Information processing method and device, electronic equipment and storage medium

By creating a knowledge base and combining entity extraction and private keyword matching, the complexity and inefficiency of marking labeling information is solved, automatic marking is achieved, accuracy and efficiency are improved, and operational costs are reduced.

CN118939649APending Publication Date: 2024-11-12CHINA TELECOM CORP LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202410917609.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-09
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, the marking process of marking information is complicated and the processing efficiency is low, so it is impossible to effectively improve the accuracy and efficiency of marking.

Method used

By creating a knowledge base, combining entity extraction and private keyword matching, automatic marking of target information is achieved, marking efficiency is improved, and the accuracy of marking is further improved through entity extraction and private keyword matching.

Benefits of technology

It realizes automatic marking of benchmarking information, improves marking efficiency and accuracy, reduces operating costs, and supports question-and-answer query, greatly reducing the threshold for operators.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118939649A_ABST
    Figure CN118939649A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides an information processing method and device, electronic equipment and a storage medium. The information processing method comprises the following steps: performing entity extraction on target information to obtain entity information corresponding to the target information; matching the entity information corresponding to the target information with a private keyword contained in a pre-created knowledge base to obtain a private keyword corresponding to the target information; obtaining attribute information of a private keyword corresponding to the target information from a knowledge base, and generating label information corresponding to the target information according to the attribute information; and adding the entity information corresponding to the target information, the private keyword corresponding to the target information and the label information corresponding to the target information into the knowledge base. In the embodiment of the invention, automatic marking of the target information can be realized by creating the knowledge base and combining entity extraction and private keyword matching, the marking efficiency is improved, and the marking accuracy can be further improved by entity extraction and private keyword matching.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to an information processing method, device, electronic device and storage medium. Background Art

[0002] With the rapid development of economy, the bidding industry is also developing rapidly, and the country is vigorously supporting the application of electronic bidding. Today, the bidding information is public and the amount of information is huge and complicated, so it is very important to mark the bidding information, which helps to classify and manage the bidding information.

[0003] In the related art, the signal information is usually marked by manual marking, but this method has a complicated processing process and low processing efficiency. Summary of the invention

[0004] In view of the above problems, the embodiments of the present application propose an information processing method, device, electronic device and storage medium to improve the processing efficiency of information marking.

[0005] According to one aspect of an embodiment of the present application, there is provided an information processing method, the method comprising:

[0006] Performing entity extraction on the target information to obtain entity information corresponding to the target information;

[0007] Matching the entity information corresponding to the target information with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information;

[0008] Acquire attribute information of the private keyword corresponding to the target information from the knowledge base, and generate tag information corresponding to the target information according to the attribute information;

[0009] The entity information corresponding to the target information, the private keyword corresponding to the target information, and the tag information corresponding to the target information are added to the knowledge base.

[0010] Optionally, the method further includes: obtaining target information marked with exceptions within a set time period, obtaining a private keyword corresponding to the target information marked with exceptions from the knowledge base as a target private keyword; obtaining the number of target information items affected by the target private keyword; and when the number exceeds a preset number, setting the target private keyword to invalid.

[0011] Optionally, obtaining the number of target information items affected by the target private keyword includes: obtaining the number of target information items with marked anomalies corresponding to the target private keyword in the knowledge base as the number of target information items affected by the target private keyword.

[0012] Optionally, the method further includes: obtaining target information of marked anomalies within a set time period, obtaining entity information corresponding to the target information of marked anomalies from the knowledge base as candidate entity information; selecting target entity information from the candidate entity information according to the frequency of occurrence of the candidate entity information; obtaining private keywords to be added based on the target entity information, and adding the private keywords to be added to the knowledge base.

[0013] Optionally, obtaining the private keywords to be added according to the target entity information includes: combining multiple target entity information with the highest frequency in the target entity information to obtain combined entity information; matching the combined entity information with the target information with the marked exception, and using the successfully matched combined entity information as the target combined entity information; and obtaining the private keywords to be added according to the target entity information and the target combined entity information.

[0014] Optionally, acquiring the private keyword to be added according to the target entity information and the target combined entity information includes: determining the target combined entity information and target entity information that does not belong to the target combined entity information as the private keyword to be added.

[0015] Optionally, obtaining the private keywords to be added according to the target entity information and the target combined entity information includes: selecting the private keywords to be added from the target combined entity information and the target entity information that does not belong to the target combined entity information according to a preset priority order and a preset quantity; wherein the priority of the target combined entity information is higher than that of the target entity information.

[0016] Optionally, the method further includes: after receiving the query question, using a large language model to obtain a query answer corresponding to the query question based on the knowledge base, and returning the query answer.

[0017] Optionally, the method also includes: summarizing the target information in the knowledge base according to the corresponding label information to obtain summary information; using a large language model to obtain a query answer corresponding to the query question based on the knowledge base, including: using a large language model to obtain a query answer corresponding to the query question based on the summary information in the knowledge base; when obtaining the query answer fails, obtaining the query answer corresponding to the query question based on other information in the knowledge base except the summary information.

