Information extraction method and system, computing equipment and electronic equipment

Abstract

The invention discloses an information extraction method and system, computing equipment and electronic equipment, and relates to the field of large model technology and information processing. The method comprises the following steps: acquiring a to-be-extracted object which needs to be subjected to information extraction; based on a plurality of information extraction models corresponding to the to-be-extracted object, information extraction is carried out on the to-be-extracted object, multiple pieces of object information of the to-be-extracted object are obtained, different information extraction models are different in application field, and the multiple pieces of object information correspond to the information extraction models; the multiple pieces of object information are fused on the basis of information weights corresponding to different information extraction models, target information of the to-be-extracted object is obtained, and the information weights are used for representing the correlation degree between the different information extraction models and the to-be-extracted object. According to the method and the device, the technical problem that the accuracy of extracting the selling point information of the specified object is relatively low in the related technology is solved.

Description

Technical Field

[0001] The present application relates to large model technology and information processing fields, and specifically, to an information extraction method, system, computing device and electronic device. Background Art

[0002] Currently, when users browse products, sellers usually show users the selling points of different products to attract users' purchasing interest. However, there may be a large number of different types of products in a product application, and some large models currently configured in the product application for extracting product selling points are often only applicable to the selling point extraction process of a single type or similar types of products. When it is necessary to extract selling points for other types of products, incorrect extraction results usually occur.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide an information extraction method, system, computing device, and electronic device to at least solve the technical problem of poor accuracy in extracting selling point information of a specified object in the related art.

[0005] According to one aspect of an embodiment of the present application, there is provided an information extraction method, including: obtaining an object to be extracted for which information extraction is required; extracting information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, to obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; fusing the multiple object information based on information weights corresponding to different information extraction models, to obtain target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted.

[0006] According to another aspect of an embodiment of the present application, another information extraction method is provided, including: responding to an input instruction on an operation interface, displaying an object to be extracted for which information extraction is required on the operation interface; responding to a processing instruction on the operation interface, displaying target information of the object to be extracted on the operation interface, wherein the target information is used to characterize information weights corresponding to different information extraction models, and information obtained by fusing multiple object information, and multiple object information is used to characterize information obtained by extracting information from the objects to be extracted based on multiple information extraction models, and different information extraction models have different application fields, and information weights are used to characterize the degree of association between different information extraction models and the objects to be extracted.

[0007] According to another aspect of an embodiment of the present application, another information extraction method is provided, including: obtaining an object to be extracted that needs to be extracted by calling a first interface, wherein the first interface includes a first parameter, and the parameter value of the first parameter includes the object to be extracted; extracting information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, and obtaining multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; fusing the multiple object information based on information weights corresponding to different information extraction models, and obtaining target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted; outputting the target information by calling a second interface, wherein the second interface includes a second parameter, and the parameter value of the second parameter includes the target information.

[0008] According to one aspect of an embodiment of the present application, an information extraction device is also provided, including: an object acquisition module, used to acquire an object to be extracted that requires information extraction; an information extraction module, used to extract information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, and obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; an information fusion module, used to fuse the multiple object information based on information weights corresponding to different information extraction models, and obtain target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted.

[0009] According to another aspect of an embodiment of the present application, another information extraction device is provided, including: a first display module, used to respond to input instructions acting on an operation interface, and display an object to be extracted that needs to be extracted on the operation interface; a second display module, used to respond to processing instructions acting on the operation interface, and display target information of the object to be extracted on the operation interface, wherein the target information is used to characterize information weights corresponding to different information extraction models, and information obtained by fusing multiple object information, and multiple object information is used to characterize information obtained by extracting information from the objects to be extracted based on multiple information extraction models, and different information extraction models have different application fields. Information weights are used to characterize the degree of association between different information extraction models and the objects to be extracted.

[0010] According to another aspect of an embodiment of the present application, an information extraction device is also provided, including: a first calling module, used to obtain an object to be extracted that needs to be extracted by calling a first interface, wherein the first interface includes a first parameter, and the parameter value of the first parameter includes the object to be extracted; an object extraction module, used to extract information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, and obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; an object fusion module, used to fuse the multiple object information based on information weights corresponding to different information extraction models, and obtain target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted; a second calling module, used to output target information by calling a second interface, wherein the second interface includes a second parameter, and the parameter value of the second parameter includes the target information.

[0011] According to another aspect of an embodiment of the present application, an information extraction system is also provided, including: a client, used to send an object to be extracted that requires information extraction; a server, connected to the client, and used to extract information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, to obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; based on the information weights corresponding to the different information extraction models, the multiple object information are merged to obtain the target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted; the client is also used to output the target information.

[0012] According to another aspect of the embodiments of the present application, a computing device is further provided, including: a memory storing an executable program; and a processor for running the program, wherein the method in each embodiment of the present application is executed when the program is running.

[0013] According to another aspect of the embodiments of the present application, an electronic device is also provided, including: a memory storing an executable program; a processor connected to the memory via a bus and used to run the program, wherein the method of each embodiment of the present application is executed when the program is running.

[0014] According to another aspect of an embodiment of the present application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in each embodiment of the present application.

[0015] According to another aspect of the embodiments of the present application, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present application is implemented.

[0016] According to another aspect of an embodiment of the present application, a computer program product is also provided, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present application is implemented.

[0017] According to another aspect of the embodiments of the present application, a computer program is also provided, and when the computer program is executed by a processor, the methods in the various embodiments of the present application are implemented.

