An information extraction method, device, equipment and medium combining RPA and AI

By combining RPA and AI, using pre-configured matching templates and extraction rules, key information in financial text is automatically extracted, solving the problem of low information acquisition efficiency in the existing technology and achieving efficient and accurate information extraction.

CN111881664BActive Publication Date: 2024-11-01BEIJING LAIYE NETWORK TECH CO LTD +1
View PDF 11 Cites 0 Cited by

Patent Information

Application Number
CN202010786436.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-06-30
Filing Date
2020-08-06
Publication Date
2024-11-01
Estimated Expiration
2040-08-06

AI Technical Summary

Technical Problem

In the process of financial management, it is difficult for the existing technology to efficiently identify and extract key information from massive texts, resulting in low information acquisition efficiency and the machine cannot understand the recognition results, which requires manual search and classification.

Method used

Combining RPA and AI, the information extractor to be used to automatically extract the extracted text by pre-configuring the matching template, and automatically identify and extract information in the text using configuration fields and extraction rules (such as text matching rules, vocabulary matching rules, and regular matching rules). .

Benefits of technology

It improves the efficiency and accuracy of information extraction, reduces manual participation, can accurately extract information in text, and improves the accuracy of information extraction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111881664B_ABST
    Figure CN111881664B_ABST
Patent Text Reader

Abstract

The embodiment of the present invention discloses an information extraction method, device, equipment and medium combining RPA and AI. The method includes: obtaining a text to be extracted; inputting the text to be extracted into an extractor configured with a matching template to obtain the extracted information in the text to be extracted. By applying the scheme provided by the embodiment of the present invention, information can be extracted from the text to be extracted based on the extractor. Since the matching template is pre-configured in the extractor, after the text to be extracted is input into the extractor, the information matching the matching template in the text to be extracted can be automatically extracted. The entire extraction process does not require manual participation, thereby improving the efficiency of information extraction. Moreover, compared with manual information extraction, information extraction based on matching templates can accurately extract the extracted information in the text to be extracted, thereby improving the accuracy of information extraction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of information extraction technology, and more specifically, to an information extraction method, apparatus, device, and medium that combines RPA (Robotic Process Automation) and AI (Artificial Intelligence). Background Technology

[0002] Artificial Intelligence (AI) is a technical science that studies and develops theories, methods, technologies, and application systems to simulate, extend, and expand human intelligence. With the maturation of AI technology, Natural Language Processing (NLP) has also developed rapidly. Natural language processing is an important direction in the field of AI. It enables effective communication between humans and computers using natural language. Robotic Process Automation (RPA) is a new type of AI-powered virtual process automation robot used to simulate human operations on a computer, automatically executing process tasks according to rules. RPA can be widely applied in various fields requiring process automation, such as financial management.

[0003] In financial management, it is often necessary to identify various invoices, receipts, contracts, and documents. One known method is to use image recognition software to identify the documents and obtain the results. However, these methods only provide the recognition results; the machine cannot understand the content. Users still need to search through massive amounts of data, searching for keywords and then organizing and classifying the information—a cumbersome process with low information retrieval efficiency. Summary of the Invention

[0004] This invention provides a method, apparatus, device, and medium for information extraction that combines RPA and AI to improve the efficiency of information extraction. The specific technical solution is as follows.

[0005] In a first aspect, embodiments of the present invention provide an information extraction method combining RPA and AI, the method comprising:

[0006] S1. Obtain the text to be extracted;

[0007] S2. Input the text to be extracted into the extractor with the configured matching template to obtain the extraction information from the text to be extracted.

[0008] Optionally, the matching template includes: configuration fields and extraction rules, and step S2 specifically includes:

[0009] S21. Input the text to be extracted into the extractor with the configured matching template, and extract the initial field in the text to be extracted that matches the configured field.

[0010] S22. The initial field that matches the extraction rule in the initial field is used as the extraction information in the text to be extracted.

[0011] Optionally, the extraction rules include at least one of the following: text matching rules, vocabulary matching rules, regular expression matching rules, and arbitrary text matching rules; step S22 specifically includes:

[0012] S221. Match each of the initial fields with each of the extraction rules in turn;

[0013] S222. For each of the initial fields, determine whether the initial field matches any of the extraction rules;

[0014] S223. When the initial field matches any of the extraction rules, the initial field is used as the extraction information in the text to be extracted.

[0015] Optionally, step S222 specifically includes:

[0016] S2221. For each of the initial fields, determine whether the initial field conforms to the text sentence pattern determined in the text matching rules;

[0017] S2222. For each of the initial fields, determine whether the initial field is included in the vocabulary that has been configured in the vocabulary matching rules;

[0018] S2223. For each of the initial fields, determine whether the expression of the initial field is consistent with the regular expression configured in the regular expression matching rules.

