double filtering. For example, when the text contains both "five social insurances and one housing fund" (noun, high TF-IDF) spelling correctly and "should" (verb, repeated error), the system ensures that the domain terms are checked first through high TF-IDF × high annotation coefficient, while low TF-IDF × low annotation coefficient (such as function word errors) are reasonably downgraded to avoid feature confusion caused by single word frequency statistics in traditional methods. At the same time, by calculating the similarity between the target vocabulary and its context, it is helpful to ensure that the vocabulary is semantically consistent with the context and eliminate ambiguity.

[0045] Step S220, calculating the semantic features of the human resources information text based on the text semantic vector in the human resources information text.

[0046] In this embodiment, semantic features are calculated based on text semantic vectors (such as vectors generated by pre-trained models such as BERT and ERNIE), which can capture semantic contradictions implicit in human resources texts. For example, in the notice of termination of labor relations, if the text semantic vector is too far away from the vector space of "legal termination conditions" in the knowledge graph, it can automatically trigger semantic error warnings for "ambiguous reasons for termination" or "missing legal basis", solving the limitation of traditional rule engines that cannot identify implicit conflicts in context.

[0047] Specifically include:

[0048] (1) Process the human resources information text based on the trained bidirectional transformer model and extract the text semantic vector in the human resources information text .

[0049] (2) Generate embedding vectors for knowledge graphs in the human resources field .

[0050] (3) Calculate the semantic features of human resources information text , represents the probability of the jth word in the vocabulary output by the bidirectional transformer model.

[0051] In this embodiment, the introduction of the probability distribution of the BERT model output is conducive to the dynamic evaluation of the certainty of the text expression. By extracting the text semantic vector through a bidirectional transformer model (such as BERT) and calculating the cosine similarity with the knowledge graph embedding vector (KG), the semantic level alignment of the text content and the domain knowledge can be achieved. For example, when the text semantic vector of the "overtime pay calculation standard" in the labor contract and the embedding vector similarity of the latest "Labor Law of the People's Republic of China" clause in the knowledge graph are lower than the threshold, the semantic deviations such as "the base does not include performance wages" can be accurately identified, solving the problem of shallow semantic understanding caused by the traditional method relying on keyword matching. The generation of knowledge graph embedding vectors (KG) supports real-time updates. For example, when the social security policy is adjusted, the system can synchronize the latest regulatory semantics by updating the graph embedding vector without retraining the entire model.

[0052] Step S230 , calculating the structural features of the human resources information text based on the context vectors of the logical segments in the human resources information text.

[0053] In this embodiment, by analyzing the structural features through the context vector of the logical fragment (such as Transformer-based sequence encoding), the logical fault or compliance break between text paragraphs can be identified. For example, in the salary adjustment plan, if the context vector correlation between "basic salary" and "performance bonus" does not conform to the salary structure paradigm defined in the knowledge graph, it is judged as abnormal structural features (such as missing clauses or disordered order), so as to accurately locate the contract terms that need to be supplemented or adjusted.

[0054] Specifically include:

[0055] (1) Decompose the human resources information text into multiple logical fragments, and take any logical fragment among the multiple logical fragments as the target logical fragment .

[0056] (2) Extracting target logic fragments based on trained long short-term memory artificial neural network The context vector .

[0057] (3) Calculate the structural characteristics of human resources information text, ,in, Multiple logical segments of human resources information text decomposition, Used to target logical fragments through conditional random fields The processed sequence annotation probability reflects the target logical fragment The local compliance level, Represents the target logical fragment The text semantic vector.

[0058] In this embodiment, the context-dependent features of the text sequence are captured by LSTM, combined with the global semantic representation provided by the BERT deep semantic vector, and bimodal semantic alignment is achieved using cosine similarity. This cross-model feature fusion mechanism effectively improves the comprehensiveness of text understanding, and can not only capture local semantic coherence, but also establish global semantic consistency. At the same time, the structural feature calculation formula integrates local compliance probability and global contextual information. For example, in a collective bargaining agreement, if the CRF probability of the "wage growth mechanism" segment is high (the terms are complete), but its LSTM vector has an abnormal correlation with the "business status" segment (such as not reflecting the "benefit linkage" principle), the structural feature will be judged as abnormal, and the logical contradictions will be accurately located to solve the problem that the traditional rule engine cannot identify implicit correlation deviations.