[0018] According to another aspect of an embodiment of the present application, there is provided an information processing device, the device comprising:

[0019] An extraction module, used to perform entity extraction on the target information to obtain entity information corresponding to the target information;

[0020] A matching module, used for matching the entity information corresponding to the target information with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information;

[0021] A generating module, used for acquiring attribute information of the private keyword corresponding to the target information from the knowledge base, and generating label information corresponding to the target information according to the attribute information;

[0022] The writing module is used to add the entity information corresponding to the target information, the private keyword corresponding to the target information and the tag information corresponding to the target information into the knowledge base.

[0023] Optionally, the device also includes: a first acquisition module, used to acquire target information with marking anomalies within a set time period, and acquire a private keyword corresponding to the target information with marking anomalies from the knowledge base as a target private keyword; a second acquisition module, used to acquire the number of target information items affected by the target private keyword; and a setting module, used to set the target private keyword to invalid when the number exceeds a preset number.

[0024] Optionally, the second acquisition module is specifically used to acquire the number of target information items with marked anomalies corresponding to the target private keyword in the knowledge base as the number of target information items affected by the target private keyword.

[0025] Optionally, the device also includes: a third acquisition module, used to acquire target information of marked anomalies within a set time period, and acquire entity information corresponding to the target information of marked anomalies from the knowledge base as candidate entity information; a fourth acquisition module, used to select target entity information from the candidate entity information according to the frequency of occurrence of the candidate entity information; and an adding module, used to acquire private keywords to be added according to the target entity information, and add the private keywords to be added to the knowledge base.

[0026] Optionally, the adding module includes: an information combination unit, used to combine multiple target entity information with the highest frequency of appearance in the target entity information to obtain combined entity information; an information matching unit, used to match the combined entity information with the target information with the marked exception, and use the successfully matched combined entity information as the target combined entity information; a keyword acquisition unit, used to acquire the private keywords to be added based on the target entity information and the target combined entity information.

[0027] Optionally, the keyword acquisition unit is specifically configured to determine the target composite entity information and target entity information that does not belong to the target composite entity information as private keywords to be added.

[0028] Optionally, the keyword acquisition unit is specifically used to select private keywords to be added from the target composite entity information and target entity information that does not belong to the target composite entity information according to a preset priority order and a preset quantity; wherein the priority of the target composite entity information is higher than that of the target entity information.

[0029] Optionally, the device further includes: a query module, which is used to, after receiving a query question, use a large language model to obtain a query answer corresponding to the query question based on the knowledge base, and return the query answer.

[0030] Optionally, the device also includes: a summary module, used to summarize the target information in the knowledge base according to the corresponding label information to obtain summary information; the query module is specifically used to use a large language model to obtain a query answer corresponding to the query question based on the summary information in the knowledge base; when the query answer fails to be obtained, the query answer corresponding to the query question is obtained based on other information in the knowledge base except the summary information.

[0031] According to another aspect of an embodiment of the present application, an electronic device is provided, comprising a processor and a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium; when the computer program is executed by the processor, the processor executes the information processing method as described in any one of the above items.

[0032] According to another aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the processor executes the information processing method as described in any one of the above items.

[0033] In an embodiment of the present application, entity extraction is performed on the target information to obtain entity information corresponding to the target information; the entity information corresponding to the target information is matched with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information; the attribute information of the private keywords corresponding to the target information is obtained from the knowledge base, and the label information corresponding to the target information is generated according to the attribute information; the entity information corresponding to the target information, the private keywords corresponding to the target information, and the label information corresponding to the target information are added to the knowledge base. It can be seen from this that in an embodiment of the present application, by creating a knowledge base, combining entity extraction and private keyword matching, automatic labeling of the target information can be achieved, the labeling efficiency can be improved, and the labeling accuracy can be further improved by entity extraction and private keyword matching.

[0034] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution of the embodiment of the present application, the drawings required for use in the description of the embodiment of the present application will be briefly introduced below. Obviously, the drawings described below are only some drawings of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 It is a step flow chart of an information processing method according to an embodiment of the present application.

[0037] Figure 2 It is a step flow chart of another information processing method of an embodiment of the present application.

[0038] Figure 3 It is a system architecture diagram of an embodiment of the present application.

[0039] Figure 4 It is a schematic diagram of a processing process of combining a dialogue language model with a language chain architecture according to an embodiment of the present application.

[0040] Figure 5 It is a schematic diagram of an information processing process of an embodiment of the present application.

[0041] Figure 6 It is a structural block diagram of an information processing device according to an embodiment of the present application.

[0042] Figure 7 It is a structural block diagram of an electronic device according to an embodiment of the present application.

[0043] Figure 8 It is a structural block diagram of a computer-readable storage medium in an embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, 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 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 are within the scope of protection of this application.

[0045] The information processing method in the embodiment of the present application can be applied to marking processing of various applicable target information.

[0046] Exemplarily, the target information may include, but is not limited to: bidding information, hotel information, shopping mall information, etc. For example, bidding information tagging usually involves sorting, classifying and marking bidding information so that users can quickly find and understand relevant information. The main purpose of bidding information tagging is to improve the searchability and readability of bidding information, so that users can quickly find bidding information that meets their needs. Through tagging, bidding information can be classified, sorted and marked, which is convenient for users to filter and compare.

