116 results about "Sentence pair" patented technology

The invention provides a sentence semantic perception-based data management method and related equipment, and relates to the technical field of cache management, and the method comprises the following steps: segmenting a long text to obtain a plurality of sentences and a sorting sequence of the sentences; calculating importance scores of the tokens corresponding to the sentences based on the sentences and the sorting sequence of the sentences; key tokens in the sentences are screened out according to the importance scores of the tokens corresponding to the sentences and the dynamic budget corresponding to the sentences, the key tokens are stored in a reserved set of the corresponding sentences, and then the local weighted attention score of each token is calculated; performing weighted fusion based on the local weighted attention score of each token in the sentence and the fusion weight of a preset attention head to obtain a sentence-level semantic vector; and constructing semantic index cache and key value pair (KV) content cache based on the sentence-level semantic vectors of the sentences and the corresponding reserved sets. Therefore, in combination with a KV cache management method of context-dependent dynamic loading, a more efficient and more intelligent reasoning process is realized.

The invention discloses a teaching course recommendation method and system based on English learning data, particularly relates to the field of semantic processing, is used for solving the problem of poor pertinence of a traditional English learning course, and comprises the following steps: aiming at systematic difference of native language and English expression on a syntactic structure, constructing a language structure difference map and extracting a structure offset label; the method comprises the following steps: performing dependency syntactic analysis on semantic consistent sentence pairs of multi-language aligned corpora to generate a structural difference feature set; on the basis, a semantic-grammar dimension mapping graph is constructed in a classified mode, and a general structure expression vector is established for graph nodes. A teaching course is divided into knowledge point units in combination with a context label, and a mapping relation between a structure label and the course is established. The system performs structure analysis and semantic matching on sentences input by the learner, identifies structure migration type expression errors, and recommends accurate teaching content according to context and structure labels.

System and method for aspect-level sentiment classification. The system includes a computing device, the computing device has a processer and a storage device storing computer executable code. The computer executable code is configured to: receive an aspect term-sentence pair; embed the aspect term-sentence pair; parse the sentence using multiple parsers to obtain dependency trees, and perform edge union to obtain a merged graph; combine the embedding and the merged graph to obtain a relation graph; perform a relation graph neural network on the relation graph; extract hidden representation of the aspect term from updated relation neural network; and classify the aspect term based on the extracted representation to obtain a predicted classification label of the aspect term. During training, the computer executable code is further configured to calculate a loss function based on the predicted label and the ground truth label, and adjust parameters of models.

The invention relates to the technical field of natural language processing, in particular to a long text matching method and system based on key information and difference characteristics, and solves the problems that in the prior art, key information is dispersed due to noise interference in long text processing, and semantics of words or phrases in long text processing are more fuzzy and diversified. According to the method, a long text is preprocessed to obtain a test set sentence pair, a training set sentence pair and a verification set sentence pair, then a training model is obtained through sentence-level information entropy screening, word-level dynamic filtering, semantic difference enhancement and adaptive feature fusion, and finally after verification is conducted through the verification set sentence pair, the test set sentence pair predicts and outputs a result. The system comprises a long text matching system data preprocessing unit, a long text matching system function module training unit and a long text matching system function module result output unit. Key information is extracted for learning in long text matching, a large amount of computing power does not need to be consumed, and the matching effect is improved.

The invention discloses a bidirectional information retrieval enhancement generation method for a large language model, and belongs to the technical field of artificial intelligence. In order to overcome the defects that noise is introduced and key evidences are omitted due to the fact that traditional RAG only executes'query-document 'one-way retrieval, a two-way semantic perception retrieval enhancement generation model and a two-stage training framework are constructed, wherein in the first stage, the positive / negative example distance is increased in an embedded space in a contrast learning self-supervision mode; in the second stage, fine-grained correlation discrimination is carried out on query-document bidirectional sentences through supervised dichotomy, and probabilistic correlation scores are output; in the reasoning stage, the bidirectional probabilities are fused according to Bayesian to obtain final relevancy, document reordering is carried out, and plug and play can be achieved without fine adjustment of LLM in the whole process. According to the method, the accuracy and consistency of single-hop and multi-hop questions and answers and fact checking tasks are remarkably improved, and the method has the advantages of light weight and low deployment cost.