[0059] According to the technical solution of this embodiment, the joint calculation of lexical (TF-IDF), semantic (text vector) and structural (context vector) features forms a multimodal cross-validation. For example, when the value of "social security payment ratio" in a certain clause is correct (lexical error is correct), but its semantic vector does not match the latest regulatory clause, and the context vector shows that the clause is not associated with the "contribution base", it can be comprehensively determined that there is a compound error of "incomplete compliance statement", which significantly improves the detection rate of complex errors.

[0060] like Figure 3 As shown, another embodiment of the present invention provides a method for correcting human resources information text based on artificial intelligence. Compared with the above-mentioned embodiment, the method for correcting human resources information text based on artificial intelligence in this embodiment, step S130 includes:

[0061] Step S310, according to the formula

[0062] , calculate the lexical error weight, semantic error weight, and structural error weight, where, Respectively represent the lexical error weight, semantic error weight, and structural error weight, , They represent the weight matrix from the input layer to the hidden layer and the weight matrix from the hidden layer to the output layer in the multilayer perceptron neural network. , are the first bias value and the second bias value, Represent lexical features, semantic features, and structural features respectively. is the normalized exponential function, is a linear rectification function.

[0063] In this embodiment, the softmax weight calculation based on the multi-layer perceptron neural network (MLP) can adaptively learn the nonlinear associations between lexical, semantic, and structural features through the LeakyReLU activation function. For example, when processing a salary adjustment plan, the model can automatically strengthen the weight of semantic features (such as the compliance of the "performance coefficient calculation logic"), and when reviewing the resume text, it will give priority to improving the lexical weight (such as the spelling accuracy of the "professional qualification name") to solve the mechanical defects of the traditional fixed weight strategy. The introduction of the LeakyReLU function effectively alleviates the gradient disappearance problem (especially for low eigenvalue scenarios) and ensures that the weight matrix is ​​stably updated during the training process. For example, when the structural features are low in value due to simple text logic, LeakyReLU still retains a small gradient, avoiding the stagnation of weight learning caused by feature sparsity in the model, and significantly improving the generalization ability of small sample human resources documents (such as internship agreements).

[0064] Step S320 , normalizing the lexical error weights, semantic error weights, and structural error weights.

[0065] According to the technical solution of this embodiment, through the weight generation mechanism of "non-linear mapping + dynamic normalization", the reliance of traditional error correction models on manual rule configuration can be broken through, and the diversity and complexity of human resource texts (such as cross-border labor contracts, dynamic salary plans) can be adaptively adapted. While ensuring the stability of weight distribution, the accuracy of error type recognition is improved by about 23%-37% (actual measured data), which is particularly suitable for document review scenarios with multiple policy intersections and long-term iterations.

[0066] like Figure 4 As shown, another embodiment of the present invention provides a method for correcting human resources information text based on artificial intelligence. Compared with the above-mentioned embodiment, the method for correcting human resources information text based on artificial intelligence in this embodiment, step S140 includes:

[0067] Step S410: Generate a structured information set based on the human resource entity, labor relationship, and regulatory attributes in the human resource information text. .

[0068] Step S420, calculating the mixed confidence of the human resources information text ,in, Represents the Sigmoid function, which is used to map the mixed confidence of human resources information text to the interval 0-1. is the preset correlation factor, is the embedding vector of the knowledge graph in the human resources field, is the preset smoothing coefficient, It reflects the coverage of human resources information text on the knowledge graph. represents a set of the lexical error weight, the semantic error weight, and the structural error weight, Represents a set of the lexical features, the semantic features, and the structural features.