[0047] In the embodiments of the present application, it is considered that in the process of processing the target information, if labeling is performed through simple regular matching, the labeling accuracy is low, and there is no self-learning and evolution capability, and personnel are required to continuously correct it, and some parts cannot be labeled, or all the labels are manually labeled, which is time-consuming and labor-intensive. In response to the above problems, the embodiments of the present application propose to create a knowledge base, combine entity extraction and private keyword matching, realize automatic labeling of target information, improve labeling efficiency, and through entity extraction and private keyword matching, the labeling accuracy can be further improved. The following is a detailed introduction.

[0048] Reference Figure 1 , showing a step flow chart of an information processing method in an embodiment of the present application.

[0049] like Figure 1 As shown, the information processing method may include the following steps:

[0050] Step 101: perform entity extraction on target information to obtain entity information corresponding to the target information.

[0051] In the embodiment of the present application, a knowledge base can be pre-built.

[0052] Exemplarily, the knowledge base may store private keywords (the private keywords refer to private keywords that match entity information corresponding to the target information), attribute information corresponding to the private keywords, entity information corresponding to the target information, private keywords corresponding to the target information, label information corresponding to the target information, summary information, etc. Of course, the knowledge base may also store basic industry information, regional information, friendly business information, etc., which is not limited in this embodiment.

[0053] Private keywords and the attribute information corresponding to them can be completed by batch analysis and precipitation of historical target information and manual supplementation, and then continuously updated and iterated through the operation mechanism. The attribute information corresponding to private keywords can include but is not limited to: industry, region, whether it is information-based, whether it is government-related, supplier, etc.

[0054] In the embodiment of the present application, the private keywords in the knowledge base may include private keywords of different categories, such as industry-related private keywords, government-related private keywords, information-related private keywords, etc. Each type of private keyword may correspond to its own single private keyword and combined private keyword, such as a single private keyword such as "government affairs", and a combined private keyword such as "government affairs||system", etc. The specific settings may be based on actual needs, and this embodiment does not limit this.

[0055] For example, the suppliers can be shown in Table 1 below:

[0056] Table 1

[0057] Operators Operator segmentation Supplier Name Operator 1 Operator 1 Company 1 Operator 1 Operator 1 Company 2 Operator 1 Operator 1 Company 3 Operator 2 Operator 2 (Integrated) Company 4 Operator 2 Operator 2 (Integrated) Company 5 Operator 2 Operator 2 (Construction) Company 6 Operator 3 Operator 3 (Digital) Company 7 Operator 4 Operator 4 Company 8 government government Government 1 …… …… ……

[0058] For example, whether informationization can be shown in Table 2 below:

[0059] Table 2

[0060] First level directory Secondary directory Keywords Whether it is informationized Data, information and communication convergence services other Equipment Management System yes Network access Network access optical fiber yes Data, information and communication convergence services Cloud Business Intelligent Networking yes Network access Network access Tower yes Internet of Things Internet of Things Monitoring system equipment yes Data, information and communication convergence services other Government Affairs||System yes …… …… …… ……

[0061] Among them, the data, information and communication convergence services can be DICT services (services that are the deep integration and application of Data Technology (DT), Information Technology (IT) and Communication Technology (CT)).

[0062] For example, the industry can be shown in Table 3 below:

[0063] Table 3

[0064] Industry Industry Keywords Matching rules Hospital %Hospital Industry keyword left fuzzy matching Health Center % Health Center Industry keyword left fuzzy matching Emergency Management Ministry of Emergency Management Industry keyword exact match Financial Industry %Securities% Industry keyword fuzzy matching Health %Disease Prevention% Industry keyword fuzzy matching …… …… ……

[0065] The entity information corresponding to the target information, the private keywords corresponding to the target information, and the label information corresponding to the target information may be added to the knowledge base after the target information is processed to obtain its corresponding entity information, private keywords, and label information.

[0066] The summary information can be obtained by the system performing multi-dimensional statistical analysis and recording and summarizing by industry, region, etc. on a regular basis.

[0067] For any piece of target information, entity extraction can be performed on the target information to obtain entity information corresponding to the target information.

[0068] Entity extraction (also known as named entity recognition) is an important task in natural language processing, which aims to identify entities with specific meanings from text data, such as names of people, places, and organizations. The main purpose of entity extraction is to extract entities with specific meanings from text. These entities are usually key information points in the text and are crucial for tasks such as understanding text content and building knowledge graphs. Through entity extraction, we can convert unstructured text data into structured data for subsequent analysis and application.

[0069] In the embodiment of the present application, any applicable method can be used to extract entities from the target information.

[0070] For example, a rule-based approach can be used to extract entities from target information. This approach mainly relies on rule templates developed by linguistics experts and domain experts to identify entities. For example, entities such as names of people and places in text can be identified by defining specific vocabulary lists, grammatical rules, and contextual information. However, this approach requires a lot of manpower and time, and is difficult to adapt to changes in different fields and text styles.