[0018] In an embodiment of the present application, an object to be extracted that requires information extraction is obtained; information is extracted from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, so as to obtain multiple object information of the object to be extracted; multiple object information is fused based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted. By means of ensemble learning, object information corresponding to different information extraction models is fused according to the information weights corresponding to different information extraction models to improve the accuracy of the obtained target information, thereby solving the technical problem of poor accuracy in extracting selling point information of a specified object in the related art.

[0019] It is easy to notice that the above general description and the following detailed description are only for the purpose of exemplifying and explaining the present application, and do not constitute a limitation of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0021] Figure 1 is a schematic diagram of an application scenario of an information extraction method according to an embodiment of the present application;

[0022] Figure 2 is a flow chart of an information extraction method according to an embodiment of the present application;

[0023] Figure 3 is a schematic diagram of an information extraction process according to an embodiment of the present application;

[0024] Figure 4 is a flow chart of another information extraction method according to an embodiment of the present application;

[0025] Figure 5 is a flow chart of another information extraction method according to an embodiment of the present application;

[0026] Figure 6 is a structural block diagram of an information extraction device according to an embodiment of the present application;

[0027] Figure 7 is a structural block diagram of another information extraction device according to an embodiment of the present application;

[0028] Figure 8 is a structural block diagram of another information extraction device according to an embodiment of the present application;

[0029] Fig. 9 is a structural block diagram of an information extraction system according to an embodiment of the present application;

[0030] Fig.10 is a structural block diagram of a computing device according to an embodiment of the present application;

[0031] Fig.11 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0033] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0034] The technical solution provided in this application is mainly implemented by large-scale model technology. The large model here refers to a deep learning model with large-scale model parameters, which can usually contain hundreds of millions, tens of billions, hundreds of billions, trillions or even more than 10 trillion model parameters. The large model can also be called a foundation model / foundation model. The large model is pre-trained through large-scale unlabeled corpus to produce a pre-trained model with more than 100 million parameters. This model can adapt to a wide range of downstream tasks, and the model has good generalization ability, such as large-scale language model (Large Language Model, LLM), multi-modal pre-training model (multi-modal pre-training model), etc.

[0035] It should be noted that, when the large model is actually applied, the pre-trained model can be fine-tuned by a small number of samples, so that the large model can be applied to different tasks. For example, the large model can be widely used in natural language processing (Natural Language Processing, referred to as NLP), computer vision, speech processing and other fields, and can be specifically applied to computer vision tasks such as visual question answering (Visual Question Answering, referred to as VQA), image description (Image Capt ion, referred to as IC), image generation, etc., and can also be widely used in text-based sentiment classification, text summary generation, machine translation and other natural language processing tasks. Therefore, the main application scenarios of the large model include but are not limited to digital assistants, intelligent robots, search, online education, office software, e-commerce, intelligent design, etc. In the embodiment of the present application, the data processing is performed by multiple information extraction models and integrated learning fusion models in the information extraction scenario as an example for explanation.

[0036] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following explanations:

[0037] Ensemble learning: Improve the overall prediction performance and robustness by combining the prediction results of multiple models.

[0038] Big Language Model: A deep learning-based AI model specifically designed for processing and generating natural language text.

[0039] According to an embodiment of the present application, an information extraction method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0040] Considering the huge number of model parameters of the large model and the limited computing resources of the mobile terminal, the above method provided in the embodiment of the present application can be applied to Figure 1 The application scenarios shown are not limited to these. Figure 1 is a schematic diagram of an application scenario of an information extraction method according to an embodiment of the present application. Figure 1 In the application scenario shown, the large model is deployed in the server 10, and the server 10 can be connected to one or more client devices 20 through a local area network connection, a wide area network connection, an Internet connection, or other types of data networks. The client device 20 here may include but is not limited to: a smart phone, a tablet computer, a laptop computer, a PDA, a personal computer, a smart home device, a vehicle-mounted device, etc. The client device 20 can interact with the user through a graphical user interface to implement the call of the large model, thereby implementing the method provided in the embodiment of the present application.

[0041] In an embodiment of the present application, a system composed of a client device and a server can perform the following steps: the client device executes obtaining an object to be extracted that needs information extraction. The server executes information extraction on the object to be extracted based on multiple information extraction models corresponding to the object to be extracted; and multiple object information is merged based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted.

[0042] It should be noted that, with the rapid development of high-performance computing units, in other application scenarios, the above method provided in the embodiment of the present application can also be applied to the model all-in-one machine. In an optional embodiment, a plurality of models are built into the model all-in-one machine, and the user can choose to adjust a model as needed to obtain the user's own model, so that the high-performance computing unit built into the model all-in-one machine can directly call the adjusted model to execute the above method provided in the embodiment of the present application. In another optional embodiment, a trained model is built into the large model all-in-one machine, so that the high-performance computing unit built into the model all-in-one machine can directly call the model to execute the above method provided in the embodiment of the present application.

[0043] Furthermore, when users need to train their own models, they can also upload their own data sets through the client, which are sent by the client to the server, so that the server can adjust the pre-trained model with the data set to obtain the user's own model and then deploy it to the production environment. In order to facilitate users' needs for model adjustment, the server can provide complete adjustment tools, development frameworks and processes, and can support multiple adjustment strategies, so that the adjusted model can better adapt to applications in different fields and achieve a high degree of customization.

[0044] Under the above operating environment, this application provides Figure 2The information extraction method shown. Figure 2 FIG. 1 is a flow chart of an information extraction method according to an embodiment of the present application. Figure 2 As shown, the method may include the following steps:

[0045] Step S202, obtaining the object to be extracted for which information extraction is required.

[0046] The above-mentioned objects to be extracted may refer to objects with certain information value, and may include but are not limited to: physical goods in real scenes such as clothing, vehicles, furniture, and virtual goods in virtual environments such as digital collections, accounts, and online courses.