The invention provides a key sentence extraction method combining keyword guidance and large language model near-end strategy optimization. The key sentence extraction method comprises the following steps: constructing a keyword-key sentence pair; the relevancy is evaluated by using a joint matching model, and a reward value is generated; introducing a KL divergence to measure the difference between the training model and the reference model, and estimating the value score of the current state in combination with a state value network; and optimizing the guidance model through a near-end strategy to realize extraction of key sentences. According to the method provided by the embodiment of the invention, the keyword-key sentence pair is constructed, the relevancy of the keyword-key sentence pair is evaluated by using the joint matching model, the reward value is generated, the KL divergence is introduced to measure the difference between the training model and the reference model, the value score of the current state is estimated by combining the state value network, and the guidance model is optimized through the near-end strategy. According to the method, the key sentences are extracted, the problems of high dependence on annotation data and high training cost of a large language model are solved to a great extent, and the key sentence extraction effect is remarkably improved.

Method and apparatus for dynamic remediation. A set of records associated with a project is accessed. The set of records is processed using one or more natural language processing techniques to generate textual data comprising a plurality of pairs of sentences corresponding to one or more topics associated with the project. An issue for at least one topic associated with the project is identified based on the textual data, comprising identifying a pair of sentences that comprises a first sentence and a second sentence, calculating a sentence similarity score by comparing the first and second sentences using a similarity metric, and determining that the sentence similarity score satisfies one or more criteria. In response to determining that the one or more criteria are satisfied, a project meeting for the issue is scheduled based at least in part on a criticality of the issue.

The application relates to a sentence vector generation method and device, a matching method and device and a storage medium. The sentence vector generation method comprises the following steps: preprocessing a received original sentence to obtain a filtered stop word sentence and a filtered stop word and entity label marked sentence; processing the filtered stop word sentence to obtain a first sentence vector, wherein the first sentence vector comprises attribute information of characters included in the filtered stop word sentence; processing the filtered stop word and entity label marked sentence and the original sentence to obtain a second sentence vector, wherein the second sentence vector comprises information of the entity label; and a target vector of the original sentence comprises the first sentence vector and the second sentence vector. The target vector of the sentence obtained by the method can contain more effective information, which is beneficial to improving the accuracy of the result output by an intelligent question and answer system.

The invention provides a text information recognition method and device, equipment and a storage medium. In some embodiments of the present disclosure, a first paragraph text and a second paragraph text of an audit report text are obtained; performing sentence segmentation processing on the first paragraph text and the second paragraph text to obtain a first sentence of the first paragraph text and a second sentence of the second paragraph text; combining the first clause and the second clause pairwise to obtain a sentence pair; encoding each sentence pair to obtain a sentence vector corresponding to each sentence pair; performing semantic emotion recognition on the sentence vector corresponding to each sentence pair to obtain a consistency result of each sentence pair; determining consistency information of the first paragraph text and the second paragraph text according to a consistency result of each sentence pair; based on semantic emotion recognition in the field of natural language processing, the audit report text is subjected to consistency auditing automatically, the labor cost is reduced, the auditing efficiency is improved, and the auditing accuracy is improved.

The present invention discloses a method, medium, and device for improving automatic machine translation quality assessment using retrieval. The method comprises: a retrieval phase: for machine translation quality assessment sentence pairs, relevant parallel sentence pairs are retrieved from a database for the words to be assessed in the machine translation quality assessment sentence pairs; and a machine translation quality assessment phase: the retrieved parallel sentence pairs are encoded and incorporated into a machine translation quality assessment model. The present invention can directly and effectively utilize relevant parallel sentence pairs, while also alleviating the problem of sparse training data for machine translation quality assessment. It also better explains the reasons why the model makes relevant decisions, eliminates the need for model retraining, and avoids the drawback of end-to-end models forgetting training data during training, thereby improving the performance of machine translation quality assessment models.

The invention discloses a grammar error correction method based on large model syntactic preference optimization. According to the grammar error correction method, syntactic relevance between retrieval example sentence pairs and syntactic relevance between the retrieval example sentence pairs are cooperatively utilized. On one hand, a Monte Carlo tree search algorithm is used for exploring the influence of syntactic structure differences among retrieval examples on error correction performance, search path selection is adjusted based on dynamic weights, and syntactic perception corpora containing hidden syntactic association are constructed. On the other hand, by constructing a syntactic preference alignment mechanism, syntactic knowledge contained in syntactic perception corpora is effectively utilized, dynamic adjustment of grammar error correction model parameters is achieved, and then the performance of the grammar error correction model in a complex grammar error correction task is improved. According to the method, a comprehensive contrast experiment is carried out on two Chinese error correction data sets and two English error correction data sets. Experimental results show that the method can effectively integrate the structured syntactic features in the retrieval examples, and effectively relieve the over-correction phenomenon of the grammar error correction model.