[0071] For example, statistical-based methods can be used to extract entities from target information. This type of method uses machine learning technology to identify entities in text by training classification models (such as support vector machines, decision trees, etc.). Usually, these methods require manual design of features, such as part-of-speech tagging, context information, grammatical dependencies, etc., and then use these features to train the model. Compared with rule-based methods, statistical-based methods have better adaptability and generalization capabilities, but they also require a large amount of labeled data for training.

[0072] For example, deep learning-based methods can be used to extract entities from target information. With the rise of deep learning technology, neural network models have been widely used in entity extraction tasks. Such methods can automatically learn complex feature representations from data, reducing reliance on manual feature engineering. At present, entity extraction methods based on deep learning have become mainstream and have achieved significant performance improvements.

[0073] Exemplarily, the W2NER (Word to Word Named Entity Recognition) model can be used to perform entity extraction on the target information, thereby obtaining the entity information corresponding to the target information. The W2NER model introduces two word-to-word relationships: NNW (Next-Neighboring-Word) and THW (Tail-Head-Word). Among them, NNW represents the adjacent positions of two words in the same entity, while THW represents two words in the same entity, and are the end and beginning of the entity respectively. Using the W2NER model for entity extraction can transform the target information into a predicted word-to-word relationship type, and obtain entity information after interpretation. This can solve the problem that other entity extraction schemes may split the proper nouns in the communication field, resulting in inaccurate recommendation effects.

[0074] In the embodiment of the present application, in order to further improve the accuracy of entity extraction, the W2NER model can also be trained in advance using target information samples, so that the W2NER model can process the target information more accurately. The specific process of W2NER model training can be processed according to actual experience, and this embodiment will not be discussed in detail here.

[0075] Step 102: Match the entity information corresponding to the target information with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information.

[0076] In the embodiment of the present application, after obtaining the entity information corresponding to the target information, the entity information corresponding to the target information is matched with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information.

[0077] The entity information corresponding to the target information is text data, and the private keywords contained in the knowledge base are also text data. Therefore, the entity information corresponding to the target information can be matched with the private keywords contained in the knowledge base to obtain the private keywords that match the entity information corresponding to the target information, and the private keywords that match the entity information corresponding to the target information are used as the private keywords corresponding to the target information.

[0078] In the embodiment of the present application, the entity information corresponding to the target information can be matched with the private keywords contained in the pre-created knowledge base in any applicable manner. For example, fuzzy matching method, complete matching method, etc. can be used for matching, or matching can be based on a deep learning model, matching based on vocabulary overlap, matching based on a shallow semantic topic model, matching based on vocabulary similarity, matching based on a similarity measure of word embedding, matching based on a convolutional neural network, etc. For the specific matching process, those skilled in the art can process it according to actual experience, and this embodiment will not be discussed in detail here.

[0079] Step 103: Acquire attribute information of the private keyword corresponding to the target information from the knowledge base, and generate tag information corresponding to the target information according to the attribute information.

[0080] The knowledge base stores attribute information corresponding to each private keyword. After obtaining the private keyword corresponding to the target information, the attribute information of the private keyword corresponding to the target information can be obtained from the knowledge base, and the label information corresponding to the target information can be generated according to the attribute information of the private keyword corresponding to the target information.

[0081] In the embodiment of the present application, at least one attribute information in the attribute information of the private keyword corresponding to the target information may be used as the label information corresponding to the target information.

[0082] Step 104: Add the entity information corresponding to the target information, the private keyword corresponding to the target information, and the tag information corresponding to the target information to the knowledge base.

[0083] After obtaining the entity information corresponding to the target information, the private keywords corresponding to the target information and the tag information corresponding to the target information, the entity information corresponding to the target information, the private keywords corresponding to the target information and the tag information corresponding to the target information are added to the knowledge base so that the relevant data of the target information can be summarized and maintained.

[0084] In an embodiment of the present application, by creating a knowledge base and combining entity extraction and private keyword matching, automatic labeling of target information can be achieved to improve labeling efficiency, and through entity extraction and private keyword matching, the accuracy of labeling can be further improved.

[0085] Reference Figure 2 , shows a step flow chart of another information processing method according to an embodiment of the present application.

[0086] like Figure 2 As shown, the information processing method may include the following steps:

[0087] Step 201: perform entity extraction on target information to obtain entity information corresponding to the target information.

[0088] Step 202: Match the entity information corresponding to the target information with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information.

[0089] Step 203: Acquire attribute information of the private keyword corresponding to the target information from the knowledge base, and generate tag information corresponding to the target information according to the attribute information.

[0090] Step 204: Add the entity information corresponding to the target information, the private keyword corresponding to the target information, and the tag information corresponding to the target information to the knowledge base.

[0091] Step 205: Update the knowledge base.

[0092] In the embodiment of the present application, a knowledge base operation mechanism is constructed to update the knowledge base in a timely manner.

[0093] In an optional implementation, the process of updating the knowledge base may include the following steps A1 to A3:

[0094] Step A1, obtaining target information marked with abnormalities within a set time period, and obtaining a private keyword corresponding to the target information marked with abnormalities from the knowledge base as a target private keyword.