[0019] Optionally, the matching template is obtained in the following way:

[0020] S3. Configure the field list within the extractor;

[0021] S4. Configure the resource list within the extractor; the resource list includes a vocabulary and regular expressions;

[0022] S5. Construct an initial template. Configure the initial template based on the field list and the resource list to obtain the matching template.

[0023] Optionally, step S3 specifically includes:

[0024] S31. When a command to add a field is received, the field addition interface is displayed;

[0025] S32. Receive and store the input field names and remarks.

[0026] Optionally, step S3 specifically includes:

[0027] S33. When a batch field addition instruction is received, download the field import template from the preset storage location;

[0028] S34. Display the field import template and receive the field names and remarks entered in the field import template;

[0029] S35. When an upload file instruction is received, the field is stored in the import template.

[0030] Optionally, step S4 specifically includes:

[0031] S41. When a dictionary import instruction is received, the file selection interface is displayed;

[0032] S42. Receive the file selection instruction and upload the selected vocabulary.

[0033] Optionally, step S4 specifically includes:

[0034] S43. When a dictionary open command is received, the selected dictionary is displayed;

[0035] S44. Receive the edit operation on the selected vocabulary and store the modified vocabulary.

[0036] Optionally, step S4 specifically includes:

[0037] S45. When a new regular expression command is received, the interface for adding a regular expression is displayed;

[0038] S46. Receive and store the input regular expression description and regular expression.

[0039] Optionally, step S5 specifically includes:

[0040] S51. Construct an initial template, determine and store the fields selected from the field list;

[0041] S52. Display the rule setting interface, receive and store the input matching text, the regular expression selected from the resource list, and the vocabulary, and obtain the matching template.

[0042] Optionally, after step S5, the method further includes:

[0043] S6. Receive test text;

[0044] S7. When a test instruction is received, information is extracted from the test text, and first information or second information is displayed in a preset area; the first information includes: information extraction result; the second information includes: a matching template that matches the test text, and field names and field values ​​in the matching template that match the information extraction result.

[0045] Secondly, embodiments of the present invention provide an information extraction device combining RPA and AI, the device comprising:

[0046] The text acquisition module is used to acquire the text to be extracted;

[0047] The information extraction module is used to input the text to be extracted into an extractor with a pre-configured matching template to obtain the extracted information from the text to be extracted.

[0048] Optionally, the matching template includes: configuration fields and extraction rules, and the information extraction module specifically includes:

[0049] The field extraction submodule is used to input the text to be extracted into the extractor with the configured matching template, and extract the initial field in the text to be extracted that matches the configured field.

[0050] The information matching submodule is used to select the initial field that matches the extraction rule from the initial field as the extraction information in the text to be extracted.

[0051] Optionally, the extraction rules include at least one of the following: text matching rules, vocabulary matching rules, regular expression matching rules, and arbitrary text matching rules; the information matching submodule specifically includes:

[0052] The field matching subunit is used to match each of the initial fields with each of the extraction rules in turn;

[0053] The rule judgment subunit is used to determine whether each initial field matches any of the extraction rules.

[0054] The information extraction subunit is used to extract information from the text to be extracted when the initial field matches any of the extraction rules.

[0055] Optionally, the rule-determining subunit is specifically used for:

[0056] For each of the initial fields, determine whether the initial field conforms to the text sentence pattern defined in the text matching rules;

[0057] For each initial field, determine whether the initial field is included in the vocabulary that has been configured in the vocabulary matching rules;

[0058] For each initial field, determine whether the expression of the initial field is consistent with the regular expression configured in the regular expression matching rules.

[0059] Optionally, the device further includes:

[0060] The field configuration module is used to configure the field list within the extractor;

[0061] The resource configuration module is used to configure the resource list within the extractor; the resource list includes a vocabulary and regular expressions.

[0062] The template configuration module is used to build an initial template, configure the initial template based on the field list and the resource list, and obtain the matching template.

[0063] Optionally, the field configuration module specifically includes:

[0064] The Field Addition submodule is used to display the field addition interface when a field addition command is received;

[0065] The field storage submodule is used to receive and store the input field names and remarks.

[0066] Optionally, the field configuration module specifically includes:

[0067] The batch add submodule is used to download field import templates from a preset storage location when a batch add field instruction is received;

[0068] The template display submodule is used to display the field import template and receive the field names and remarks entered in the field import template.

[0069] The template storage submodule is used to store the fields and import them into the template when an upload file instruction is received.

[0070] Optionally, the resource configuration module specifically includes:

[0071] The vocabulary import submodule is used to display the file selection interface when a vocabulary import instruction is received;

[0072] The vocabulary storage submodule is used to receive file selection instructions and upload the selected vocabulary.

[0073] Optionally, the resource configuration module specifically includes:

[0074] The vocabulary opening submodule is used to display the selected vocabulary when a vocabulary opening command is received;

[0075] The vocabulary editing submodule is used to receive editing operations on the selected vocabulary and store the modified vocabulary.