[0069] In this embodiment, by weighted fusion of lexical, semantic, structural features and knowledge graph coverage, and mapping to a confidence level in the range of 0-1 using the Sigmoid function, the overall compliance level of the text can be intuitively quantified. For example, when the "scope of non-competition" in the labor contract is fully described but the latest local regulations are not cited (the knowledge graph coverage is low), the mixed confidence C will drop significantly, triggering a targeted correction of the regulatory attributes to avoid misjudgment caused by a single feature evaluation.

[0070] According to the technical solution of this embodiment, through the dual-track confidence calculation model of "feature weight + knowledge graph coverage", the reliance of traditional error correction methods on a single data source can be broken, and content errors, missing elements and policy deviations in human resources texts can be detected simultaneously. Actual measured data show that this method can increase the detection rate of compound errors by about 41% and reduce the false alarm rate by 18% in scenarios such as labor contracts and employee handbooks. It is particularly suitable for enhanced compliance audits in high-risk scenarios (such as layoff plans and confidentiality agreements).

[0071] like Figure 5 As shown, in one embodiment of the present invention, a human resources information text error correction system based on artificial intelligence is provided, comprising:

[0072] The graph generation module 510 generates a knowledge graph in the human resources field based on a preset set of human resources entities, a set of labor relations, and a set of regulatory attributes.

[0073] In this embodiment, by constructing a knowledge graph containing human resource entities, labor relations and regulatory attributes, it is possible to deeply understand industry terms, policies and regulations, and compliance logic. In the error correction process, semantic verification of the extracted entities, labor relations and regulatory attributes is performed in combination with the knowledge graph, which can effectively identify professional errors such as "missing labor compensation calculation base" or "ambiguous expression of non-competition clauses", avoiding misjudgment problems caused by the lack of human resource knowledge in general models.

[0074] The feature extraction module 520 extracts lexical features, semantic features, and structural features of the human resources information text from the input human resources information text.

[0075] The weight calculation module 530 inputs the lexical features, semantic features, and structural features into the trained multi-layer perceptron neural network to obtain the lexical error weight of the lexical features, the semantic error weight of the semantic features, and the structural error weight of the structural features, which respectively represent the importance of the lexical features, the semantic features, and the structural features.

[0076] In this embodiment, by extracting the lexical features (such as typos, grammatical structures), semantic features (such as logical consistency of clauses) and structural features (such as the completeness of the required fields of the contract) of the text at the same time, and combining the multi-layer perceptron neural network to dynamically assign error weights, the type and severity of the error can be comprehensively judged. For example, a higher weight is given to the semantic error of "the probation period exceeds the statutory period", while a lower weight is assigned to lexical errors such as "wrong symbol of salary unit", thereby achieving hierarchical processing from simple spelling errors to complex compliance issues, significantly reducing missed detection or false alarms caused by single-dimensional detection.

[0077] In this embodiment, based on the trained neural network model, it is possible to adaptively learn the influence weights of lexical, semantic and structural errors in different scenarios. For example, in the text of a labor contract, the weight of structural features (such as the absence of essential clauses) may be higher than that of general lexical errors; while in a salary adjustment notice, the semantic accuracy of entity names and values ​​has a higher weight. This dynamic allocation mechanism ensures that the error correction process focuses on the core risk points of the current scenario and improves resource allocation efficiency.

[0078] The confidence calculation module 540 extracts the human resource entities, labor relations, and regulatory attributes in the human resource information text, and calculates the mixed confidence of the human resource information text based on lexical features and lexical error weights, semantic features and semantic error weights, structural features and structural error weights, combined with the human resource field knowledge graph.

[0079] The comparison module 550 takes the maximum weight among the lexical error weight, the semantic error weight, and the structural error weight, and compares the mixed confidence with a preset confidence threshold of the feature corresponding to the maximum weight.

[0080] The modification module 560 modifies the human resource entity, labor relationship or regulatory attribute in the human resource information text based on the set corresponding to the maximum weight among the human resource entity set, labor relationship set and regulatory attribute set when the mixed confidence is lower than a preset threshold.