[0095] In the embodiment of the present application, after the target information is processed through the above steps 201 to 204, the label information corresponding to the target information can be displayed and synchronized to the information administrator, who will regularly review the label information corresponding to each target information and mark the target information with the review result of abnormal labeling. The abnormal labeling can refer to inaccurate labeling, wrong labeling, etc.

[0096] In an embodiment of the present application, target information with marking anomalies within a set time period is obtained from a knowledge base (that is, target information marked as marking anomalies). The knowledge base stores private keywords corresponding to each target information. Therefore, the private keyword corresponding to the target information with marking anomalies can be obtained from the knowledge base, and the private keyword corresponding to the target information with marking anomalies can be used as the target private keyword.

[0097] The specific value of the set time period can be set according to actual needs, and this embodiment does not limit this.

[0098] Step A2, obtaining the number of target information items affected by the target private keyword.

[0099] Exemplarily, the process of obtaining the number of target information items affected by the target private keyword may include: obtaining the number of target information items with marked exceptions corresponding to the target private keyword in the knowledge base, and using the number of target information items with marked exceptions corresponding to the target private keyword as the number of target information items affected by the target private keyword.

[0100] For example, the private keyword corresponding to a target information marked with an exception is "government affairs||system". If there are 8 target information marked with an exception among the target information corresponding to the private keyword "government affairs||system" in the knowledge base, then the number of target information affected by the private keyword "government affairs||system" is 8.

[0101] Step A3: when the number of entries exceeds a preset number, setting the target private keyword to be invalid.

[0102] Setting the target private keyword to invalid does not mean physical deletion, but invalidation. The invalid private keyword will no longer be used to match the entity information corresponding to the target information. For the private keyword set to invalid, it can be restored according to actual needs, that is, the invalidation setting can be eliminated.

[0103] The specific value of the preset number of items can be set according to actual needs, and this embodiment does not impose any limitation on this.

[0104] In an optional implementation, the process of updating the knowledge base may include the following steps B1 to B3:

[0105] Step B1, obtaining target information with abnormal marking within a set time period, and obtaining entity information corresponding to the target information with abnormal marking from the knowledge base as candidate entity information.

[0106] In an embodiment of the present application, target information with marking anomalies within a set time period is obtained from a knowledge base (that is, target information marked as marking anomalies). The knowledge base stores entity information corresponding to each target information. Therefore, the entity information corresponding to the target information with marking anomalies can be obtained from the knowledge base, and the entity information corresponding to the target information with marking anomalies can be used as candidate entity information.

[0107] The specific value of the set time period can be set according to actual needs, and this embodiment does not limit this.

[0108] Step B2: selecting target entity information from the candidate entity information according to the occurrence frequency of the candidate entity information.

[0109] For each candidate entity information, the occurrence frequency of the candidate entity information in all the candidate entity information may be counted, and the target entity information may be selected from the candidate entity information according to the occurrence frequency of the candidate entity information.

[0110] For example, candidate entity information with a frequency exceeding a preset frequency can be selected from the candidate entity information as target entity information. The specific value of the preset frequency can be set according to actual needs, and this embodiment does not limit this.

[0111] Exemplarily, at least one candidate entity information with the highest occurrence frequency may be selected from the candidate entity information as the target entity information.

[0112] Step B3, obtaining the private keywords to be added according to the target entity information, and adding the private keywords to be added to the knowledge base.

[0113] In an optional implementation manner, the process of acquiring the private keyword to be added according to the target entity information may include: determining the target entity information as the private keyword to be added.

[0114] In an optional implementation, the process of acquiring the private keyword to be added according to the target entity information may include the following steps C1 to C3:

[0115] Step C1, combining multiple target entity information with the highest occurrence frequency in the target entity information to obtain combined entity information.

[0116] Exemplarily, for the multiple target entity information with the highest appearance frequency in the target entity information, any at least two of the target entity information may be combined to obtain combined entity information.

[0117] For example, the target entity information "government affairs" and the target entity information "system" can be combined into the combined entity information "government affairs||system", and so on.

[0118] Step C2, matching the combined entity information with the target information of the marking anomaly, and using the successfully matched combined entity information as the target combined entity information.

[0119] Exemplarily, in the process of matching the combined entity information with the target information of the marking anomaly, it can be determined whether the combined entity information is included in any target information of the marking anomaly. If so, it can be considered that the combined entity information matches successfully, otherwise it can be considered that the combined entity information matches unsuccessfully.

[0120] Step C3: acquiring the private keyword to be added according to the target entity information and the target combined entity information.

[0121] Exemplarily, the process of acquiring the private keywords to be added according to the target entity information and the target combined entity information may include: determining the target combined entity information and target entity information not belonging to the target combined entity information as the private keywords to be added.

[0122] Exemplarily, the process of obtaining the private keywords to be added based on the target entity information and the target combined entity information may include: selecting the private keywords to be added from the target combined entity information and the target entity information that does not belong to the target combined entity information according to a preset priority order and a preset quantity; wherein the priority of the target combined entity information is higher than that of the target entity information.

[0123] Specifically, a preset number of entity information may be selected from the target combined entity information and the target entity information that does not belong to the target combined entity information according to a preset priority order as the private keywords to be added.