[0076] Optionally, the resource configuration module specifically includes:

[0077] The Regular Expression Add Submodule is used to display the Regular Expression Add interface when a new regular expression command is received;

[0078] The regular expression storage submodule is used to receive and store the input regular expression descriptions and regular expressions.

[0079] Optionally, the template configuration module specifically includes:

[0080] The template building submodule is used to build an initial template, determine and store the fields selected from the field list;

[0081] The template configuration submodule is used to display the rule setting interface, receive and store the input matching text, the regular expression selected from the resource list and the vocabulary, and obtain the matching template.

[0082] Optionally, the device further includes:

[0083] The text receiving module is used to receive test text.

[0084] The text testing module is used to extract information from the test text when a test instruction is received, and display first information or second information in a preset area; the first information includes: information extraction result; the second information includes: a matching template that matches the test text, and field names and field values ​​in the matching template that match the information extraction result.

[0085] Thirdly, embodiments of the present invention provide a computing device, the device comprising:

[0086] Memory containing executable program code;

[0087] A processor coupled to memory;

[0088] Specifically, the processor calls the executable program code stored in the memory to execute the information extraction method combining RPA and AI as described in the first aspect above.

[0089] Fourthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, characterized in that the program is executed by a processor using the information extraction method combining RPA and AI as described in the first aspect above.

[0090] As can be seen from the above, the information extraction method, apparatus, device, and medium combining RPA and AI provided in this embodiment of the invention can extract information from the text to be extracted based on an extractor. Since the extractor is pre-configured with a matching template, after the text to be extracted is input into the extractor, information matching the matching template can be automatically extracted from the text. The entire extraction process does not require manual intervention, thereby improving information extraction efficiency. Furthermore, compared with manual information extraction, information extraction based on a matching template can accurately extract the information from the text to be extracted, improving the accuracy of information extraction. Of course, implementing any product or method of this invention does not necessarily require achieving all the advantages described above simultaneously.

[0091] The innovative aspects of this invention include:

[0092] 1. Information extraction is performed based on an extractor. Since the extractor is pre-configured with matching templates, once the text to be extracted is input, it can automatically extract information that matches the template. The entire extraction process requires no manual intervention, thus improving information extraction efficiency. Furthermore, compared to manual information extraction, template-based information extraction can accurately extract information from the text, improving the accuracy of information extraction.

[0093] 2. By setting configuration fields and extraction rules in the matching template, information can be extracted from the text to be extracted based on multiple conditions, thereby improving the accuracy of information extraction.

[0094] 3. After configuring the matching template, test the accuracy of information extraction of the matching template by using test text. If the accuracy of the matching template does not meet the requirements, it can be reconfigured until the accuracy meets the requirements before it is configured in the extractor, thereby improving the accuracy of information extraction from the text to be extracted. Attached Figure Description

[0095] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0096] Figure 1 This is a schematic diagram of a matching template according to an embodiment of the present invention;

[0097] Figure 2 A flowchart illustrating an information extraction method combining RPA and AI provided in an embodiment of the present invention;

[0098] Figure 3 This is another flowchart illustrating the information extraction method combining RPA and AI provided in an embodiment of the present invention.

[0099] Figure 4 This is a schematic diagram of the field list page according to an embodiment of the present invention;

[0100] Figure 5 This is a schematic diagram of the interface for adding fields in an embodiment of the present invention;

[0101] Figure 6 This is a schematic diagram illustrating the batch addition of fields in an embodiment of the present invention;

[0102] Figure 7 This is a schematic diagram of the field editing interface according to an embodiment of the present invention;

[0103] Figure 8 This is a schematic diagram of the resource list according to an embodiment of the present invention;

[0104] Figure 9 This is a schematic diagram of the file selection interface according to an embodiment of the present invention;

[0105] Figure 10 This is a schematic diagram of the interface for creating a new vocabulary list according to an embodiment of the present invention;

[0106] Figure 11 This is a schematic diagram of the vocabulary value editing interface according to an embodiment of the present invention;

[0107] Figure 12 This is a schematic diagram of the regular expression-added interface in an embodiment of the present invention;

[0108] Figure 13 This is a schematic diagram of the template list interface according to an embodiment of the present invention;

[0109] Figure 14 This is a schematic diagram of the matching rule configuration interface according to an embodiment of the present invention;

[0110] Figure 15 This is a schematic diagram of the text test results according to an embodiment of the present invention;

[0111] Figure 16 A schematic diagram of an information extraction device combining RPA and AI provided in an embodiment of the present invention;

[0112] Figure 17 This is a schematic diagram of the structure of a computing device provided in an embodiment of the present invention. Detailed Implementation

[0113] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0114] It should be noted that the terms "comprising" and "having," and any variations thereof, in the embodiments and drawings of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0115] In the description of this invention, the term "text to be extracted" refers to the text from which information needs to be extracted, and can be any type of document, such as contract documents, various invoices, etc.