[0047] In an optional scheme of the present embodiment, considering that when a user browses an object, such as purchasing a product, the product recommendation system will usually recommend a large number of products that are related to the user's needs to the user at one time. The corresponding user can browse the information presented by these products, such as videos, pictures, texts, other users' comments and other information, and choose the desired products by themselves. In order to accurately attract the user's purchasing interest and improve the user's shopping experience, the recommendation system will usually display the key information that can demonstrate the value of these products, such as the characteristics and selling points of these products, to the user, so that the user can quickly understand the product to determine whether the product can meet their own needs, thereby improving the user's experience in purchasing products. Therefore, how to accurately extract product information related to user needs from the products is one of the key points of recommending products to users. Considering that a product may have a lot of information, and the area used to display product information in the operation interface may be relatively limited, and the user's browsing needs may be different each time, therefore, when displaying products, the products that actually need to be extracted may be products that can meet the user's interests or needs, and the information extracted and displayed from these products may also be product information related to the user's interests and needs. Correspondingly, the information extraction system can first determine the objects to be extracted that currently need to be extracted based on the user's interests, needs and other information, so as to ensure that the products and product information displayed to the user can meet the user's interests or needs as much as possible.

[0048] It should be noted that the above method of determining the objects to be extracted is only an exemplary display. In addition, the user can also actively specify the objects to be extracted for which information extraction is required. The information extraction system can also directly determine the products in the product library as the objects to be extracted for which information extraction is required. The specific method of obtaining the objects to be extracted can be decided by the user according to actual conditions and is not limited here.

[0049] Step S204: extract information from the objects to be extracted based on multiple information extraction models corresponding to the objects to be extracted, and obtain multiple pieces of object information of the objects to be extracted.

[0050] Among them, different information extraction models have different application fields, and multiple object information corresponds to multiple information extraction models respectively.

[0051] The above-mentioned multiple information extraction models may refer to models selected from different application fields, and the model architectures of different information extraction models have certain differences.

[0052] In an optional solution of the present embodiment, considering that the information fields to which the object information of different objects to be extracted may belong may be different, if a single information extraction model is used to extract information from different types of objects to be extracted, the extracted object information may be inaccurate. For example, when the commodity to be extracted is a clothing commodity, the information extraction system can use the corresponding clothing information extraction model to extract information from the clothing commodity to obtain commodity information with higher accuracy, such as clothing style, texture, origin, user evaluation and other information. When the commodity to be extracted is a vehicle commodity, the clothing information extraction model is continued to be used to extract information from the vehicle commodity, and the corresponding extracted vehicle information will be more inclined to the description content of clothing, such as the color and style of the vehicle surface, and deviate from the structural information of the vehicle itself, such as the vehicle model, the characteristics of the vehicle-mounted equipment and other information. In addition, considering that the color and style of the vehicle surface information are actually different from the color and style of the clothing surface information, there will be a large difference in expression, resulting in the use of the clothing information extraction model to extract information from the vehicle commodity, which will greatly affect the accuracy of the extraction result. Therefore, multiple information extraction models with different application fields can be configured in the information extraction system to avoid the situation where information cannot be accurately extracted using a single information extraction model. Correspondingly, when extracting information, the information extraction system can assign corresponding information extraction models to the objects to be extracted according to the types of the objects to be extracted, thereby improving the accuracy of information extraction for the objects to be extracted using the corresponding information extraction models.

[0053] Considering that even if the training accuracy of the information extraction model is very high, there may still be errors in the extraction of information during the actual application process. When the clothing information extraction model is used to extract the origin information of clothing products, the brand name may be mistakenly identified as the origin. However, the origin information extraction model designed for the origin information can accurately extract the corresponding clothing origin from the clothing products. Therefore, in order to further improve the accuracy of the extracted object information, the information extraction system can introduce the above-mentioned multiple information extraction models when extracting information from the extracted object, that is, use the above-mentioned multiple information extraction models to extract information from the extracted object, and obtain the object information output by different information extraction models to ensure that accurate object information can be included in these multiple object information. Further, as shown above, considering that users will purchase goods according to their own preferences or needs, therefore, when extracting information, the user's preferences or needs can be further introduced, so that the user can intuitively judge whether the object is needed through the extracted object information, thereby improving the user experience.

[0054] Step S206: fuse multiple object information based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted.

[0055] Among them, the information weight is used to characterize the degree of association between different information extraction models and the objects to be extracted.

[0056] In an optional scheme of the present embodiment, considering that different information extraction models may have different accuracy in extracting information from the object to be extracted, after obtaining the object information corresponding to different information extraction models, the information extraction model can further obtain information weights corresponding to different information extraction models, and use the obtained information weights to reflect the degree of association between different information extraction models and the object to be extracted, thereby reflecting the accuracy of different information extraction models in extracting information from the object to be extracted, and then use the obtained information weights to fuse the extracted multiple object information to obtain the target information of the object to be extracted with higher accuracy.

[0057] Furthermore, as shown above, considering that the extraction capabilities of different information extraction models in different information dimensions may be different, for example, model A has a higher recognition accuracy for the style of a product, model B has a higher recognition accuracy for the origin of a product, and model C has a higher extraction accuracy for user reviews of a product. Therefore, in order to ensure the accuracy when the above-mentioned multiple object information is fused, the above-mentioned information weights may include weights of different information extraction models in different information dimensions. Correspondingly, when the above-mentioned multiple object information is fused, these information weights can be used to fuse the information in different information dimensions in different object information. For example, the output result of the information extraction model with the largest weight in the corresponding information dimension is selected as the object information of the object to be extracted in the information dimension, or the output results of multiple information extraction models with larger weights in the corresponding information dimension are first selected, and then the same results in the selected output results are summarized, and the object information of the object to be extracted in the information dimension is determined according to the number of different results after summarization, so as to maximize the accuracy of the target information obtained by the above fusion.