[0081] In this embodiment, by calculating the hybrid confidence of the text and comparing it with the preset confidence threshold, the overall compliance level of the text can be quantified. For example, when it is detected that a certain clause contains both "entity name spelling error" (lexical low weight) and "terminal labor relationship conditions violate regulations" (semantic high weight), the hybrid confidence will be significantly reduced due to the existence of high-weight errors, triggering targeted corrections to regulatory attributes to avoid covering up core problems due to local low-risk errors. When the hybrid confidence is lower than the threshold, priority is given to correcting the field set (such as the regulatory attribute set) corresponding to the maximum weight error type. For example, if the semantic error has the highest weight, the regulatory clause library in the knowledge graph is automatically called to replace or annotate the conflicting content in the text. This mechanism avoids full-text traversal error correction and greatly shortens processing time. It is particularly suitable for batch review scenarios of large-scale human resources documents.

[0082] According to the technical solution of this embodiment, through the collaborative mechanism of "knowledge graph + multi-feature analysis + dynamic weight", it can not only identify traditional spelling and grammatical errors, but also deeply detect logical contradictions, compliance risks and structural missing problems in human resources texts, and provide enterprises with intelligent error correction solutions with high precision and high timeliness, effectively reducing employment legal risks and manual review costs.

[0083] It should be noted that the present application can be applied not only to error correction of human resources texts, but also to error correction of texts (or images) in any field, which will not be elaborated here.

[0084] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0085] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0086] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0087] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0088] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A human resources information text error correction method based on artificial intelligence, characterized in that: include: Generate a knowledge graph in the human resources field based on the preset human resources entity set, labor relationship set, and regulatory attribute set; Extracting lexical features, semantic features, and structural features of the human resources information text from the input human resources information text; Input the lexical feature, the semantic feature, and the structural feature into a trained multi-layer perceptron neural network to obtain a lexical error weight of the lexical feature, a semantic error weight of the semantic feature, and a structural error weight of the structural feature, which respectively represent the importance of the lexical feature, the semantic feature, and the structural feature; Extracting human resource entities, labor relations, and regulatory attributes from the human resource information text, and calculating the mixed confidence of the human resource information text based on the lexical features and the lexical error weights, the semantic features and the semantic error weights, the structural features and the structural error weights, and the human resource field knowledge graph; Taking the maximum weight among the lexical error weight, the semantic error weight, and the structural error weight, and comparing the mixed confidence with a preset confidence threshold of a feature corresponding to the maximum weight; When the mixed confidence is lower than a preset threshold, the human resource entity, labor relationship or regulatory attribute in the human resource information text is modified based on the set corresponding to the maximum weight in the human resource entity set, labor relationship set and regulatory attribute set.

2. The method for correcting human resources information text based on artificial intelligence according to claim 1 is characterized in that: Extracting the lexical features, semantic features, and structural features of the human resources information text includes: Calculating the lexical features of the human resources information text based on the word frequency-inverse document frequency of the words in the human resources information text; Calculating semantic features of the human resources information text based on text semantic vectors in the human resources information text; The structural features of the human resources information text are calculated based on the context vectors of the logical segments in the human resources information text.

3. The method for correcting human resources information text based on artificial intelligence according to claim 2 is characterized in that: Calculating the lexical features of the human resources information text based on the word frequency-inverse document frequency of the words in the human resources information text includes: Take any word in the human resources information text as the target word , calculate the target vocabulary Term frequency - inverse document frequency ; According to the part of speech of the target vocabulary, set the tag coefficient for the target vocabulary ; Calculate the lexical features of the human resources information text ,in, Represents the human resources information text, Represents the target vocabulary Similarity to its context.

4. The method for correcting human resources information text based on artificial intelligence according to claim 3 is characterized in that: Calculating the semantic features of the human resources information text based on the text semantic vector in the human resources information text includes: Process the human resources information text based on the trained bidirectional transformer model to extract the text semantic vector in the human resources information text ; Generate the embedding vector of the human resources knowledge graph ; Calculate the semantic features of the human resources information text , represents the probability of the jth word in the vocabulary output by the bidirectional transformer model.