[0116] In the description of this invention, the term "matching template" refers to a template configured in the extractor for extracting information from the text to be extracted. Condition set In other words, the matching template specifies which fields in the text to be extracted; only fields that meet all the conditions in the matching template can be extracted.

[0117] In the description of this invention, the term "extracted information" refers to all the keyword information extracted from the text to be extracted, which can be text, numbers, English letters, etc.

[0118] In the description of this invention, the term "initial field" refers to a field extracted from the text that matches a field matching rule. This field only matches the field matching rule but has not yet matched the extraction rule, and therefore cannot be output as final extracted information. Only initial fields that match the extraction rule can be determined as final extracted information.

[0119] In the description of this invention, the term "text matching rule" refers to the definition of specific text as matching conditions for a particular sentence structure.

[0120] In the description of this invention, the term "vocabulary matching rule" refers to the requirement that for enumerable information such as city names, country names, and company names, they and their corresponding synonyms need to be added to a dictionary. This way, when performing information matching, the extracted information can be searched based on the dictionary to determine whether the extracted information meets the requirements.

[0121] In the description of this invention, the term "regular expression matching rule" refers to the use of regular expressions to represent information with a common format, such as time, date, and currency. When performing matching, it is determined whether the extracted information meets the given regular expression, thereby determining whether the extracted information meets the requirements.

[0122] In the description of this invention, the term "arbitrary text matching rule" refers to the assumption that any extracted content satisfies the rule for non-enumerable information such as project names and locations during matching, since the information is non-enumerable.

[0123] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0124] Example 1

[0125] In the process of financial management, it is usually necessary to identify keywords contained in various invoices, receipts, contracts, documents, etc. In other words, it is necessary to identify the required keyword information from the entire document.

[0126] In this embodiment of the invention, a matching template can be pre-configured in the extractor. For example... Figure 1 As shown, this illustrates a matching template. Each matching template includes field and rule configurations. Only information in the input text that matches both the configured fields and rules will be selected for extraction.

[0127] Example 2

[0128] Figure 2 This is a flowchart illustrating an information extraction method combining RPA and AI provided in an embodiment of the present invention. The method is applied to electronic devices and specifically includes the following steps.

[0129] S210, Obtain the text to be extracted.

[0130] In this embodiment of the invention, when information extraction from any text is required, the user can transmit the text to a predetermined storage location on the electronic device. When the electronic device detects new text in the predetermined storage location, it can use this new text as the text to be extracted. The text to be extracted can be of any type, such as contracts, invoices, etc.

[0131] S220: Input the text to be extracted into the extractor with the configured matching template to obtain the extraction information from the text to be extracted.

[0132] In this embodiment of the invention, a matching template can be pre-configured in the extractor. After the electronic device obtains the text to be extracted, it can extract information from the text using the matching template.

[0133] In one implementation, the matching template of this embodiment may include: a configuration field and extraction rules. The configuration field defines which fields the current extractor needs to extract from the text, and the final extraction result can only output the defined fields. The extraction rules define which fields among the extracted fields need to be output, and only fields that match the extraction rules can be output as extracted information.

[0134] Specifically, when an electronic device extracts information, it can first input the text to be extracted into an extractor with a pre-configured matching template, and extract the initial fields in the text that match the configured fields; then, the initial fields that match the extraction rules are used as the extracted information in the text to be extracted.

[0135] In one implementation, the extraction rules include at least one of the following: text matching rules, vocabulary matching rules, regular expression matching rules, and arbitrary text matching rules. When the electronic device matches the initial fields with the extraction rules, specifically, it can first match each initial field sequentially with each extraction rule; then, for each initial field, it can determine whether the initial field matches any extraction rule; when the initial field matches any extraction rule, the initial field is used as the extracted information from the text to be extracted.

[0136] In other words, as long as the initial field matches any rule in the extraction rules, it can be determined that the initial field matches the extraction rules, and the initial field can be identified as the information to be extracted from the text.

[0137] When determining whether each initial field matches the extraction rules, in one implementation, the electronic device can, for each initial field, determine whether the initial field conforms to the text sentence structure defined in the text matching rules; if so, the initial field can be determined to be the information to be extracted from the text. For each initial field, it can be determined whether the initial field is included in the vocabulary already configured in the vocabulary matching rules; if so, the initial field can be determined to be the information to be extracted from the text. For each initial field, it can be determined whether the expression of the initial field is consistent with the regular expression already configured in the regular expression matching rules; if so, the initial field can be determined to be the information to be extracted from the text.