[0058] To facilitate understanding of the above information extraction process, Figure 3 is a schematic diagram of an information extraction process according to an embodiment of the present application, such as Figure 3 As shown in FIG, when performing information extraction, the information extraction system can first obtain the information currently displayed by the object to be extracted, such as the image information and text information of the text to be extracted, and then input the obtained object to be extracted into multiple information extraction models, that is, Figure 3 In the large language models such as LLM-1, LLM-2, and LLM-N, these models are used to extract information from the extracted objects to obtain multiple object information, namely Figure 3 The object information such as Tag-1, Tag-2, Tag-N in the image is extracted, and finally the information weights corresponding to different information extraction models are used, that is, Figure 3 w1, w2, wN in , fuse the information of these multiple objects, so as to obtain the target information of the object to be extracted with higher accuracy. When performing information fusion, an integrated learning fusion model can be used to fuse the above multiple object information according to different information weights to improve the fusion efficiency and accuracy.

[0059] In an embodiment of the present application, an object to be extracted that requires information extraction is obtained; information is extracted from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, so as to obtain multiple object information of the object to be extracted; multiple object information is fused based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted. By means of ensemble learning, object information corresponding to different information extraction models is fused according to the information weights corresponding to different information extraction models to improve the accuracy of the obtained target information, thereby solving the technical problem of poor accuracy in extracting selling point information of a specified object in the related art.

[0060] In an embodiment of the present application, multiple object information is fused to obtain target information of the object to be extracted, including: performing correlation detection on the object to be extracted in multiple information extraction models respectively, and obtaining correlation detection results corresponding to different information extraction models; based on the correlation detection results, determining information weights corresponding to different information extraction models; inputting multiple information weights and multiple object information into an integrated learning fusion model, and using the integrated learning fusion model to fuse multiple object information to obtain target information.

[0061] In an optional solution of this embodiment, in order to accurately assign corresponding information weights to different information extraction models to ensure the accuracy of fusing multiple object information, when performing information fusion, the information extraction system can first perform correlation detection on the object to be extracted with different information extraction models, for example, by analyzing the extraction of different information dimensions by different information extraction models from multiple information dimensions of the object to be extracted, such as the accuracy of the extraction results of other objects in the corresponding information dimensions, to determine whether different information extraction models are suitable for information extraction of the object to be extracted, so as to obtain correlation detection results that can reflect the association between the object to be extracted and different information extraction models. Then, after obtaining the correlation detection results, the information weights corresponding to the different information extraction models are determined according to the correlation detection results corresponding to the different information extraction models, for example, a higher weight is assigned to the information extraction model with higher correlation, and a lower weight is assigned to the information extraction model with lower correlation, and finally, the object information extracted by the different information extraction models is fused using the determined information weights to obtain the corresponding target information. In order to improve the accuracy of the fused target information as much as possible, the above-mentioned integrated learning fusion model can be introduced during the fusion, and the above-mentioned multiple information weights and multiple object information can be input into the integrated learning fusion model, so as to use the integrated learning fusion model to fuse the multiple object information, thereby stably obtaining the target information with higher accuracy.

[0062] In an embodiment of the present application, multiple information weights and multiple object information are input into an integrated learning fusion model, and the multiple object information is fused using the integrated learning fusion model, including: using the integrated learning fusion model to fuse multiple object information according to multiple information weights to obtain initial information of the object to be extracted; in response to receiving an information adjustment instruction, adjusting the initial information based on the information adjustment instruction to obtain target information; in response to not receiving an information adjustment instruction, determining that the initial information is the target information.

[0063] In an optional scheme of the present embodiment, when using an integrated learning fusion model to fuse multiple object information, the information extraction system can first use the integrated learning fusion model to fuse multiple object information according to multiple information weights to obtain the initial information of the above-mentioned object to be extracted, and then output the initial information in a preset operation interface. The staff, that is, the personnel who verify the information of the object to be extracted, can adjust the fused initial information to ensure the accuracy of the initial information finally displayed to the user. The corresponding information extraction model can adjust the initial information according to the information adjustment instruction when receiving the information adjustment instruction to obtain accurate target information. If no information adjustment instruction is received, the initial information can be directly determined as the above-mentioned target information.

[0064] In an embodiment of the present application, the object to be extracted is subjected to correlation detection in multiple information extraction models respectively, and correlation detection results corresponding to different information extraction models are obtained, including: category identification is performed on the object to be extracted to obtain the object category to which the object to be extracted belongs; the object category is matched with the application fields of multiple information extraction models respectively to obtain multiple field matching results; the multiple field matching results are analyzed to obtain the confidence of different information extraction models when extracting information from the object to be extracted; based on the confidence, the correlation detection results corresponding to different information extraction models are constructed.

[0065] In an optional scheme of the present embodiment, in order to accurately determine the information weights corresponding to different information extraction models, when performing correlation detection on the object to be extracted and different information extraction models, the information extraction system can first perform category recognition on the object to be extracted to determine the object category of the object to be extracted, and then match the identified object category with the application field of different information extraction models to obtain corresponding multiple field brand results, such as determining whether the object category belongs to the application field, or whether the object category has a relationship with the application field. After obtaining multiple field matching results, the information extraction system can further analyze the multiple field matching results to obtain the confidence of different information extraction models when extracting information from the object to be extracted, that is, the reliability and credibility of different information extraction models when extracting information from objects of the object category, and finally construct the correlation detection results corresponding to different information extraction models based on the determined confidence.

