A target localization named entity recognition and classification system based on multi-layer feature fusion

Abstract

The present invention relates to the field of natural language processing technology, and discloses a target location named entity recognition and classification system based on multi-layer feature fusion, including: a full-text feature extraction module for extracting named entity features with context features to provide semantic information for subsequent processing. The high-level feature extraction module extracts multi-level semantic features from the text through local convolution operations and fuses these features to obtain text information of different granularities. Then, the fusion module fuses the context features and high-level features together to generate a comprehensive feature set, providing strong support for the recognition module. The recognition module locates the named entity based on these features and calculates the start position and length of the named entity. Finally, the classification module classifies the named entity using the location information, and feeds back the final results such as the location information, length, and category to the user through the output module. Furthermore, it can accurately determine the location, length, and category of each named entity in the target text.

Description

Technical Field

[0001] The present invention relates to the technical field of natural language processing, and in particular, to an object location named entity recognition and classification system based on multi-level feature fusion. Background Art

[0002] Named entity recognition (NER) is a fundamental and crucial task in natural language processing (NLP). Its main objective is to automatically identify entities with specific semantics from text, such as person names, place names, organization names, time, dates, quantities, currency units, etc. NER technology is widely used in fields such as information extraction, search engines, question answering systems, machine translation, sentiment analysis, etc., and is particularly significant when dealing with a large amount of unstructured text data. By identifying entities in text, NER lays the foundation for subsequent tasks such as semantic understanding, relation extraction, and knowledge graph construction.

[0003] Currently, traditional named entity recognition technologies are gradually being replaced by neural network-based models. These models can capture richer context information through training and learning, thereby improving the recognition accuracy. However, although existing deep learning methods have achieved certain success in processing standard text, there are still some challenges when dealing with complex text (such as text containing a large number of synonyms, context ambiguities, or irregular structures). For example, insufficient context information is a major problem faced by traditional named entity recognition technologies. Named entities often rely on context information to determine their specific meanings and categories. Secondly, most traditional named entity recognition technologies rely on manual feature extraction and simple dictionary matching, which limits their ability to express text. In deep learning methods, although word vectors are used to enhance the feature representation ability of the model, these methods still have problems, that is, they cannot effectively utilize feature information at multiple levels (such as character-level, word-level, sentence-level features, etc.) simultaneously, resulting in limitations in performance when dealing with complex text. In addition, in many actual application scenarios, named entities do not always appear in simple word forms, but may be nested in longer sentences or interfered by other entity information, resulting in difficult positioning. For such complex text, traditional named entity recognition technologies usually cannot obtain effective location information at an early stage, thus affecting the overall recognition accuracy.

[0004] Therefore, there is an urgent need to invent a named entity recognition and classification technology to solve the problems of context ambiguity, insufficient feature fusion, and difficult precise positioning of named entities in complex text in the prior art. Summary of the Invention

[0005] In view of this, the present invention proposes a target location named entity recognition and classification system based on multi-layer feature fusion, aiming to solve the problems of context ambiguity, insufficient feature fusion, and difficult accurate location of named entities in complex texts in the current technology.

[0006] The present invention proposes a target location named entity recognition and classification system based on multi-layer feature fusion, including:

[0007] A full-text feature extraction module, configured to extract named entity features with context features from the input text;

[0008] A high-level feature extraction module, configured to extract multi-level features in the input text based on local convolution, and extract text features corresponding to the granularity of each level of features. The high-level feature extraction module is also configured to fuse based on the text features corresponding to the granularity of each level of features to generate a comprehensive feature set;

[0009] A fusion module, electrically connected to the full-text feature extraction module and the high-level feature extraction module respectively. The fusion module is configured to splice and fuse between the context features and the comprehensive feature set to generate recognition features;

[0010] A recognition module, electrically connected to the fusion module. The recognition module is configured to perform location recognition of named entities in the input text based on the recognition features, and determine the start position offset and length of each named entity;

[0011] A classification module, electrically connected to the recognition module. The classification module is configured to classify named entities according to the results located by the recognition module, and determine the categories of each named entity;

[0012] An output module, electrically connected to the classification module. The output module is configured to output the position information, length, and category of each named entity in the input text.

[0013] Further, when the full-text feature extraction module is configured to extract named entity features with context features from the input text, it includes:

[0014] The full-text feature extraction module is also configured to perform word segmentation processing on the input text and generate a plurality of semantic unit sequences;

[0015] The full-text feature extraction module is also configured to encode a plurality of the semantic unit sequences and extract a feature sequence including context dependency relationships;

[0016] The full-text feature extraction module is further configured to obtain the dependencies between several semantic unit sequences based on the self-attention mechanism, and determine the named entity features with context features in the input text based on the dependencies between several semantic unit sequences and the feature sequence including context dependencies.