The present application relates to a text sentence processing method, apparatus, computer device, and storage medium, comprising: obtaining sample text sentences containing entity pairs and relationship labels for the entity pairs; extracting positive and negative sentence pairs from the sample text sentences based on the relationship labels, and performing positive and negative sampling processing to obtain a training set; inputting the training set into a relationship extraction model to be trained to generate a loss value including a contrast loss value; the contrast loss value is used to characterize the difference between the similarity of the sentences in the positive sentence pairs and the similarity of the sentences in the negative sentence pairs; adjusting the parameters of the relationship extraction model based on the loss value, and returning to the step of extracting positive and negative sentence pairs from the sample text sentences based on the relationship labels to perform iterative training until a training stop condition is met to obtain a relationship extraction model; the relationship extraction model is used to identify entity relationships between entity pairs in the text sentences. The present method can effectively improve the accuracy of entity relationship extraction.

The invention relates to a large model Mian machine translation method based on language knowledge retrieval, and belongs to the field of natural language processing. Comprising the following steps: constructing tree root nodes of a tree-shaped retrieval structure for parallel corpora from Chinese to Muranju language; dividing the parallel corpora from Chinese to Myanforn into long sentence pairs and short sentence pairs according to sentence lengths; constructing a preliminary tree-shaped retrieval structure through the short sentence pair; inserting the long sentence pair into a tree-shaped retrieval structure; in the retrieval stage, firstly, a to-be-translated text is coded through an LASER model and then matched with related words and sentence pairs in a tree retrieval structure, candidate sentence pairs are selected through a BM25 algorithm, and the cosine similarity between the candidate sentence pairs and the to-be-translated text is further calculated; before the large language model is used for final translation, the text reordering model is used for reordering the context prompt sentence pairs obtained through retrieval, and the context prompt sentence pairs and the to-be-translated text are input into the large language model to obtain a final translation result. The accuracy and smoothness of the translation result can be improved.

Taking a translation task from Naxi Dongba to Chinese as an example, the invention provides a weak semantic low-resource character machine translation method based on semantic enhancement, which comprises the following steps: S1, designing a Naxi Dongba encoding system, and establishing a Naxi Dongba electronic dictionary; s2, a sufficient number of Naxi Dongba text-Chinese parallel sentence pairs are collected and marked, and a Naxi Dongba text-Chinese parallel corpus is constructed; s3, dividing the data set into a fine adjustment data set and a test data set, and further dividing the fine adjustment data set into a training set and a verification set; s4, constructing a semantic enhancement model based on fine tuning and custom word list embedding; s5, providing an iterative reverse translation method combined with word replacement, and constructing an extended data set; s6, constructing a weak semantic low-resource text machine translation model based on semantic enhancement, adopting an increment updating mechanism, taking the high-quality pseudo-parallel corpus generated in the step S5 as increment, inputting the increment into the semantic enhancement model in the step S3, and adjusting and optimizing the weight of the model through parameters; and S7, inputting the Naxi Dongba coded sentences to be translated into the updated model for translation, and outputting a result. According to the method, translation research from the Naxi Dongba text to Chinese is carried out based on traditional expert experience, automatic translation of the Naxi Dongba text can be achieved, meanwhile, the method has the capacity of continuous learning and adapting to new data, the machine translation effect of weak-semantic low-resource characters is improved, and technical support is provided for research in related fields.

This application proposes a training method, a Chinese sentence simplification method, and a device for the Chinese sentence simplification model. The training method for the Chinese sentence simplification model includes: obtaining a dataset of complex sentence-simple sentence pairs containing supervisory signals and a Chinese monolingual pre-training model; selecting a simple sentence from the current complex sentence-simple sentence pair as a positive sample in each training batch, and randomly selecting a preset number of simple sentences from other sentence pairs in the same training batch as negative samples; projecting the complex sentence, positive sample, and negative sample into a vector representation space, and obtaining the hidden layer vectors of the last layer of the encoder respectively; calculating the contrastive learning loss, and calculating the cross-entropy loss of the desired simple sentence through the decoder; and jointly training the Chinese monolingual pre-training model by minimizing the contrastive learning loss and cross-entropy loss of the simple sentence output by the Chinese monolingual pre-training model. The simplified model obtained by this method can improve the controllability and fidelity of the generated simplified sentences.