[0066] In an embodiment of the present application, the above method also includes: performing feature extraction on the object to be extracted to obtain a description method for describing the object features of the object to be extracted; based on the description method, obtaining multiple corresponding initial extraction models from a model library; evaluating multiple initial extraction models from multiple evaluation dimensions to obtain model evaluation scores of different initial extraction models; based on the model evaluation scores, selecting multiple information extraction models from multiple initial extraction models.

[0067] In an optional scheme of the present embodiment, in order to avoid affecting the accuracy of the target information obtained by fusion, when selecting an information extraction model for extracting information of multiple objects, the information extraction system can first perform feature extraction on the object to be extracted to obtain a description method for describing the object features of the object to be extracted, such as text description, picture description, video description, etc., and then obtain multiple initial extraction models that match the extracted description method from the model library, and evaluate these multiple initial extraction models from multiple evaluation dimensions, such as accuracy, efficiency, recall rate, etc. of the initial extraction model to obtain model evaluation scores of different initial extraction models, and finally use the determined model evaluation scores to screen the multiple initial extraction models, such as selecting by proportion, selecting by quantity, etc., to screen and obtain the above-mentioned information extraction model.

[0068] In an embodiment of the present application, the above method also includes: obtaining multiple training samples and information extraction results corresponding to the multiple training samples from a full sample database, wherein the full sample database is used to store the information extraction results of different training samples; training multiple first extraction models based on multiple training samples to obtain multiple training extraction results corresponding to different first extraction models; adjusting different first extraction models based on multiple training extraction results and information extraction results corresponding to different training samples to obtain corresponding second extraction models, wherein the second extraction model includes multiple initial extraction models; and constructing a model library based on the second extraction models corresponding to different first extraction models.

[0069] In an optional scheme of the present embodiment, in order to be able to quickly select the above-mentioned multiple information extraction models, the information extraction system can first select multiple training samples and information extraction results corresponding to these multiple training samples from the full sample database, that is, a database used to store sample information of a large number of training samples and corresponding information extraction results, and then use these multiple training samples to train multiple models, that is, the above-mentioned first extraction models to obtain training extraction results output by different first extraction models, and then use the training extraction results corresponding to the above-mentioned multiple training samples and the information extraction results corresponding to different training samples to adjust different first models to obtain multiple second extraction models with higher accuracy, and finally use these multiple second extraction models to construct the above-mentioned model library. Correspondingly, the multiple trained second extraction models can include the above-mentioned multiple initial extraction models.

[0070] In an embodiment of the present application, the above method also includes: obtaining multiple first objects from an initial database, wherein the multiple first objects include multiple training samples; performing information annotation on the multiple first objects to obtain multiple information annotation results corresponding to different first objects, wherein the multiple information annotation results include information extraction results; performing data cleaning on the multiple information annotation results to obtain multiple target annotation results; based on the multiple target annotation results and multiple second objects, constructing a full sample database, wherein the multiple second objects are used to characterize objects in the multiple first objects that match the multiple information annotation results.

[0071] In an optional scheme of this embodiment, in constructing a full sample database, the information extraction system may first select multiple first objects from an initial database, such as a product library corresponding to a product shopping website, or a database actively provided by a staff member, and the multiple first objects may include the multiple training samples for model training mentioned above. After selecting multiple first objects, the information extraction system may further annotate the multiple first objects, such as by using an information annotation model or manually annotating them, to obtain multiple information annotation results corresponding to different first objects, and the corresponding multiple information annotation results may include information extraction results corresponding to different training samples. After annotating the multiple information annotation results corresponding to different first objects, the information extraction system may further clean the multiple information annotation results to obtain multiple target annotation results with higher accuracy, and select the second objects corresponding to the multiple target annotation results from the first objects, and finally use the multiple target annotation results and the corresponding second objects to construct the above-mentioned full sample database.

[0072] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0073] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0074] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0075] According to an embodiment of the present application, another information extraction method is also provided. Figure 4 is a flow chart of another information extraction method according to an embodiment of the present application, such as Figure 4 As shown, the steps of the method may include:

[0076] Step S402, in response to an input instruction on the operation interface, displaying objects to be extracted for which information extraction is required on the operation interface.

[0077] Step S404, in response to the processing instruction acting on the operation interface, the target information of the object to be extracted is displayed on the operation interface.

[0078] Among them, the target information is used to characterize the information weights corresponding to different information extraction models, and the information is obtained by fusing multiple object information. Multiple object information is used to characterize the information obtained by extracting information from the objects to be extracted based on multiple information extraction models. Different information extraction models have different application fields, and the information weights are used to characterize the degree of association between different information extraction models and the objects to be extracted.

[0079] In an optional scheme of the present embodiment, when an input instruction is detected, the information extraction system may first display the objects to be extracted that currently need to be extracted to the user in the operation interface, and then, when a processing instruction is detected, use multiple information extraction models to extract information from the objects to be extracted to obtain object information corresponding to different information extraction models, and finally use the information weights corresponding to different information extraction models to fuse the multiple object information to obtain target information with higher accuracy.

[0080] The specific information extraction process can be as shown above and will not be repeated here.

[0081] According to an embodiment of the present application, another information extraction method is also provided. Figure 5 is a flow chart of another information extraction method according to an embodiment of the present application, such as Figure 5 As shown, the steps of the method may include:

[0082] Step S502: acquiring an object to be extracted for which information extraction is required by calling a first interface.

[0083] The first interface includes a first parameter, and a parameter value of the first parameter includes an object to be extracted.

[0084] Step S504: extract information from the objects to be extracted based on multiple information extraction models corresponding to the objects to be extracted, and obtain multiple pieces of object information of the objects to be extracted.