[0138] The above judgment process is merely an example and does not limit the order in which the rules are judged. In other words, in practical applications, the above judgment order can be arbitrary, and this embodiment of the invention does not impose any specific limitations on the above judgment order.

[0139] As described above, in this embodiment of the invention, information extraction can be performed on the text to be extracted using an extractor. Since the extractor is pre-configured with a matching template, after the text to be extracted is input into the extractor, information matching the matching template can be automatically extracted. The entire extraction process does not require manual intervention, thereby improving information extraction efficiency. Furthermore, compared with manual information extraction, information extraction based on the matching template can accurately extract the information from the text to be extracted, improving the accuracy of information extraction.

[0140] Example 3

[0141] Figure 3 This is another flowchart illustrating the information extraction method combining RPA and AI provided in this embodiment of the invention, specifically a flowchart of the matching template construction process. This method is applied to electronic devices and includes the following steps.

[0142] S310. Configure the field list within the extractor.

[0143] The field configuration within the extractor defines which fields the current extractor needs to extract from the text, and the final extraction result can only output the defined fields.

[0144] like Figure 4 As shown, the field list page displays the field name, notes, and operations for each field. When configuring the field list, you can add a single field or add multiple fields in batches.

[0145] When a user needs to add a field, they can click... Figure 4 The "Add Field" button shown indicates that the electronic device will receive the instruction to add a field. The electronic device can then display the field addition interface, such as... Figure 5 As shown.

[0146] exist Figure 5 The interface shown allows users to enter field names and remarks. Field names can be, for example, contract name, contract number, contract amount, etc.; remarks can include multiple field values, such as characters and numbers. After entering the field names and remarks, users can click... Figure 5 Once the "Save" button is displayed, the electronic device can receive and store the entered field name and remarks. At this point, the field configuration is successful.

[0147] When users need to add fields in batches, they can... Figure 4 Select the "Add Fields in Batch" button from the "Add New Fields" dropdown menu shown, as follows: Figure 6As shown, the electronic device can receive a batch field addition instruction. In this case, it can download the field import template from the preset storage location and display the field import template. The user can then enter the field names and remarks in the field import template. After clicking the "Save Field Import Template" button, the electronic device can save the field import template.

[0148] When the user clicks Figure 6 When the "Batch Export Fields" button is shown, the electronic device can automatically download an Excel file containing all field information.

[0149] Users can modify the configured fields. Specifically, for example... Figure 4 As shown, when a user clicks the "Edit" button after a field, the electronic device can display something like this. Figure 7 The interface shown displays the field name and related notes for that field. Users can then modify the field name or notes. After modification, click... Figure 7 The "Save" button shown allows your electronic device to save the modified fields.

[0150] S320. Configure the resource list within the extractor; the resource list includes a vocabulary and regular expressions.

[0151] The resource configuration in the extractor is used to define the matching resources used in the matching rules, such as... Figure 8 As shown, the resources in this embodiment of the invention are divided into two categories: vocabulary and regular expressions.

[0152] When a user wants to import a vocabulary list, they can click as follows: Figure 8 The left arrow button next to the dictionary in the interface shown indicates that the electronic device can receive the dictionary import command. In this case, the electronic device can display a file selection interface, such as... Figure 9 As shown. Afterwards, the user can select files, and once the selected files are confirmed, the electronic device can upload the selected vocabulary.

[0153] Importing a thesaurus allows for the batch addition of existing thesauruses. In this embodiment of the invention, users can also create new thesauruses; specifically, when a user clicks on... Figure 8 When the "+" button appears after the vocabulary list in the interface shown, the electronic device can display the new vocabulary list interface, such as... Figure 10 As shown. In the case of Figure 10 In the interface, users can enter a dictionary name and a dictionary description. After clicking the "OK" button, the electronic device can save the entered dictionary name and dictionary description.

[0154] Optionally, in this embodiment of the invention, multiple vocabulary values ​​can be added to each vocabulary list. Specifically, when a user clicks on a vocabulary list, the electronic device can display the selected vocabulary list, and the user can then edit the vocabulary list content in the displayed interface.

[0155] In one implementation, such as Figure 11 As shown, when a user edits a vocabulary value, they can add multiple expressions for that value, such as multiple expressions for a particular product. Furthermore, a standard vocabulary value can be set among the multiple expressions for a single vocabulary value. During rule matching, if any expression of the current vocabulary value is matched, the standard vocabulary value is output.

[0156] Regular expression configuration specifies the regular expressions that may be used in the matching template. The validity of these regular expressions needs to be validated upon saving. In one implementation, when the user clicks... Figure 8 When the "+" button appears after the regular expression in the interface shown, the electronic device can display a new regular expression interface, such as... Figure 12 As shown. In the case of Figure 12 In the interface, users can enter regular expression descriptions and regular expressions. After clicking the "Save" button, the electronic device can receive and store the entered regular expression descriptions and regular expressions.

