Information processing device, information processing method, and program

Abstract

To provide an information processing device, an information processing method, and a program that can correct text according to user attributes.SOLUTION: The information processing device of one embodiment comprises: first acquisition means for acquiring text data and user attributes of a user utilizing the text data; read means for reading language rules corresponding to the user attributes acquired by the acquisition means from reference data that defines language rules for the languages used according to the type of user attributes; generation means for generating query information containing an instruction sentence that instructs the correction of the text contained in the text data based on the language rules read by the read means; second acquisition means for inputting the text data and the query information generated by the generation means to a generation AI that is configured to perform corrections to a sentence according to the contents of the instruction in the query information and generate the corrected sentence; and output means for outputting the corrected sentence acquired by the second acquisition means.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] An embodiment of the present invention relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] Conventionally, in order to go paperless, etc., paper documents with printed or handwritten characters have been converted into image data using scanners, cameras, etc. Also, there is a character recognition technology called OCR (Optical Character Recognition / Reader) that recognizes the character portion of the image data document and converts it into character data.

[0003] In the above-mentioned OCR, if a misrecognition occurs during character recognition, the character is converted into erroneous character data, so the converted character data must be checked and corrected manually. Therefore, a technology for correcting character data using an AI model has been proposed (for example, Reference 1).

[0004] Incidentally, documents handled by users in an office or the like may contain expressions using terminology specific to the organization or department to which the user belongs. However, in the above-mentioned conventional technology, when correcting character data, the attributes of the user, such as the organization or department to which the user performing OCR belongs, are not taken into consideration. Therefore, while the conventional technology can correct clerical errors, the correction may not be based on the user's attributes, and there is room for improvement in terms of convenience. Summary of the Invention [Problem to be solved by the invention]

[0005] The problem to be solved by the present invention is to provide an information processing device, an information processing method, and a program that are capable of correcting text in accordance with user attributes. [Means for solving the problem]

[0006] The information processing device of the embodiment comprises a first acquisition means for acquiring text data and user attributes of a user who uses the text data; a reading means for reading language rules corresponding to the user attributes acquired by the acquisition means from reference data that specifies language rules for the language used according to the type of user attribute; a generation means for generating inquiry information including an instruction sentence that instructs the correction of text included in the text data based on the language rules read by the reading means; a second acquisition means for inputting the text data and the inquiry information generated by the generation means into a generation AI that is functioned to make corrections to the text according to the instructions in the inquiry information and generate the corrected corrected text, and acquiring the corrected text generated by the generation AI; and an output means for outputting the corrected text acquired by the second acquisition means. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a schematic diagram showing a schematic configuration of an information processing system according to an embodiment. [Figure 2] FIG. 2 is a block diagram illustrating an example of a hardware configuration of an edge device according to the embodiment. [Figure 3] FIG. 3 is a block diagram illustrating an example of a hardware configuration of an edge server according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of a data configuration of a reference DB according to the embodiment. [Figure 5] FIG. 5 is a block diagram illustrating an example of a functional configuration of an edge device according to the embodiment. [Figure 6] FIG. 6 is a block diagram illustrating an example of a functional configuration of the edge server according to the embodiment. [Figure 7] FIG. 7 is a sequence diagram illustrating an example of a control process of the information processing system according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0008] Hereinafter, with reference to the drawings, an embodiment of the information processing system S will be described. Note that in the embodiment described below, an edge server 2 installed in an office or the like will be described as an example of an information processing device, but the present invention is not limited to the embodiment.

[0009] 1 is a schematic diagram showing a schematic configuration of an information processing system S according to an embodiment. As shown in FIG. 1, the information processing system S includes an edge device 1 and an edge server 2.

[0010] The edge device 1 and the edge server 2 are communicably connected via a network Na such as a LAN (Local Area Network). Note that the number of edge devices 1 connected to the edge server 2 is not limited to the illustrated example.

[0011] The edge device 1 is a terminal device used by a user of the information processing system S. The edge device 1 also refers to various devices that serve as a point of contact between the information processing system S and the user, or a system that includes such devices. For example, the edge device 1 is a scanner device or a multi-function peripheral / printer (MFP) that is installed in an image forming device such as a facsimile machine.

[0012] The edge device 1 exchanges various types of information with the edge server 2. Specifically, the edge device 1 uses a card reader 109 (see FIG. 2 ), which will be described later, to acquire user attributes of the user operating the edge device 1 from a medium containing information that identifies an individual, such as an employee ID card. Here, the user attributes are, for example, identification information for identifying the organization or department to which the user operating the edge device 1 belongs. Note that the user attributes are not limited to this example. For example, the user attributes may be the job title or position title of the user operating the edge device 1. Furthermore, for example, the user attributes may be the user ID of the user.