[0124] The specific numerical value of the above preset number can be set according to actual needs, and this embodiment does not impose any limitation on this.

[0125] Of course, the target composite entity information and the target entity information not belonging to the target composite entity information may be filtered and re-evaluated to determine the private keywords to be added.

[0126] Step 206: After receiving the query question, using the large language model, obtain the query answer corresponding to the query question based on the knowledge base, and return the query answer.

[0127] In the embodiment of the present application, a large language model (LLM) and a knowledge base can be combined to provide a query function related to target information. Specifically, after receiving a query question, a query answer corresponding to the query question is obtained based on the knowledge base using a large language model, and the query answer is returned.

[0128] In an optional implementation, the target information in the knowledge base may be summarized according to the corresponding tag information at regular intervals to obtain summary information, such as summarizing the amount, quantity, and informatization amount by industry; summarizing the amount and quantity of informatization of different industries and the amount involved in the government by region, etc.

[0129] Correspondingly, the process of using a large language model to obtain a query answer corresponding to the query question based on the knowledge base may include: using a large language model to obtain a query answer corresponding to the query question based on the summary information in the knowledge base; when obtaining a query answer based on the summary information in the knowledge base fails, obtaining a query answer corresponding to the query question based on other information in the knowledge base except the summary information.

[0130] In the embodiment of the present application, in addition to supporting traditional query methods, question-and-answer queries are also supported. Enter the corresponding query question, such as querying the annual revenue of the government affairs industry and the medical industry in a certain area, etc. Through the large language model and knowledge base, the summary information in the knowledge base is queried first. If no relevant results are found in the summary information, other information in the knowledge base is queried again. Finally, after integration according to the prompt template corresponding to the large language model, the query answer is given. This method can quickly integrate responses and improve query speed and accuracy.

[0131] Reference Figure 3 , showing a system architecture diagram of an embodiment of the present application. Figure 3 Take the target information as the signal information as an example. Figure 3 As shown, the system architecture includes service models such as Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS), and Infrastructure-as-a-Service (IaaS).

[0132] In SaaS, there are portal layer, capability opening layer, and unified data access layer, and the unified data access layer includes application layer and data layer. The portal layer can include personal computer (PC) and application (App). In the application layer, functions such as data collection, signal query, signal analysis, abnormal operation mechanism, basic and interface management can be provided. In the data layer, signal data (i.e. signal information) and local knowledge base can be included.

[0133] In PaaS, platform management and scheduling, microservice framework, database, conversation language model (such as ChatGLM-6B (ChatGLM-6B is based on the GLM (General Language Model) architecture, a powerful, open source and Chinese-optimized conversation language model)), distributed cache, distributed small files, distributed task scheduling, cloud computing service tools (such as AWS (Amazon Web Services) tools, etc.) and other functions can be provided.

[0134] In IaaS, physical machines, cloud hosts, storage resources, network resources, etc. can be provided.

[0135] In the embodiment of the present application, the conversational language model can be applied in combination with the language chain architecture, such as the ChatGLM-6B+LangChain (language chain) architecture, and on the basis of the architecture, an entity extraction process, such as the W2NER entity extraction process, is introduced, and the target information is processed in combination with the knowledge base, so that the target information can be accurately marked, and an abnormal analysis mechanism is provided at the same time. By extracting hot words, high-frequency words, and other analysis and comparisons, private keywords are continuously iterated and improved, the accuracy of marking is improved, and the problem of inaccurate data judgment, large amounts of data cannot be judged, and a large number of people need to intervene manually for a long time is solved. In combination with a large language model to perform query operations, more accurate question and answer queries are supported, making queries simpler and greatly reducing the threshold for managers to use them.

[0136] Reference Figure 4 , which shows a schematic diagram of the processing process of a dialogue language model combined with a language chain architecture in an embodiment of the present application. Figure 4 As shown, in the dialogue language model combined with the language chain architecture, the unstructured data loader (Unstructured Loader) is used to process the local file (Local Documents), such as the target information, to obtain the text (Text); then the text (Text) is segmented by the text splitter (Text Splitter) to obtain text chunks (Text Chunks); then the text chunks (Text Chunks) are embedded (Embedding) to obtain the embedding vector, and the embedding vector is stored (Vector Store); the embedding (Embedding) operation is performed on the query question (Query) to obtain the query embedding vector (Query Vector); the vector similarity (Vector Similarity) of the query embedding vector (Query Vector) and the stored embedding vector (Vector Store) is calculated to obtain the related text chunks (Related Text Chunks); the prompt information (Prompt) is generated according to the prompt template (Prompt Template); and the large language model (LLM) is used to give the answer (Answer) according to the prompt information (Prompt).

[0137] Reference Figure 5 , which shows a schematic diagram of an information processing process in an embodiment of the present application. Figure 5As shown, the target information is segmented, segmentation is constructed, and then entity extraction is performed to obtain entity information, and the entity information is matched and labeled with private keywords in the knowledge base to obtain label information, and the entity information is text-vectorized to obtain vector data, and the vector data and label information are added to the database, and the data in the database is processed and converted to obtain a multi-dimensional summary result (i.e., summary information). After the user enters a query question, the query question is text-vectorized, and then a similarity search is performed on the query question based on the knowledge base to obtain relevant documents, and prompt information is obtained according to the prompt template, and the query answer is replied based on the prompt information using a large language model. The operation mechanism is executed on the knowledge base, including extracting high-frequency words (i.e., target entity information), combining high-frequency words (i.e., combining entity information), and processing according to the corresponding rules. For details, please refer to the relevant description above.