[0157] S330. Construct an initial template. Configure the initial template based on the field list and resource list to obtain a matching template.

[0158] The template function in an extractor is a fundamental feature. An extractor can be configured with multiple templates, each corresponding to a text segment extraction rule. When text is input, the extractor automatically matches the templates in order and outputs the extracted content from the templates according to the matching order.

[0159] When configuring matching templates, electronic devices can display a template list interface, such as... Figure 13 As shown, when the user clicks the "New Model Version" button, the electronic device can display the following: Figure 14 The interface shown. In, as... Figure 14 In the interface shown, users can configure matching rules. Specifically, users can enter matching text, select regular expressions and vocabulary from the resource list, and set whether to perform fuzzy matching. Users can also select the required fields from the field list. After configuring the fields and matching rules, the template can be published to obtain the matching template.

[0160] In one implementation, to improve the accuracy of template configuration and thus the accuracy of information extraction, the electronic device can also test the configured matching template.

[0161] Specifically, the electronic device can receive test text; this test text can be a .txt file. For example... Figure 15 As shown, when the user clicks the "Start Test" button, the electronic device can receive the test instruction, extract information from the test text, and display the first information or the second information in a preset area; the first information includes: the information extraction result; the second information includes: the matching template that matches the test text, and the field name and field value in the matching template that match the information extraction result.

[0162] like Figure 15 As shown, the left side is the text display area, which displays the entered test text. Clicking on the text area allows you to edit the text, and clicking "Start Test" again will change the result on the right.

[0163] Figure 15 The right side displays the visualization results, showing the templates matched in the text, as well as the field names and values ​​output from those templates. When the visualization results view is selected, hovering the mouse over the template area highlights the matched text content in the text display area; hovering the mouse over the field area in the template highlights the matched output field content in the text area. Clicking "View Template" opens a new page with details for that template. The JSON results view displays the final test results.

[0164] In this embodiment, by setting configuration fields and extraction rules in the matching template, information can be extracted from the text to be extracted based on multiple conditions, thereby improving the accuracy of information extraction. After the matching template is configured, the accuracy of information extraction of the matching template is tested by test text. If the accuracy of the matching template does not meet the requirements, it can be reconfigured until the accuracy meets the requirements before it is configured in the extractor, thereby improving the accuracy of information extraction from the text to be extracted.

[0165] Example 4

[0166] Figure 16 This is a schematic diagram of an information extraction device combining RPA and AI provided in an embodiment of the present invention. The device includes:

[0167] The text acquisition module 610 is used to acquire the text to be extracted.

[0168] The information extraction module 620 is used to input the text to be extracted into an extractor with a configured matching template to obtain the extracted information from the text to be extracted.

[0169] Optionally, the matching template includes: configuration fields and extraction rules, and the information extraction module 620 specifically includes:

[0170] The field extraction submodule is used to input the text to be extracted into the extractor with the configured matching template, and extract the initial field in the text to be extracted that matches the configured field.

[0171] The information matching submodule is used to select the initial field that matches the extraction rule from the initial field as the extraction information in the text to be extracted.

[0172] Optionally, the extraction rules include at least one of the following: text matching rules, vocabulary matching rules, regular expression matching rules, and arbitrary text matching rules; the information matching submodule specifically includes:

[0173] The field matching subunit is used to match each of the initial fields with each of the extraction rules in turn;

[0174] The rule judgment subunit is used to determine whether each initial field matches any of the extraction rules.

[0175] The information extraction subunit is used to extract information from the text to be extracted when the initial field matches any of the extraction rules.

[0176] Optionally, the rule-determining subunit is specifically used for:

[0177] For each of the initial fields, determine whether the initial field conforms to the text sentence pattern defined in the text matching rules;

[0178] For each initial field, determine whether the initial field is included in the vocabulary that has been configured in the vocabulary matching rules;

[0179] For each initial field, determine whether the expression of the initial field is consistent with the regular expression configured in the regular expression matching rules.

[0180] Optionally, the device further includes:

[0181] The field configuration module is used to configure the field list within the extractor;

[0182] The resource configuration module is used to configure the resource list within the extractor; the resource list includes a vocabulary and regular expressions.

[0183] The template configuration module is used to build an initial template, configure the initial template based on the field list and the resource list, and obtain the matching template.

[0184] Optionally, the field configuration module specifically includes:

[0185] The Field Addition submodule is used to display the field addition interface when a field addition command is received;

[0186] The field storage submodule is used to receive and store the input field names and remarks.

[0187] Optionally, the field configuration module specifically includes:

[0188] The batch add submodule is used to download field import templates from a preset storage location when a batch add field instruction is received;

[0189] The template display submodule is used to display the field import template and receive the field names and remarks entered in the field import template.

[0190] The template storage submodule is used to store the fields and import them into the template when an upload file instruction is received.