The application discloses a bidirectional information retrieval enhancement generation method for a large language model, and belongs to the technical field of artificial intelligence. In order to overcome the defects of traditional RAG that only performs one-way retrieval from 'query to document' and introduces noise and omits key evidence, a bidirectional semantic perception retrieval enhancement generation model and a two-stage training framework are constructed. In the first stage, the positive / negative example distance is pulled apart in the embedding space in a contrast learning self-supervised manner. In the second stage, a supervised two-classification is used to finely distinguish the relevance of the query-document bidirectional sentence pair, and an output probability correlation is obtained. In the reasoning stage, the bidirectional probability is fused according to Bayes to obtain the final correlation degree, and the document is reordered. The whole process can realize plug and play without fine-tuning the LLM. The application significantly improves the accuracy and consistency of single-hop, multi-hop question answering and fact checking tasks, and has the advantages of light weight and low deployment cost.

The invention provides a sentence semantic matching method and device for sentence subject interaction, and relates to the technical field of natural language processing. The method comprises the following steps of: training a sentence semantic matching model, namely extracting a subject of a to-be-matched sentence pair by a feature aggregation network to obtain a distillation subject feature, extracting a keyword of the to-be-matched sentence pair by a subject extraction layer to obtain a subject information sequence, extracting a subject of the subject information sequence by the feature aggregation network to obtain an optimal subject feature, the classification layer carries out classification according to the distillation objective characteristic and the optimal objective characteristic to obtain a matching label; and respectively calculating loss values of the distillation pursuit feature and the optimal pursuit feature, and optimizing a sentence semantic matching model according to the loss values. According to the method, the semantic relation between sentences is comprehensively captured by deeply fusing the extracted contextual subject features of the sentences and the optimal subject features, the similarity of sentence pairs in overall semantics is more accurately judged, matching tags are generated, and an optimized sentence semantic matching model is trained.

The application belongs to the field of databases and provides a code conversion method, device, equipment and medium based on statement pair equivalent replacement, which comprises the following steps: obtaining a to-be-converted statement including a first operator and a first logical statement in PLSQL code to be converted, the first operator being used for performing a logical NOT operation, and the logical operation result of the first logical statement being Null; obtaining a target logical operation result of the to-be-converted statement; removing the first operator and converting the first logical statement into a second logical statement with a logical operation result of the target logical operation result; and replacing the second logical statement with the to-be-converted statement to convert the obtained target PLSQL code into target Java code. According to the technical scheme of the embodiment, when the logical operation result of the first logical statement in the to-be-converted statement is Null, the to-be-converted statement can be replaced by the second logical statement in a pair-equivalent manner, so that the operation logic remains unchanged after being converted into a Java statement and the normal operation of the system is ensured.

The invention provides a subtitle processing method and system, electronic equipment and a storage medium, and is applied to the technical field of data processing. A plurality of sentences and original subtitle sentence durations thereof are generated according to a subtitle file; wherein the sentence is composed of at least one subtitle, and the original subtitle sentence duration is the sum of the original subtitle duration of all subtitles forming the sentence; translating the sentences to obtain translations of the sentences; based on the sentences and the duration of the original subtitle sentences, performing minimum deviation segmentation on translations of the sentences to obtain segmentation schemes of the sentences; wherein the segmentation scheme comprises a translation fragment of each subtitle corresponding to the sentence; according to the method and the device, the translation fragment of each subtitle is dubbed, and the dubbing rate of the dubbing of the translation fragment of each subtitle is adjusted based on the original subtitle duration of the subtitle, so that the aim of high-precision time sequence alignment is fulfilled under the conditions of avoiding sentence splitting and semantic incoherence and ensuring semantic integrity and a watching process.