[0017] Further, the high-level feature extraction module includes:

[0018] The local feature extraction unit captures adjacent token features based on a one-dimensional convolutional network;

[0019] The multi-layer convolutional module extracts semantic features of different scales by gradually expanding the number of channels and downsampling layer by layer;

[0020] The normalization and activation module performs batch normalization and non-linear activation after the convolutional operation to enhance the feature expression ability.

[0021] Further, the fusion module includes:

[0022] The high-level feature downsampling unit is configured to perform downsampling on the high-level features generated by the high-level feature extraction module so that the dimension of the comprehensive feature set is consistent with that of the context features;

[0023] The feature splicing unit is configured to splice the context features and the downsampled high-level features along the feature dimension;

[0024] The weighted fusion unit is configured to fuse the spliced features by weighting and generate the recognition features.

[0025] Further, when the recognition module is configured to perform location recognition of named entities in the input text based on the recognition features and determine the start position offset and length of each named entity, it includes:

[0026] The recognition module is further configured to calculate the start position offset and length of the named entity through linear transformation;

[0027] y0 = (x center - x token ) + Δ offset ;

[0028] y1 = Length;

[0029] where y0 is the start position offset of the entity, y1 is the length of the entity, x center is the center point of the entity, x token is the center point of the text, and Δ offset is the offset;

[0030] The recognition module is further configured to determine the feature distribution of the named entity based on the start position offset and length of the named entity.

[0031] Further, when the classification module is configured to classify named entities according to the positioning result of the recognition module, it includes:

[0032] The classification module is further configured to calculate the class probability distribution using the softmax function according to the named entity positioning result provided by the recognition module:

[0033] y2 = p(ci|x)(Wx + b);

[0034] where y2 is the class probability distribution, W is the weight matrix, b is the bias, and P(ci∣x) is the mathematical function of the probability distribution, where:

[0035]

[0036] where zi is the score of the named entity ci, which is expressed as the exponential sum of the scores for all classes and is used for normalization so that the sum of the probabilities of all classes is 1, and n is the total number of classes;

[0037] The classification module is further configured to determine the class label of the named entity according to the class probability distribution and pass the classification result to the output module.

[0038] Further, when the output module is configured to output the position information, length, and category to which each named entity in the input text belongs, it includes:

[0039] The output module is further configured to obtain the input format of the downstream task and convert the start position offset and length of each named entity in the input text and the category information of each named entity into the input format of the downstream task for output.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: By setting up a full-text feature extraction module and a high-level feature extraction module, named entity features of the input text are extracted respectively from the context dependence relationship and multi-level semantic granularity, and these features are spliced and fused through a fusion module. This comprehensive processing method of multi-level features improves the accuracy of named entity recognition, enabling the system to capture entity features in a complex semantic environment more precisely. Secondly, the start position offset and length of the named entity are accurately located through the recognition module, and the category of the named entity is recognized with the help of the classification module. This modular design with function separation not only improves the efficiency of named entity recognition but also enhances the stability and generalization ability of the system when dealing with multi-category named entities, and is applicable to text processing tasks in complex scenarios. Finally, the design of the output module can directly output the position information, length, and category of the named entity, providing rich and intuitive data support for subsequent natural language processing tasks. This comprehensive information output method facilitates the integration of the system with other text analysis modules, improves the practicality and usability of the overall application, and is applicable to the named entity recognition requirements in multiple fields, such as information retrieval, question answering systems, and knowledge graph construction, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0042] Figure 1 FIG. is a functional block diagram of a target location named entity recognition and classification system based on multi-layer feature fusion provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] Hereinafter, exemplary embodiments of the present disclosure will be described in more detail with reference to the drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art. It should be noted that, without conflict, the embodiments and features in the embodiments of the present invention can be combined with each other. Hereinafter, the present invention will be described in detail with reference to the drawings and in combination with the embodiments.

[0044] As Figure 1As shown, in some embodiments of the present application, this embodiment provides a target location named entity recognition and classification system based on multi-layer feature fusion, including: a full-text feature extraction module, a high-level feature extraction module, a fusion module, an identification module, a classification module, and an output module.