[0085] Among them, different information extraction models have different application fields, and multiple object information corresponds to multiple information extraction models respectively.

[0086] Step S506: fuse multiple object information based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted.

[0087] Among them, the information weight is used to characterize the degree of association between different information extraction models and the objects to be extracted.

[0088] Step S508: output the target information by calling the second interface.

[0089] The second interface includes a second parameter, and a parameter value of the second parameter includes target information.

[0090] In an optional scheme of the present embodiment, the information extraction system can first obtain the first parameter by calling the first interface, that is, obtain the object to be extracted that currently needs to be extracted, and then use multiple information extraction models to extract information from the object to be extracted to obtain object information corresponding to different information extraction models, and finally use the information weights corresponding to different information extraction models to fuse the multiple object information to obtain target information with higher accuracy, and output the second parameter by calling the second interface, that is, output the target information, for user convenience to view.

[0091] The specific information extraction process can be as shown above and will not be repeated here.

[0092] According to an embodiment of the present application, there is also provided an information extraction device for implementing the above-mentioned information extraction method. Figure 6 is a structural block diagram of an information extraction device according to an embodiment of the present application. Figure 6 As shown, the device includes: an object acquisition module 602, an information extraction module 604 and an information fusion module 606.

[0093] Among them, the object acquisition module 602 is used to obtain the objects to be extracted that need to be extracted; the information extraction module 604 is used to extract information from the objects to be extracted based on multiple information extraction models corresponding to the objects to be extracted, and obtain multiple object information of the objects to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; the information fusion module 606 is used to fuse the multiple object information based on the information weights corresponding to different information extraction models, and obtain the target information of the objects to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the objects to be extracted.

[0094] Furthermore, the information fusion module 606 includes: a correlation detection unit, which is used to perform correlation detection on the objects to be extracted in multiple information extraction models respectively, and obtain correlation detection results corresponding to different information extraction models; a weight determination unit, which is used to determine the information weights corresponding to different information extraction models based on the correlation detection results; an information fusion unit, which is used to input multiple information weights and multiple object information into an integrated learning fusion model, and use the integrated learning fusion model to fuse multiple object information to obtain target information.

[0095] Furthermore, the information fusion unit is also used to: utilize an integrated learning fusion model to fuse multiple object information according to multiple information weights to obtain initial information of the object to be extracted; in response to receiving an information adjustment instruction, adjust the initial information based on the information adjustment instruction to obtain target information; in response to not receiving an information adjustment instruction, determine that the initial information is the target information.

[0096] Furthermore, the correlation detection unit is also used to: perform category recognition on the object to be extracted to obtain the object category to which the object to be extracted belongs; match the object category with the application fields of multiple information extraction models respectively to obtain multiple field matching results; analyze the multiple field matching results to obtain the confidence of different information extraction models when extracting information from the object to be extracted; based on the confidence, construct correlation detection results corresponding to different information extraction models.

[0097] Furthermore, the device also includes: a feature extraction module, which is used to extract features of the object to be extracted and obtain a description method for describing the object features of the object to be extracted; a model acquisition module, which is used to obtain corresponding multiple initial extraction models from a model library based on the description method; a model evaluation module, which is used to evaluate multiple initial extraction models from multiple evaluation dimensions to obtain model evaluation scores of different initial extraction models; and a model selection module, which is used to select multiple information extraction models from multiple initial extraction models based on the model evaluation scores.

[0098] Furthermore, the above-mentioned device also includes: a sample acquisition module, which is used to obtain multiple training samples and information extraction results corresponding to the multiple training samples from a full sample database, wherein the full sample database is used to store the information extraction results of different training samples; a model training module, which is used to train multiple first extraction models based on multiple training samples to obtain multiple training extraction results corresponding to different first extraction models; a model adjustment module, which is used to adjust different first extraction models based on multiple training extraction results and information extraction results corresponding to different training samples to obtain corresponding second extraction models, wherein the second extraction model includes multiple initial extraction models; a model library construction module, which is used to construct a model library based on the second extraction models corresponding to different first extraction models.

[0099] Furthermore, the above-mentioned device also includes: a first acquisition module, used to obtain multiple first objects from an initial database, wherein the multiple first objects include multiple training samples; an information labeling module, used to perform information labeling on the multiple first objects to obtain multiple information labeling results corresponding to different first objects, wherein the multiple information labeling results include information extraction results; a data cleaning module, used to perform data cleaning on the multiple information labeling results to obtain multiple target labeling results; a database construction module, used to construct a full sample database based on the multiple target labeling results and multiple second objects, wherein the multiple second objects are used to characterize objects in the multiple first objects that match the multiple information labeling results.

[0100] It should be noted that the object acquisition module 602, the information extraction module 604 and the information fusion module 606 correspond to steps S202 to S206 in the above embodiment, and the three modules and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules or units can be hardware components or software components stored in a memory and processed by one or more processors, and the above modules can also be run in the server 10 provided in the above embodiment as part of the device.

[0101] According to an embodiment of the present application, another information extraction device is also provided. Figure 7 is a structural block diagram of another information extraction device according to an embodiment of the present application, such as Figure 7 As shown, the device includes: a first display module 702 and a second display module 704 .

[0102] Among them, the first display module 702 is used to respond to input instructions on the operation interface, and display the objects to be extracted that need to be extracted on the operation interface; the second display module 704 is used to respond to processing instructions on the operation interface, and display the target information of the objects to be extracted on the operation interface, wherein the target information is used to characterize the information weights corresponding to different information extraction models, and the information is obtained by fusing multiple object information. The multiple object information is used to characterize the information obtained by extracting information from the objects to be extracted based on multiple information extraction models. Different information extraction models have different application fields, and the information weights are used to characterize the degree of association between different information extraction models and the objects to be extracted.