The application provides an aspect-level sentiment analysis method and device based on contrast learning. The method comprises the following steps: S1, generating a plurality of prompt sentence pairs based on a plurality of preset aspect sentiment pairs; S2, inputting a to-be-analyzed sentence into an aspect-level sentiment analysis model based on contrast learning to obtain an analysis result, wherein the aspect-level sentiment analysis model comprises: an enhancement module, which combines the to-be-analyzed sentence with a question or answer prompt sentence in different prompt sentence pairs to obtain different to-be-analyzed enhanced sentences; a pre-training coding layer, which obtains a sentence representation vector and a word vector of the to-be-analyzed enhanced sentence; a first activation function layer, which obtains a first analysis result; and a second activation function layer, which marks the position of a target word matched with the aspect sentiment pair to obtain a marked sequence. The joint detection of the target, the aspect and the sentiment is realized, high-quality semantic information can be generated after the to-be-analyzed sentence is enhanced by using the question or answer prompt sentence, the problem of sparse marked data is effectively alleviated, and the sentiment analysis effect is improved.

This invention relates to a method and system for optimizing machine translation using a large language model, belonging to the field of machine translation technology. It includes: generating a corresponding first translation from a source sentence in a bilingual corpus; segmenting the source sentence into words, counting the frequency of easily misspelled words in each segment, and calculating the easily misspelled word score, obtaining an easily misspelled word set based on the score; calculating the semantic similarity between the sentence to be translated and multiple candidate examples, and calculating the quality scores of the multiple candidate examples; selecting the optimal k candidate examples from the multiple candidate examples based on semantic similarity and quality scores, and constructing them as prompt templates; obtaining training sentence pairs from the bilingual corpus based on the easily misspelled word scores to construct a training set; and using the training set and prompt templates to perform low-rank adaptive training on a large language model to obtain an optimized large language model. This invention not only improves translation quality but also enhances the interpretability of the translation process.

The invention discloses a subject information-based field bilingual sentence pair selection method, which is used for selecting a sentence pair subset related to a to-be-translated text from a large-scale bilingual corpus mixed with fields by virtue of subject correlation between a bilingual sentence pair and a target field so as to train a specific field translation system and improve the translation quality of the field text. The method comprises the following steps: firstly, learning topic vectors of phrase pairs by using context words of the phrase pairs in a bilingual corpus; secondly, for a target domain development set and a candidate bilingual sentence pair, obtaining topic vectors of the target domain development set and the candidate bilingual sentence pair by utilizing the extracted phrase pair set; and finally, the topic relevancy between the candidate bilingual sentence pairs and the field development set text is calculated, and the sentence pairs with high relevancy are preferentially selected as target field training data. The invention further discloses a domain bilingual sentence pair selection system based on the theme information. According to the method, the field-related bilingual sentence pairs are selected by means of the topic relevancy of the text, and the problem of insufficient training data in a specific field is solved.

The present invention discloses a two-stage text summarization method based on the PEGASUS model and dynamic error correction. The method obtains text data to be processed, performs preprocessing, obtains various features, fuses them, and obtains sentence weights. Sentences are selected according to the sentence weights to select key sentences for the summary. Hierarchical clustering is performed on the key sentences. Potential unregistered words are identified, marked as incorrect words, and replaced with alternative words, and the alternative words are corrected by substitution loss. The replacement text is subjected to the PEGASUS model to calculate the contrast loss. The comprehensive loss is calculated by the substitution loss and the contrast loss, and the weight coefficients, dynamic thresholds, and similarity thresholds are adjusted until the comprehensive loss reaches a minimum or the number of iterations reaches a maximum, and then a summary is output. The unregistered words are replaced with different alternative words by the "dynamic error correction mechanism" to achieve the optimal selection of the summary and improve the accuracy of the generated summary.

To solve the problem in which there is a need for a technique which makes it even easier to assign tags to a text.SOLUTION: There is provided a program for operating a computer including a computer processor and configured to manage information of a plurality of tags assigned to a text in a storage unit. The program causes the computer processor to perform the steps of: generating an explanatory text describing the tags by providing a large language model with tag information; acquiring a text to which the tags are to be assigned; identifying tags corresponding to the acquired text on the basis of the acquired text and the generated explanatory text of the tags; and presenting the identified tags to a user in association with the acquired text.SELECTED DRAWING: Figure 12

The present invention relates to a natural language semantic extraction method and system, wherein the method comprises: performing word segmentation on a target document in units of sentences to obtain word segmentation units; analyzing whether multiple word segmentation units in a sentence constitute keywords; in response to multiple word segmentation units in a sentence constituting one or more keywords, extracting the sentences containing the keywords; performing grammatical analysis on the sentences containing the keywords to obtain a syntax tree; and extracting effective keywords and corresponding semantic tags from the syntax tree. The method and system provided by the present invention extract keywords that meet the requirements of information recommendation from the relevant target document content, and obtain effective keywords through grammatical analysis of the sentences containing the keywords, splitting and reconstructing the syntax tree, so as to obtain the true intention of the target document and improve the understanding of the user's target requirements.