[0045] Specifically, the full-text feature extraction module is configured to extract named entity features with context features from the input text; the high-level feature extraction module is configured to extract multi-level features in the input text based on local convolution, and extract text features corresponding to the granularity of each level of features. The high-level feature extraction module is also configured to fuse based on the text features corresponding to the granularity of each level of features to generate a comprehensive feature set; the fusion module is electrically connected to the full-text feature extraction module and the high-level feature extraction module respectively. The fusion module is configured to splice and fuse between the context features and the comprehensive feature set to generate recognition features; the identification module is electrically connected to the fusion module. The identification module is configured to perform location recognition of named entities in the input text based on the recognition features, and determine the start position offset and length of each named entity; the classification module is electrically connected to the identification module. The classification module is configured to classify the named entities according to the positioning results of the identification module, and determine the categories of each named entity; the output module is electrically connected to the classification module. The output module is configured to output the position information, length, and category of each named entity in the input text.

[0046] It can be understood that by the collaborative work of the full-text feature extraction module and the high-level feature extraction module, the features of named entities are extracted from the global context and local semantics respectively. The full-text feature extraction module uses context information to extract semantic dependency features of the text, while the high-level feature extraction module uses local convolution to perform multi-level processing on the input text, and extracts multi-dimensional semantic information from features of different granularities. Subsequently, the high-level feature extraction module fuses these multi-level features to generate a comprehensive feature set, thus achieving a comprehensive characterization of the text features. In addition, the fusion module integrates the context features and the comprehensive feature set through a splicing operation to generate unified recognition features. The identification module performs location recognition of named entities in the input text based on these recognition features, and determines the start position offset and length of the named entities in the text. This process of feature fusion and location recognition relies on the high expression ability of features and the spatial inference ability of the model, ensuring the accuracy of recognition. Finally, the classification module receives the positioning results provided by the identification module, combines the recognized feature information, and judges and classifies the category of each named entity. Finally, the output module outputs the position information, length, and category of each named entity recognized in the text. This process utilizes the category discrimination ability of the classification module and the intuitive presentation ability of the output module to complete the entire process from feature extraction to result output, realizing a closed-loop technical architecture for named entity recognition and classification.

[0047] It can be seen that the complete process from feature extraction to entity recognition and classification is achieved through multi-module collaboration. Among them, the full-text feature extraction module and the high-level feature extraction module are combined to effectively improve the comprehensive capture ability of text semantic information; the fusion module enhances the feature expression ability and recognition accuracy through the splicing and fusion of context features and multi-level features; the recognition module and the classification module locate named entities in sequence and perform precise classification, thus realizing the precise recognition and classification of named entities in the text; the output module presents the results in a structured manner for subsequent use.

[0048] Specifically, when the full-text feature extraction module is configured to extract named entity features with context features from the input text, it includes: the full-text feature extraction module is also configured to perform word segmentation on the input text and generate a number of semantic unit sequences; the full-text feature extraction module is also configured to encode the number of semantic unit sequences and extract a feature sequence containing context dependency relationships; the full-text feature extraction module is also configured to obtain the dependency relationships between a number of semantic unit sequences based on the self-attention mechanism, and determine the named entity features with context features in the input text based on the dependency relationships between a number of semantic unit sequences and the feature sequence containing context dependency relationships.

[0049] It can be understood that the input text is segmented into a number of semantic units (such as words or sub-words), which are the basic components of the text. In this way, long texts can be converted into smaller units that are easier to process, facilitating subsequent feature extraction and analysis. Secondly, the full-text feature extraction module encodes these semantic units and extracts a feature sequence containing context dependency relationships. This means that the module not only focuses on the features of a single semantic unit, but also captures the mutual dependencies and context relationships between them, thereby improving the accuracy of named entity recognition. The encoding process may involve methods such as Word2Vec and BERT to learn the semantic representations of vocabulary. Finally, the module uses the self-attention mechanism to obtain the dependency relationships between semantic units. The self-attention mechanism allows the model to consider the influence of all other words in the input sequence when processing each word, which helps to better understand the long-range dependency relationships in the text. By combining these dependency relationships and context information, the full-text feature extraction module can effectively determine the features of named entities, improving the effect and accuracy of named entity recognition.

[0050] Specifically: the high-level feature extraction module includes: a local feature extraction unit that captures adjacent token features based on a one-dimensional convolutional network; a multi-layer convolutional module that extracts semantic features of different scales by gradually expanding the number of channels and downsampling; a normalization and activation module that performs batch normalization and non-linear activation after convolutional operations to enhance the feature expression ability.