[0138] The embodiments of the present application have the following beneficial effects:

[0139] (1) Improving labeling accuracy: Current labeling is done manually and simply through regular matching, which has low accuracy and efficiency. After introducing a labeling method based on a large language model and knowledge base, the labeling accuracy can be improved and it has the ability to continuously learn and evolve.

[0140] (2) Improve the experience and lower the threshold for operation: The current technology supports traditional conditional screening queries. After the introduction of the tagging method based on the large language model and knowledge base, in addition to supporting traditional query methods, it also supports question-and-answer quick support, greatly lowering the threshold for operators. It can also automatically generate content according to needs and preferences to improve user satisfaction.

[0141] (3) Reduce operating costs: Through the knowledge base operation mechanism, continuously improve the labeling accuracy and gradually release manpower.

[0142] Reference Figure 6 , shows a structural block diagram of an information processing device according to an embodiment of the present application.

[0143] like Figure 6 As shown, the information processing device may include the following modules:

[0144] An extraction module 601 is used to perform entity extraction on target information to obtain entity information corresponding to the target information;

[0145] A matching module 602 is used to match the entity information corresponding to the target information with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information;

[0146] A generating module 603 is used to obtain attribute information of the private keyword corresponding to the target information from the knowledge base, and generate tag information corresponding to the target information according to the attribute information;

[0147] The writing module 604 is used to add the entity information corresponding to the target information, the private keyword corresponding to the target information and the tag information corresponding to the target information into the knowledge base.

[0148] Optionally, the device also includes: a first acquisition module, used to acquire target information with marking anomalies within a set time period, and acquire a private keyword corresponding to the target information with marking anomalies from the knowledge base as a target private keyword; a second acquisition module, used to acquire the number of target information items affected by the target private keyword; and a setting module, used to set the target private keyword to invalid when the number exceeds a preset number.

[0149] Optionally, the second acquisition module is specifically used to acquire the number of target information items with marked anomalies corresponding to the target private keyword in the knowledge base as the number of target information items affected by the target private keyword.

[0150] Optionally, the device also includes: a third acquisition module, used to acquire target information of marked anomalies within a set time period, and acquire entity information corresponding to the target information of marked anomalies from the knowledge base as candidate entity information; a fourth acquisition module, used to select target entity information from the candidate entity information according to the frequency of occurrence of the candidate entity information; and an adding module, used to acquire private keywords to be added according to the target entity information, and add the private keywords to be added to the knowledge base.

[0151] Optionally, the adding module includes: an information combination unit, used to combine multiple target entity information with the highest frequency of appearance in the target entity information to obtain combined entity information; an information matching unit, used to match the combined entity information with the target information with the marked exception, and use the successfully matched combined entity information as the target combined entity information; a keyword acquisition unit, used to acquire the private keywords to be added based on the target entity information and the target combined entity information.

[0152] Optionally, the keyword acquisition unit is specifically configured to determine the target composite entity information and target entity information that does not belong to the target composite entity information as private keywords to be added.

[0153] Optionally, the keyword acquisition unit is specifically used to select private keywords to be added from the target composite entity information and target entity information that does not belong to the target composite entity information according to a preset priority order and a preset quantity; wherein the priority of the target composite entity information is higher than that of the target entity information.

[0154] Optionally, the device further includes: a query module, which is used to, after receiving a query question, use a large language model to obtain a query answer corresponding to the query question based on the knowledge base, and return the query answer.

[0155] Optionally, the device also includes: a summary module, used to summarize the target information in the knowledge base according to the corresponding label information to obtain summary information; the query module is specifically used to use a large language model to obtain a query answer corresponding to the query question based on the summary information in the knowledge base; when the query answer fails to be obtained, the query answer corresponding to the query question is obtained based on other information in the knowledge base except the summary information.

[0156] In an embodiment of the present application, by creating a knowledge base and combining entity extraction and private keyword matching, automatic labeling of target information can be achieved to improve labeling efficiency, and through entity extraction and private keyword matching, the accuracy of labeling can be further improved.

[0157] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0158] In an embodiment of the present application, an electronic device is further provided. The electronic device may include a processor and a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium; when the computer program is executed by the processor, the processor executes the information processing method of any of the above embodiments.

[0159] Reference Figure 7 , shows a structural block diagram of an electronic device according to an embodiment of the present application. Figure 7 As shown, the electronic device 70 includes a processor 701 and a computer-readable storage medium 702 , on which a computer program 7021 is stored.