[0103] It should be noted that the first display module 702 and the second display module 704 correspond to steps S402 to S404 in the above embodiment, and the two modules and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules or units can be hardware components or software components stored in a memory and processed by one or more processors, and the above modules can also be run as part of the device in the server 10 provided in the above embodiment.

[0104] According to an embodiment of the present application, another information extraction device is also provided. Figure 8 is a structural block diagram of another information extraction device according to an embodiment of the present application, such as Figure 8 As shown, the device includes: a first calling module 802, an object extraction module 804, an object fusion module 806 and a second calling module 808.

[0105] Among them, the first calling module 802 is used to obtain the object to be extracted that needs to be extracted by calling the first interface, wherein the first interface includes a first parameter, and the parameter value of the first parameter includes the object to be extracted; the object extraction module 804 is used to extract information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, and obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively corresponds to the multiple information extraction models; the object fusion module 806 is used to fuse multiple object information based on information weights corresponding to different information extraction models, and obtain target information of the object to be extracted, wherein the information weight is used to characterize the degree of association between different information extraction models and the object to be extracted; the second calling module 808 is used to output target information by calling the second interface, wherein the second interface includes a second parameter, and the parameter value of the second parameter includes target information.

[0106] It should be noted that the first calling module 802, the object extraction module 804, the object fusion module 806 and the second calling module 808 correspond to steps S502 to S508 in the above embodiment, and the four modules and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules or units can be hardware components or software components stored in a memory and processed by one or more processors, and the above modules can also be run in the server 10 provided in the above embodiment as part of the device.

[0107] According to an embodiment of the present application, an information extraction system is also provided. Fig. 9 is a structural block diagram of an information extraction system according to an embodiment of the present application, such as Fig. 9As shown, the system includes: a client 902 and a server 904 .

[0108] Among them, the client 902 is used to send the objects to be extracted that need to be extracted; the server 904 is connected to the client, and is used to extract information from the objects to be extracted based on multiple information extraction models corresponding to the objects to be extracted, and obtain multiple object information of the objects to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively corresponds to the multiple information extraction models; based on the information weights corresponding to the different information extraction models, the multiple object information is merged to obtain the target information of the objects to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the objects to be extracted; the client 902 is also used to output the target information.

[0109] It should be noted that the preferred implementation scheme involved in the above embodiments of the present application is the same as the scheme provided in the above embodiments, as well as the application scenario and implementation process, but is not limited to the scheme provided in the above embodiments.

[0110] An embodiment of the present application may provide a computing device. Fig.10 is a structural block diagram of a computing device according to an embodiment of the present application. Fig.10 As shown, the computing device 1000 may include: one or more ( Fig.10 (only one is shown) processor 1002, memory 1004, storage controller, and peripheral interfaces.

[0111] The above-mentioned computing device can be understood as an integrated intelligent terminal, including but not limited to a server, a desktop computer, a PC (Personal Computer), a model all-in-one machine, etc., and the computing device can be pre-installed with the model in the above-mentioned embodiment of the present application.

[0112] Specifically, the computing device can pre-set multiple types of models, including but not limited to models in the fields of natural language processing, visual processing, speech processing, code processing, multimodal task processing, etc., so as to provide a variety of model choices. In different product forms, the computing device can support one or more model usage methods, including but not limited to model training, model calling, model fine-tuning, model deployment, model reasoning and application, etc. In some product forms, the computing device also supports model management, including but not limited to multi-type model management (supporting the management of multiple types of models such as discriminants and generative models), model version control (supporting the control of different model versions), model evaluation (based on model evaluation tools, evaluating the performance and effect of the model), etc. In other product forms, the computing device can also create applications based on the model, provide API calling capabilities, and can call the model to the created application through the API interface, while providing application management tools to achieve management and monitoring of the application.

[0113] Furthermore, the computing device can also include data management (supporting the creation and management of model tuning data sets), a training center (providing rich training resources to help users learn and master AI technology), and basic management and control capabilities (providing enterprise-level basic management and control capabilities to ensure the security and efficient operation of the system). Through the above functions, a comprehensive, integrated AI development, training, deployment and application device is provided.

[0114] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implementing the methods in the above embodiments. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal A via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0115] The processor may call the executable program stored in the memory through the transmission device to execute any method in the above embodiments.

[0116] An embodiment of the present application may provide an electronic device. Fig.11 1 is a structural block diagram of an electronic device according to an embodiment of the present application. Fig.11As shown, the electronic device may include: an input / output device 1102 ; a memory 1104 , and a processor 1106 , wherein the processor 1106 is connected to the input / output device 1102 and the memory 1104 via a bus 1108 .

[0117] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implementing the methods in the above embodiments. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal A via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0118] The processor may call the executable program stored in the memory through the transmission device to execute any method in the above embodiments.

[0119] Those skilled in the art will understand that Fig.11 The structure shown is for illustration only, and the computing device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, and other terminal devices. Fig.11 The structure of the computing device is not limited. For example, the computing device 100 may also include Fig.11 More or fewer components (such as network interfaces, display devices, etc.) shown in the figure, or having Fig.11 Different configurations are shown.

[0120] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0121] The embodiment of the present application further provides a computer-readable storage medium. Optionally, in this embodiment, the computer-readable storage medium can be used to store the program code executed by the method provided in the above embodiment.

[0122] Optionally, in this embodiment, the above storage medium may be located in a computing device.

[0123] Optionally, in this embodiment, the computer-readable storage medium is configured to store an executable program, and when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the methods in the above embodiments.