[0051] It can be understood that the local feature extraction unit captures the features of adjacent tokens based on a one-dimensional convolutional network. The convolution operation processes each local region in the input text step by step through a sliding window, thereby capturing the local relationships between adjacent tokens. The advantage of this method is that it can effectively extract the relationships between each word and its neighboring words, especially suitable for the word sequence features in language models. The local convolutional network is particularly suitable for dealing with short-term dependencies because it can capture the connections between adjacent words within a certain window range, helping the model understand the meaning of each word in a specific context. Next, the mentioned multi-layer convolutional module extracts semantic features of different scales by gradually expanding the number of channels and downsampling layer by layer. The design idea of the multi-layer convolution comes from the hierarchical feature learning of deep neural networks. Each layer of the network can further deeply learn more abstract and representative features based on the features extracted by the previous layer. By gradually expanding the number of channels, the model can form a more detailed and comprehensive semantic expression from low-level simple features to high-level complex features. At the same time, the downsampling operation can help reduce the computational burden by reducing the size of the feature map and retain the most important information, thereby improving the computational efficiency. Under the action of the multi-layer convolution, the model can adaptively capture semantic features at different levels and process complex context relationships. Finally, the role of the normalization and activation module is to stabilize the training process through batch normalization and enhance the expressive ability of features through non-linear activation functions. Batch Normalization mainly standardizes the input of each layer, making the data distribution consistent during the training process, thereby accelerating convergence and improving the stability of the model. Especially in deep neural networks, it can effectively alleviate the problems of gradient disappearance or gradient explosion. In addition, non-linear activation functions (such as ReLU, Leaky ReLU, etc.) introduce non-linear transformations, enabling the model to learn and represent more complex non-linear relationships and enhancing the fitting ability of the neural network to complex patterns. By combining the two, the convolutional neural network can better perform feature extraction and expression.

[0052] Specifically, the fusion module includes: a high-level feature downsampling unit configured to downsample the high-level features generated by the high-level feature extraction module so that the dimensions of the comprehensive feature set and the context features are consistent; a feature concatenation unit configured to concatenate the context features and the downsampled high-level features along the feature dimension; and a weighted fusion unit configured to fuse the concatenated features in a weighted manner and generate recognition features.

[0053] It can be understood that the high-level feature downsampling unit is configured to perform downsampling on the high-level features generated by the high-level feature extraction module so that the dimension of the comprehensive feature set is consistent with that of the context features. The main purpose of the downsampling operation is to reduce the dimension of the high-level features to match the dimension of the context features, thereby facilitating subsequent fusion processing. Through downsampling, the model can effectively reduce the computational burden while retaining the key information in the high-level features. The ways of downsampling usually include max pooling, average pooling, or dimensionality reduction operations through learning, so as to ensure the streamlining of information without losing key information. Next, the function of the feature concatenation unit is to concatenate the context features and the downsampled high-level features along the feature dimension. Feature concatenation is a commonly used method in multimodal fusion. By directly concatenating features from different sources in the dimension, various information can be combined into a more comprehensive feature representation. In this process, the model can utilize the complementarity of the context features and the high-level features, and integrate the advantages of both. This concatenation operation enables the model to utilize both low-level context information and high-level abstract features simultaneously, improving the accuracy and robustness of named entity recognition. Finally, the weighted fusion unit generates the final recognition features by performing weighted fusion on the concatenated features. The technical principle of weighted fusion is that by assigning different weights to different features, the model can automatically learn the importance of each feature in the named entity recognition task. This weighted strategy enables the model to handle the importance of different features more flexibly, and can focus on certain features according to the needs of the task, thereby improving the overall recognition performance. Weighted fusion can fully exploit the potential of various features and form recognition features with high expressive ability

[0054] Specifically, when the recognition module is configured to perform localization and recognition of named entities in the input text based on the recognition features and determine the start position offset and length of each named entity, it includes: The recognition module is also configured to calculate the start position offset and length of the named entity through linear transformation;

[0055] y0 = (x center - x token ) + Δ offset ;

[0056] y1 = Length;

[0057] Where y0 is the start position offset of the entity, y1 is the length of the entity, x center is the center point of the entity, x token is the center point of the text, and Δ offset is the offset; The recognition module is also configured to determine the feature distribution of the named entity based on the start position offset and length of the named entity.