[0191] Optionally, the resource configuration module specifically includes:

[0192] The vocabulary import submodule is used to display the file selection interface when a vocabulary import instruction is received;

[0193] The vocabulary storage submodule is used to receive file selection instructions and upload the selected vocabulary.

[0194] Optionally, the resource configuration module specifically includes:

[0195] The vocabulary opening submodule is used to display the selected vocabulary when a vocabulary opening command is received;

[0196] The vocabulary editing submodule is used to receive editing operations on the selected vocabulary and store the modified vocabulary.

[0197] Optionally, the resource configuration module specifically includes:

[0198] The Regular Expression Add Submodule is used to display the Regular Expression Add interface when a new regular expression command is received;

[0199] The regular expression storage submodule is used to receive and store the input regular expression descriptions and regular expressions.

[0200] Optionally, the template configuration module specifically includes:

[0201] The template building submodule is used to build an initial template, determine and store the fields selected from the field list;

[0202] The template configuration submodule is used to display the rule setting interface, receive and store the input matching text, the regular expression selected from the resource list and the vocabulary, and obtain the matching template.

[0203] Optionally, the device further includes:

[0204] The text receiving module is used to receive test text.

[0205] The text testing module is used to extract information from the test text when a test instruction is received, and display first information or second information in a preset area; the first information includes: information extraction result; the second information includes: a matching template that matches the test text, and field names and field values ​​in the matching template that match the information extraction result.

[0206] As described above, in this embodiment of the invention, information can be extracted from the text to be extracted using an extractor. Since the extractor is pre-configured with a matching template, after the text to be extracted is input into the extractor, information matching the matching template can be automatically extracted. The entire extraction process does not require manual intervention, thus improving information extraction efficiency. Furthermore, compared to manual information extraction, information extraction based on the matching template can accurately extract the information from the text to be extracted, improving the accuracy of information extraction.

[0207] The above-described apparatus embodiments correspond to the method embodiments and have the same technical effects. For detailed explanations, please refer to the method embodiments. The apparatus embodiments are derived from the method embodiments; detailed explanations can be found in the method embodiments section, and will not be repeated here.

[0208] Example 5

[0209] Please see Figure 17 , Figure 17 This is a schematic diagram of the structure of a computing device provided in an embodiment of the present invention. Figure 17 As shown, the computing device may include:

[0210] Memory 701 storing executable program code;

[0211] Processor 702 coupled to memory 701;

[0212] The processor 702 calls the executable program code stored in the memory 701 to execute the information extraction method combining RPA and AI provided in any embodiment of the present invention.

[0213] This computing device embodiment and the above-described method embodiment are embodiments derived from the same inventive concept, and related aspects can be referred to each other. The above-described computing device embodiment corresponds to the method embodiment and has the same technical effects as the method embodiment; for detailed description, please refer to the method embodiment.

[0214] This invention discloses a computer-readable storage medium storing a computer program that causes a computer to execute the information extraction method combining RPA and AI provided in any embodiment of this invention.

[0215] In various embodiments of the present invention, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0216] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0217] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-accessible memory. Based on this understanding, the technical solution of the present invention, in essence, 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. This computer software product is stored in a memory and includes several requests to cause a computer device (which can be a personal computer, server, or network device, specifically a processor in the computer device) to execute some or all of the steps of the methods described in the various embodiments of the present invention.

[0218] This storage medium embodiment and the above method embodiment are embodiments derived from the same inventive concept, and related aspects can be referred to each other. The above storage medium embodiment corresponds to the method embodiment and has the same technical effects; for detailed description, please refer to the method embodiment.

[0219] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0220] Those skilled in the art will understand that the modules in the apparatus of the embodiments can be distributed in the apparatus of the embodiments as described in the embodiments, or they can be located in one or more devices different from this embodiment with corresponding changes. The modules of the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0221] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An information extraction method combining RPA and AI, It is characterized in that The method comprises: S1. Obtain the text to be extracted; S2, inputting the text to be extracted into an extractor configured with a matching template to obtain extraction information in the text to be extracted, wherein the matching template includes: configuration fields and extraction rules, and the extraction rules include at least one of the following: text matching rules, vocabulary matching rules, regular matching rules, and arbitrary text matching rules; The step of inputting the text to be extracted into an extractor configured with a matching template to obtain the extraction information in the text to be extracted includes: Input the text to be extracted into an extractor configured with a matching template, and extract the initial field in the text to be extracted that matches the configured field; Using the initial fields in the initial fields that match the extraction rule as the extraction information in the text to be extracted; The step of taking the initial fields in the initial fields that match the extraction rule as the extraction information in the text to be extracted includes: Matching each of the initial fields with each of the extraction rules in sequence; For each of the initial fields, determining whether the initial field matches any of the extraction rules; When the initial field matches any of the extraction rules, the initial field is used as the extraction information in the text to be extracted; The step of determining, for each of the initial fields, whether the initial field matches any of the extraction rules comprises: For each of the initial fields, determining whether the initial field conforms to the text sentence pattern determined in the text matching rule; For each of the initial fields, determining whether the initial field is included in the vocabulary configured in the vocabulary matching rule; For each of the initial fields, determine whether the expression of the initial field is consistent with the regular expression configured in the regular matching rule; The matching template is obtained in the following manner: S3, configuring the field list in the extractor, adding one or more fields; S4, configuring the resource list in the extractor; the resource list includes a vocabulary and regular expressions; S5. construct an initial template, and configure the initial template based on the field list and the resource list to obtain the matching template; S6. Receive a test text; S7. When a test instruction is received, information is extracted from the test text, and the first information or the second information is displayed in a preset area; the first information includes: information extraction results; the second information includes: a matching template that matches the test text, and a field name and field value in the matching template that matches the information extraction results.