The present application discloses a simultaneous translation delay detection method, related devices, electronic devices, and storage media. The simultaneous translation delay detection method includes: detecting the sentence simultaneous translation delay of each sentence pair during the simultaneous translation process; wherein, the sentence pair includes a first sentence in the source language and a second sentence in the target language, and the sentence simultaneous translation delay of the sentence pair includes the frame-level delay of the sentence pair; and statistically obtaining the passage simultaneous translation delay based on the sentence simultaneous translation delay. The above solution can automatically detect the simultaneous translation delay, which helps to significantly reduce the detection time and detection cost compared with manual detection.

The present invention relates to artificial intelligence and discloses a method for generating a text summary, including: performing sentence segmentation on historical reference articles to obtain a plurality of reference sentences; performing triple extraction and duplicate removal processing on the plurality of reference sentences to obtain a plurality of standard triples and performing label marking, using the data with label marking as a training data set, training a classification model using the training data set to obtain a standard classification model; inputting the article to be processed into the standard classification model to obtain a standard classification result; using the triples that meet the screening conditions in the standard classification result as target triples and performing triple splicing processing to obtain an input sequence, inputting the input sequence into a bidirectional long short-term memory network to obtain an article summary. In addition, the present invention also relates to blockchain technology, and the standard triples can be stored in the nodes of the blockchain. The present invention also proposes a text summary generating device, an electronic device, and a storage medium. The present invention can improve the accuracy of text summary generation.

The application discloses an intelligent substation virtual loop automatic checking method and system, which can be applied to the technical field of intelligent substations, obtains text descriptions of multiple virtual terminals in an intelligent substation virtual loop, then cleans each text description to obtain corresponding cleaned texts, encodes each cleaned text by using a target checking model trained based on an SBERT network model to obtain embedding vectors of each character, obtains corresponding sentence vectors based on the embedding vectors of the characters, determines multiple sentence pairs according to the sentence vectors, calculates the similarity between each sentence pair to obtain a text similarity value, and judges whether the text similarity is greater than a preset threshold value; if yes, each virtual terminal is determined to be matched, and the above method significantly improves the accuracy of automatic checking of the virtual terminals.

The invention relates to the technical field of natural language processing, and discloses a word segmentation method and device, equipment and a storage medium. The method comprises the following steps: for any sentence in a text, segmenting the sentence to obtain a plurality of word segmentation results corresponding to the sentence; for any word segmentation result in the plurality of word segmentation results, according to a word vector of any word segmentation in the word segmentation result, based on a frequency domain vector converted by the word vector and a font vector of the word segmentation, determining fusion features corresponding to the word segmentation; wherein the frequency domain vector of the segmented word is obtained by processing the word vector of the segmented word based on Fourier transform, and the font vector of the segmented word is obtained according to the stroke of the segmented word; and determining a target word segmentation result corresponding to the sentence from the plurality of word segmentation results according to the fusion feature corresponding to each word segmentation in each word segmentation result in the plurality of word segmentation results. Therefore, the accuracy of the word segmentation result can be improved.

The application provides a Chinese relation extraction method and system based on multi-modal semantic fusion, and relates to the technical field of information extraction. The method comprises: obtaining a Chinese sentence and entities corresponding to the Chinese sentence; extracting text semantics, shape semantics and structure semantics of each Chinese character in the Chinese sentence; constructing a multi-modal semantic fusion model through an improved Transformer network, encoding the shape semantics and the structure semantics respectively, splicing the encoded semantic features to obtain auxiliary features, taking the text semantics as main features, optimizing the feature distribution of the main features according to the correlation coefficient between the main features and the auxiliary features, and then obtaining the fused multi-modal semantic features; and determining the Chinese relation between the entities according to the multi-modal semantic features. In this way, the shape semantics and the structure semantics of the Chinese characters are used to enrich the context information of the Chinese sentence, which can reduce the influence of Chinese ambiguity in Chinese relation extraction and improve the Chinese relation extraction effect.