[0058] It can be understood that the starting position offset and length of the named entity are calculated through linear transformation. This process adopts a positioning method based on position offset and length to accurately determine the position of the named entity in the text. In the formula, y0 represents the starting position offset of the entity, y1 represents the length of the entity, and by calculating the distance between the center point of the entity and the center point of the text, and combining with the offset Δoffset, the model can accurately obtain the position of the named entity in the input text. Next, y0 (the starting position offset of the entity) is obtained through the calculation formula y0 = (xcenter - xtoken) + Δoffset. Here, xcenter represents the center point of the named entity, xtoken represents the center point in the text, and Δoffset is an adjustable offset used to adjust the position of the named entity. This linear transformation ensures the accurate positioning of the starting position of the named entity by measuring the distance between the center points and adjusting according to the offset. In this way, the model can capture the subtle changes in the relative position of the named entity in the text, improving the accuracy and reliability of recognition. Finally, the recognition module is also configured to further determine the feature distribution of the named entity based on the starting position offset and length of the named entity. This means that in addition to positioning the position of the named entity, the model can further analyze the feature distribution of the named entity on the basis of determining the entity range. The principle of this operation is to analyze the relationship between the recognized named entity and the context in the text by using the offset and length information, so as to deeply understand the semantic features of the entity.

[0059] Specifically, when the classification module is configured to classify the named entity according to the positioning result of the recognition module, it includes: The classification module is also configured to calculate the category probability distribution using the softmax function according to the named entity positioning result provided by the recognition module:

[0060] y2 = p(ci|x)(Wx + b);

[0061] where y2 is the category probability distribution, W is the weight matrix, b is the bias, and P(ci∣x) is the mathematical function of the probability distribution, where:

[0062]

[0063] where zi is the score of the named entity ci, is expressed as the exponential sum of the scores of all categories, used for normalization, so that the sum of the probabilities of all categories is 1, and n is the total number of categories; the classification module is also configured to determine the category label of the named entity according to the category probability distribution and pass the classification result to the output module.

[0064] It can be understood that the classification module first obtains the location information of the named entity from the recognition module, including the start position and length of the named entity. Then, based on this location information, the classification module further processes the features of each named entity and calculates the scores of the entity belonging to different categories. By passing these scores to the softmax function, the module generates a normalized probability distribution, which can accurately represent the relative possibility of each category. This method can not only evaluate the possibility of a single category, but also help the system comprehensively judge which category is more representative of the named entity. Secondly, through the calculation of softmax, the classification module can select the category with the highest probability from multiple possible categories as the final category of the named entity. This process makes the classification decision highly accurate, because the softmax function itself can effectively avoid bias when dealing with multi-classification problems, ensuring the fairness and reasonableness of the classification results. More importantly, this method can also adapt to various complex text environments, deeply distinguish the subtle differences between multiple categories, thereby improving the robustness and performance of the classification system. Finally, the classification module passes the determined category label to the output module, and the output module is responsible for feeding back the final classification result to the user or other application systems. This process ensures the efficient operation of the entire process from the input text to the final classification result, and also enables the system to provide accurate named entity classification information in practical applications.

[0065] Specifically, when the output module is configured to output the location information, length, and category to which each named entity in the input text belongs, it includes: the output module is also configured to obtain the input format of the downstream task and convert the start position offset and length of each named entity in the input text and the category information of each named entity into the input format of the downstream task for output.

[0066] In the above embodiments, the present invention extracts the named entity features of the input text by setting up a full-text feature extraction module and a high-level feature extraction module, starting from the context dependency relationship and multi-level semantic granularity respectively, and splices and fuses these features through a fusion module. This comprehensive processing method of multi-level features improves the accuracy of named entity recognition, enabling the system to more precisely capture entity features in complex semantic environments. Secondly, the start position offset and length of the named entity are accurately located through the recognition module, and the category of the named entity is recognized with the help of the classification module. This modular design with function separation not only improves the efficiency of named entity recognition but also enhances the stability and generalization ability of the system when dealing with multi-category named entities, and is applicable to text processing tasks in complex scenarios. Finally, the design of the output module can directly output the position information, length, and category of the named entity, providing rich and intuitive data support for subsequent natural language processing tasks. This comprehensive information output method facilitates the integration of the system with other text analysis modules, improves the practicality and usability of the overall application, and is applicable to the named entity recognition requirements in multiple fields, such as information retrieval, question answering systems, and knowledge graph construction.

[0067] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0068] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one or more of the flows or multiple flows and / or blocks Figure 1 one or more of the blocks or multiple blocks.

[0069] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions in the process Figure 1one or more processes and / or blocks Figure 1 the functions specified in one or more blocks.