2. The method according to claim 1, It is characterized in that The S3 specifically includes: S31. When receiving a new field instruction, display a field adding interface; S32. Receive and store the input field name and remark information.

3. The method according to claim 1, It is characterized in that The S3 specifically includes: S33, when receiving a batch adding field instruction, downloading a field import template from a preset storage location; S34, displaying the field import template, and receiving each field name and each remark information input in the field import template; S35. When receiving a file upload instruction, store the field import template.

4. The method according to claim 1, It is characterized in that The S4 specifically includes: S41, when receiving a vocabulary import instruction, displaying a file selection interface; S42, receiving a file selection instruction and uploading the selected vocabulary.

5. The method according to claim 1, It is characterized in that The S4 specifically includes: S43, when receiving a vocabulary opening instruction, displaying the selected vocabulary; S44: receiving an editing operation on the selected vocabulary, and storing the modified vocabulary.

6. The method according to claim 1, It is characterized in that The S4 specifically includes: S45, when receiving a new regular expression instruction, displaying a regular expression adding interface; S46. Receive and store the input regular description and regular expression.

7. The method according to claim 1, It is characterized in that The S5 specifically includes: S51, constructing an initial template, determining and storing the fields selected from the field list; S52, displaying a rule setting interface, receiving and storing the input matching text, the regular expression and thesaurus selected from the resource list, and obtaining the matching template.

8. An information extraction device combining RPA and AI, It is characterized in that The device comprises: A text acquisition module is used to acquire the text to be extracted; An information extraction module, used for inputting the text to be extracted into an extractor configured with a matching template to obtain extracted information in the text to be extracted, wherein the matching template includes: a configuration field and an extraction rule, and the extraction rule includes at least one of the following: a text matching rule, a vocabulary matching rule, a regular matching rule, and an arbitrary text matching rule; Wherein, the information extraction module includes: A field extraction submodule, used for inputting the text to be extracted into an extractor configured with a matching template, and extracting an initial field in the text to be extracted that matches the configured field; An information matching submodule, used for taking the initial fields in the initial fields that match the extraction rule as the extraction information in the text to be extracted; A field matching subunit, used for matching each of the initial fields with each of the extraction rules in sequence; A rule judgment subunit, used for judging, for each of the initial fields, whether the initial field matches any of the extraction rules; An information extraction subunit, used for taking the initial field as the extracted information in the text to be extracted when the initial field matches any of the extraction rules; The rule judgment subunit is specifically used for: For each of the initial fields, determining whether the initial field conforms to the text sentence pattern determined in the text matching rule; For each of the initial fields, determining whether the initial field is included in the vocabulary configured in the vocabulary matching rule; For each of the initial fields, determine whether the expression of the initial field is consistent with the regular expression configured in the regular matching rule; Wherein, the device further comprises: A field configuration module, used to configure the field list in the extractor; A resource configuration module, used to configure the resource list in the extractor; the resource list includes a vocabulary and regular expressions; A template configuration module, used for constructing an initial template, configuring the initial template based on the field list and the resource list to obtain the matching template; A text receiving module, used for receiving a test text; A text testing module is used to extract information from the test text when a test instruction is received, and to display the first information or the second information in a preset area; the first information includes: the information extraction result; the second information includes: a matching template that matches the test text, and a field name and field value in the matching template that matches the information extraction result.

9. A computing device, It is characterized in that The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the information extraction method combining RPA and AI as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the information extraction method combining RPA and AI as described in any one of claims 1 to 7 is implemented.

Citation Information

Patent Citations

  • Address information extracting method and system

    CN104537062A

  • Method and device for parsing law judgment document

    CN106815208A

  • Semantic model based text message extraction method and device

    CN107608949A

  • Zone defined text information extraction method and device

    CN107729480A

  • A method and a device for extracting a web page text based on an extraction template

    CN109033282A