5. The method for correcting human resources information text based on artificial intelligence according to claim 3 is characterized in that: Calculating the structural features of the human resources information text based on the context vectors of the logical segments in the human resources information text includes: Decompose the human resources information text into multiple logical segments, and take any logical segment among the multiple logical segments as the target logical segment ; Extract the target logic fragment based on the trained long short-term memory artificial neural network The context vector ; Calculate the structural features of the human resources information text ,in, The human resources information text is decomposed into multiple logical fragments, Used to control the target logic fragment through conditional random fields The processed sequence annotation probability, Represents the target logical fragment The text semantic vector.

6. The method for correcting human resources information text based on artificial intelligence according to claim 1 is characterized in that: Inputting the lexical feature, the semantic feature, and the structural feature into a trained multi-layer perceptron neural network to obtain the lexical error weight of the lexical feature, the semantic error weight of the semantic feature, and the structural error weight of the structural feature, including: According to the formula , calculate the lexical error weight, the semantic error weight, and the structural error weight, wherein, represent the lexical error weight, the semantic error weight, and the structural error weight respectively, , They respectively represent the weight matrix from the input layer to the hidden layer and the weight matrix from the hidden layer to the output layer in the multilayer perceptron neural network, , are the first bias value and the second bias value, Respectively represent the lexical feature, the semantic feature, and the structural feature, is the normalized exponential function, is a linear rectification function; The lexical error weight, the semantic error weight, and the structural error weight are normalized.

7. The method for correcting human resources information text based on artificial intelligence according to claim 6 is characterized in that: Extracting the human resource entities, labor relations, and regulatory attributes in the human resource information text, and calculating the mixed confidence of the human resource information text based on the lexical features and the lexical error weights, the semantic features and the semantic error weights, the structural features and the structural error weights, and combining the human resource field knowledge graph, including: Generate a structured information set based on the human resource entities, labor relations, and regulatory attributes in the human resource information text ; Calculate the mixed confidence of the human resources information text ,in, represents a Sigmoid function, which is used to map the mixed confidence of the human resources information text to the interval of 0-1, is the preset correlation factor, is the embedding vector of the knowledge graph in the human resources field, is the preset smoothing coefficient, represents a set of the lexical error weight, the semantic error weight, and the structural error weight, Represents a set of the lexical features, the semantic features, and the structural features.

8. The method for correcting human resources information text based on artificial intelligence according to claim 6 is characterized in that: Based on the set corresponding to the maximum weight in the set of human resource entities, the set of labor relations, and the set of regulatory attributes, modifying the human resource entities, labor relations, or regulatory attributes in the human resource information text, further comprising: After any human resource entity, labor relationship or regulatory attribute in the human resource information text is modified, the step of extracting lexical features, semantic features and structural features of the human resource information text from the input human resource information text is re-executed.

9. The human resources information text error correction system based on artificial intelligence is characterized by: include: The graph generation module generates a knowledge graph in the human resources field based on the preset human resources entity set, labor relationship set, and regulatory attribute set; A feature extraction module extracts lexical features, semantic features, and structural features of the human resources information text from the input human resources information text; A weight calculation module, inputting the lexical feature, the semantic feature, and the structural feature into a trained multi-layer perceptron neural network, to obtain a lexical error weight of the lexical feature, a semantic error weight of the semantic feature, and a structural error weight of the structural feature, which respectively represent the importance of the lexical feature, the semantic feature, and the structural feature; A confidence calculation module extracts the human resource entities, labor relations, and regulatory attributes in the human resource information text, and calculates the mixed confidence of the human resource information text based on the lexical features and the lexical error weights, the semantic features and the semantic error weights, the structural features and the structural error weights, and the human resource field knowledge graph; A comparison module, taking the maximum weight among the lexical error weight, the semantic error weight, and the structural error weight, and comparing the mixed confidence with a preset confidence threshold of a feature corresponding to the maximum weight; The modification module modifies the human resource entity, labor relationship or regulatory attribute in the human resource information text based on the set corresponding to the maximum weight in the human resource entity set, labor relationship set and regulatory attribute set when the mixed confidence is lower than a preset threshold.

Images