[0070] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more processes and / or blocks Figure 1 the functions specified in one or more blocks.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific implementation manners of the present invention. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A target location named entity recognition and classification system based on multi-layer feature fusion, characterized in that, Including: A full-text feature extraction module, configured to extract named entity features with context features from the input text; A high-level feature extraction module, configured to extract multi-level features in the input text based on local convolution and extract text features corresponding to the granularity of each level of features. The high-level feature extraction module is also configured to fuse based on the text features corresponding to the granularity of each level of features to generate a comprehensive feature set; A fusion module, electrically connected to the full-text feature extraction module and the high-level feature extraction module respectively. The fusion module is configured to splice and fuse between the context features and the comprehensive feature set to generate recognition features; A recognition module, electrically connected to the fusion module. The recognition module is configured to perform localization recognition of named entities in the input text based on the recognition features and determine the start position offset and length of each named entity; A classification module, electrically connected to the recognition module. The classification module is configured to classify the named entities according to the positioning result of the recognition module and determine the category of each named entity; An output module, electrically connected to the classification module. The output module is configured to output the position information, length, and category of each named entity in the input text.

2. The object localization named entity recognition and classification system based on multi-layer feature fusion according to claim 1, characterized in that, When the full-text feature extraction module is configured to extract named entity features with context features from the input text, it includes: The full-text feature extraction module is also configured to perform word segmentation on the input text and generate a number of semantic unit sequences; The full-text feature extraction module is also configured to encode a number of the semantic unit sequences and extract a feature sequence including context dependency relationships; The full-text feature extraction module is also configured to obtain the dependency relationships between a number of the semantic unit sequences based on the self-attention mechanism, and determine the named entity features with the context features in the input text based on the dependency relationships between a number of the semantic unit sequences and the feature sequence including context dependency relationships.

3. The target location named entity recognition and classification system based on multi-layer feature fusion according to claim 1, characterized in that, The high-level feature extraction module includes: A local feature extraction unit, which captures adjacent token features based on a one-dimensional convolutional network; A multi-layer convolutional module, which extracts semantic features of different scales by gradually expanding the number of channels and downsampling layer by layer; A normalization and activation module, which performs batch normalization and non-linear activation after the convolutional operation to enhance the feature expression ability.

4. The object - location named - entity recognition and classification system based on multi - layer feature fusion according to claim 1, wherein The fusion module includes: A high-level feature downsampling unit, configured to perform downsampling processing on the high-level features generated by the high-level feature extraction module so that the dimension of the comprehensive feature set is consistent with that of the context features; A feature splicing unit, configured to splice the context features and the downsampled high-level features along the feature dimension; A weighted fusion unit, configured to fuse the spliced features in a weighted manner and generate the recognition features.

5. The target location named entity recognition and classification system based on multi-layer feature fusion according to claim 1, characterized in that, When the recognition module is configured to perform localization recognition of named entities in the input text based on the recognition features and determine the start position offset and length of each named entity, it includes: The recognition module is also configured to calculate the start position offset and length of the named entity through linear transformation; y0 = (x center - x token ) + Δ offset ; y1 = Length; Among them, y0 is the offset of the start position of the entity, y1 is the length of the entity, and x center is the center point of the entity, and x token is the center point of the text, and Δ offset is the offset; The recognition module is further configured to determine the feature distribution of the named entity based on the start position offset and length of the named entity.

6. The target localization named entity recognition and classification system based on multi-layer feature fusion according to claim 1, characterized in that When the classification module is configured to classify the named entity according to the positioning result of the recognition module, it includes: The classification module is further configured to calculate the category probability distribution using the softmax function according to the named entity positioning result provided by the recognition module: y2 = p(ci|x)(Wx + b); where y2 is the category probability distribution, W is the weight matrix, b is the bias, and P(ci∣x) is the mathematical function of the probability distribution, where: where zi is the score of the named entity ci, which is expressed as the exponential sum of the scores for all categories and is used for normalization so that the sum of the probabilities for all categories is 1, and n is the total number of categories; The classification module is further configured to determine the category label of the named entity according to the category probability distribution and transfer the classification result to the output module.

7. The target location named entity recognition and classification system based on multi-layer feature fusion according to claim 6, characterized in that, When the output module is configured to output the position information, length, and category to which each named entity in the input text belongs, it includes: The output module is further configured to obtain the input format of the downstream task and convert the start position offset and length of each named entity in the input text and the category information of each named entity into the input format of the downstream task for output.

Images