[0124] The embodiment of the present application further provides a computer program product. Optionally, in this embodiment, the computer program product may include a computer program, and the computer program implements the method provided in the embodiment when executed by a processor.

[0125] The embodiments of the present application also provide a computer program product. Optionally, the computer program product may include a non-volatile computer-readable storage medium, which may be used to store a computer program, and when the computer program is executed by a processor, the method provided in the embodiments is implemented.

[0126] The embodiment of the present application further provides a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, the method provided in the above embodiment is implemented.

[0127] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0128] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0129] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0130] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0131] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc., which can store program code.

[0132] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. An information extraction method, characterized in that: include: Get the object to be extracted that needs information extraction; Extracting information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, respectively, to obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; The multiple object information is fused based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted.

2. The method according to claim 1, characterized in that The fusing the plurality of object information to obtain the target information of the object to be extracted includes: Performing correlation detection on the object to be extracted and the multiple information extraction models respectively, and obtaining correlation detection results corresponding to different information extraction models; Based on the correlation detection result, determining the information weights corresponding to different information extraction models; The multiple information weights and the multiple object information are input into an integrated learning fusion model, and the multiple object information are fused using the integrated learning fusion model to obtain the target information.

3. The method according to claim 2, characterized in that The step of inputting the plurality of information weights and the plurality of object information into an integrated learning fusion model, and fusing the plurality of object information using the integrated learning fusion model, comprises: Using the integrated learning fusion model, the multiple object information is fused according to the multiple information weights to obtain the initial information of the object to be extracted; In response to receiving the information adjustment instruction, adjusting the initial information based on the information adjustment instruction to obtain the target information; In response to not receiving the information adjustment instruction, determining the initial information as the target information.

4. The method according to claim 2, characterized in that: The step of performing correlation detection on the object to be extracted and the plurality of information extraction models respectively to obtain correlation detection results corresponding to different information extraction models includes: Performing category recognition on the object to be extracted to obtain the object category to which the object to be extracted belongs; Matching the object categories with the application fields of the multiple information extraction models respectively to obtain multiple field matching results; Analyzing the multiple field matching results to obtain the confidence of different information extraction models when extracting information from the object to be extracted; Based on the confidence level, the correlation detection results corresponding to different information extraction models are constructed.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Extracting features of the object to be extracted to obtain a description method for describing the object features of the object to be extracted; Based on the description method, obtaining a corresponding plurality of initial extraction models from a model library; Evaluate the multiple initial extraction models from multiple evaluation dimensions to obtain model evaluation scores of different initial extraction models; Based on the model evaluation scores, the multiple information extraction models are selected from the multiple initial extraction models.

6. The method according to claim 5, characterized in that The method further comprises: Obtaining multiple training samples and information extraction results corresponding to the multiple training samples from a full sample database, wherein the full sample database is used to store information extraction results of different training samples; Training multiple first extraction models based on the multiple training samples to obtain multiple training extraction results corresponding to different first extraction models; Based on the multiple training extraction results and the information extraction results corresponding to different training samples, different first extraction models are adjusted to obtain corresponding second extraction models, wherein the second extraction model includes the multiple initial extraction models; The model library is constructed based on the second extraction models corresponding to different first extraction models.

7. The method according to claim 6, characterized in that The method further comprises: Acquire a plurality of first objects from an initial database, wherein the plurality of first objects include the plurality of training samples; Performing information annotation on the multiple first objects to obtain multiple information annotation results corresponding to different first objects, wherein the multiple information annotation results include the information extraction result; Performing data cleaning on the multiple information labeling results to obtain multiple target labeling results; The full sample database is constructed based on the multiple target annotation results and multiple second objects, wherein the multiple second objects are used to characterize objects among the multiple first objects that match the multiple information annotation results.

8. An information extraction method, characterized in that: include: In response to an input instruction acting on the operation interface, an object to be extracted that needs information extraction is displayed on the operation interface; In response to a processing instruction acting on the operation interface, the target information of the object to be extracted is displayed on the operation interface, wherein the target information is used to characterize information weights corresponding to different information extraction models, and information obtained by fusing multiple object information. The multiple object information is used to characterize information obtained by extracting information from the objects to be extracted based on multiple information extraction models. Different information extraction models have different application fields, and the information weights are used to characterize the degree of association between different information extraction models and the objects to be extracted.

9. An information extraction method, characterized in that: include: Acquire an object to be extracted that needs to be extracted by calling a first interface, wherein the first interface includes a first parameter, and a parameter value of the first parameter includes the object to be extracted; Extracting information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, respectively, to obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; The plurality of object information are fused based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted; The target information is output by calling a second interface, wherein the second interface includes a second parameter, and a parameter value of the second parameter includes the target information.

10. An information extraction system, characterized in that: include: The client is used to send the objects to be extracted; A server is connected to the client and is used to extract information from the object to be extracted based on multiple information extraction models corresponding to the object to be extracted, to obtain multiple object information of the object to be extracted, wherein different information extraction models have different application fields, and the multiple object information respectively correspond to the multiple information extraction models; the multiple object information are merged based on information weights corresponding to different information extraction models to obtain target information of the object to be extracted, wherein the information weights are used to characterize the degree of association between different information extraction models and the object to be extracted; The client is also used to output the target information.

11. A computing device, characterized in that: include: A memory storing an executable program; A processor, configured to run the program, wherein the program executes the method according to any one of claims 1 to 9 when running.

12. An electronic device, characterized in that: include: A memory storing an executable program; A processor, connected to the memory via a bus, and configured to run the program, wherein the program executes the method described in any one of claims 1 to 9 when running.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 9.

14. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 9.

Images