[0160] The processor 701 is used to execute the computer program 7021 stored on the computer-readable storage medium 702. When executing the computer program 7021, the processor 701 implements the information processing method of any of the above-mentioned embodiments and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0161] The processor 701 mentioned above may include but is not limited to: a central processing unit (CPU), a network processor (NP), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0162] The computer-readable storage medium 702 mentioned above may include but is not limited to: read-only memory (ROM), random access memory (RAM), compact disc read-only memory (CD-ROM), electronic erasable programmable read-only memory (EEPROM), hard disk, floppy disk, flash memory, etc.

[0163] In an embodiment of the present application, a computer-readable storage medium is further provided, on which a computer program is stored. The computer program can be executed by a processor of an electronic device. When the computer program is executed by the processor, the processor executes the information processing method described in any of the above embodiments.

[0164] Reference Figure 8 , shows a block diagram of a computer-readable storage medium according to an embodiment of the present application. Figure 8 As shown, a computer program 801 is stored on a computer-readable storage medium 80. When the computer program 801 is executed by a processor, the processor executes the information processing method described in any of the above embodiments and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0165] The various embodiments in this specification are interrelated, and each embodiment is described in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referenced to each other.

[0166] It should be noted that all actions of acquiring signals, information or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0167] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0168] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application.

[0169] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

[0170] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0171] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0172] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as 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 devices or units, which can be electrical, mechanical or other forms.

[0173] 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.

[0174] 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.

[0175] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. In summary, the content of this specification should not be understood as limiting the present application.

Claims

1. An information processing method, characterized in that: The method comprises: Performing entity extraction on the target information to obtain entity information corresponding to the target information; Matching the entity information corresponding to the target information with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information; Acquire attribute information of the private keyword corresponding to the target information from the knowledge base, and generate tag information corresponding to the target information according to the attribute information; The entity information corresponding to the target information, the private keyword corresponding to the target information, and the tag information corresponding to the target information are added to the knowledge base.

2. The method according to claim 1, characterized in that The method further comprises: Obtain target information marked with anomalies within a set time period, and obtain a private keyword corresponding to the target information marked with anomalies from the knowledge base as a target private keyword; Obtain the number of target information items affected by the target private keyword; When the number of entries exceeds a preset number, the target private keyword is set to be invalid.

3. The method according to claim 2, characterized in that Obtaining the number of target information items affected by the target private keyword includes: The number of target information items with marked anomalies corresponding to the target private keyword in the knowledge base is obtained as the number of target information items affected by the target private keyword.

4. The method according to claim 1, characterized in that: The method further comprises: Acquire target information with abnormal marking within a set time period, and acquire entity information corresponding to the target information with abnormal marking from the knowledge base as candidate entity information; Selecting target entity information from the candidate entity information according to the occurrence frequency of the candidate entity information; Acquire the private keywords to be added according to the target entity information, and add the private keywords to be added to the knowledge base.

5. The method according to claim 4, characterized in that Acquiring the private keyword to be added according to the target entity information includes: Combining multiple target entity information with the highest occurrence frequency in the target entity information to obtain combined entity information; Matching the combined entity information with the target information of the marking anomaly, and taking the successfully matched combined entity information as the target combined entity information; According to the target entity information and the target combined entity information, the private keyword to be added is acquired.

6. The method according to claim 5, characterized in that According to the target entity information and the target combined entity information, the private keyword to be added is obtained, including: The target composite entity information and target entity information that does not belong to the target composite entity information are determined as private keywords to be added.

7. The method according to claim 5, characterized in that According to the target entity information and the target combined entity information, the private keyword to be added is obtained, including: Selecting the private keywords to be added from the target composite entity information and the target entity information that does not belong to the target composite entity information according to a preset priority order and a preset number; The priority of the target combination entity information is higher than that of the target entity information.

8. The method according to claim 1, characterized in that The method further comprises: After receiving the query question, the query answer corresponding to the query question is obtained based on the knowledge base using the large language model, and the query answer is returned.

9. The method according to claim 8, characterized in that The method further includes: aggregating the target information in the knowledge base according to the corresponding tag information to obtain summary information; Utilizing a large language model to obtain a query answer corresponding to the query question based on the knowledge base, including: utilizing a large language model to obtain a query answer corresponding to the query question based on the summary information in the knowledge base; when obtaining a query answer fails, obtaining a query answer corresponding to the query question based on other information in the knowledge base except the summary information.

10. An information processing device, characterized in that: The device comprises: An extraction module, used to perform entity extraction on the target information to obtain entity information corresponding to the target information; A matching module, used for matching the entity information corresponding to the target information with the private keywords contained in the pre-created knowledge base to obtain the private keywords corresponding to the target information; A generating module, used for acquiring attribute information of the private keyword corresponding to the target information from the knowledge base, and generating label information corresponding to the target information according to the attribute information; The writing module is used to add the entity information corresponding to the target information, the private keyword corresponding to the target information and the tag information corresponding to the target information into the knowledge base.

11. An electronic device, characterized in that: The electronic device includes a processor and a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium; when the computer program is executed by the processor, the processor executes the information processing method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the information processing method according to any one of claims 1 to 9.

Citation Information

Patent Citations

  • Tag information-based user preference entity classification method and system

    CN106599098A

  • Data processing method and computer program product

    CN115048966A

  • Bank flow marking method and device

    CN116720944A