[0013] Furthermore, when the edge device 1 receives a command to execute a reading process from an operation unit 107 (see FIG. 2 ) described later by a user operating the edge device 1, the edge device 1 executes reading of a document made of a paper medium or the like, and acquires an image of the document. Then, the edge device 1 executes character recognition processing (hereinafter also referred to as OCR processing) on ​​the acquired image, and extracts text contained in the image as character data (hereinafter also referred to as sentences). Note that known technology can be applied to the OCR processing.

[0014] Note that the image to be subjected to the OCR process is not limited to the image acquired by the reading process. For example, the edge device 1 may perform the OCR process on an image stored in the device itself or an image acquired from an external server connected to the edge device 1 via a network (not shown).

[0015] Furthermore, the edge device 1 generates a document file converted into a general-purpose file format based on the text extracted by the OCR process. For example, the edge device 1 generates a file in Portable Document Format (PDF) format (hereinafter also referred to as a PDF file) based on the text extracted by the OCR process. The edge device 1 then transmits the user attributes and the PDF file to the edge server 2. Note that it is preferable to generate the PDF file in a format that allows the text extracted by the OCR process contained in the PDF file to be searched, i.e., that allows the text to be extracted from the PDF file.

[0016] Furthermore, when the edge device 1 receives a storage completion notification from the edge server 2, the edge device 1 displays a message on a display unit 106 (see FIG. 2) to be described later, informing the user of the edge device 1 that storage has been completed. The storage completion notification will be described in detail later.

[0017] The edge server 2 is a server that manages documents and is installed in an office, for example. The edge server 2 corrects the text extracted by the OCR process according to the user attributes. The edge server 2 also changes the storage destination (output destination) of the corrected text according to the user attributes.

[0018] Next, a description will be given of the hardware configuration of the edge device 1. Fig. 2 is a block diagram showing an example of the hardware configuration of the edge device 1 according to the embodiment.

[0019] As shown in FIG. 2, the edge device 1 includes a CPU (Central Processing Unit) 101, which is an example of a processor, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, and a memory unit 104.

[0020] The CPU 101 comprehensively controls each part of the edge device 1. The ROM 102 stores various programs. The RAM 103 is a workspace where programs and various data are expanded.

[0021] The memory unit 104 is a non-volatile memory such as a hard disk drive (HDD) or flash memory that retains stored information even when the power is turned off. The memory unit 104 includes a control program unit 1041.

[0022] The control program unit 1041 stores a control program for controlling the edge device 1. The CPU 101, ROM 102, RAM 103, and memory unit 104 are connected to one another via a bus 110. The CPU 101, ROM 102, and RAM 103 constitute a control unit 100 having a computer configuration. That is, the control unit 100 executes a control process for the edge device 1, which will be described later, by the CPU 101 operating in accordance with the control program unit 1041 stored in the ROM 102 or memory unit 104 and expanded in the RAM 103.

[0023] The control unit 100 is connected to a communication unit 105 , a display unit 106 , an operation unit 107 , a reading unit 108 , and a card reader 109 via a bus 110 .

[0024] The communication unit 105 is a communication interface such as a LAN I / F (Interface), and is connected to the network Na. The communication unit 105 transmits and receives various information to and from the edge server 2, for example, via the network Na.

[0025] The display unit 106 is a display device such as an LCD (Liquid Crystal Display). The display unit 106 displays various information under the control of the CPU 101. The operation unit 107 is an input device such as a keyboard or a pointing device. The operation unit 107 outputs operation contents input via the input device to the CPU 101. The operation unit 107 may be a touch panel provided on the display unit 106.

[0026] The reading unit 108 is a scanner device using a CCD (Charge Coupled Device) sensor, a CIS (Contact Image Sensor), etc. The reading unit 108 reads a document placed on a placement unit (not shown) of the edge device 1.

[0027] The card reader 109 is, for example, a magnetic card reader that reads information from a card medium such as a magnetic card. The card reader 109 reads user attributes from the card medium. Note that the card reader 109 is not limited to a magnetic card, and may also read information from, for example, an IC card.

[0028] Next, a description will be given of the hardware configuration of the edge server 2. Fig. 3 is a block diagram showing an example of the hardware configuration of the edge server 2 according to the embodiment.

[0029] As shown in FIG. 3, the edge server 2 includes a CPU 201, which is an example of a processor, a ROM 202, a RAM 203, and a memory unit 204.

[0030] The CPU 201 comprehensively controls each unit of the edge server 2. The ROM 202 stores various programs. The RAM 203 is a workspace where programs and various data are expanded.

[0031] The memory unit 204 is a non-volatile memory such as a HDD or flash memory that retains stored information even when the power is turned off. The memory unit 204 includes a control program unit 2041, a reference DB 2042, a calibration LLM 2043, and a PDF file storage unit 2044.

[0032] The control program unit 2041 stores a control program for controlling the edge server 2. The CPU 201, ROM 202, RAM 203, and memory unit 204 are connected to one another via a bus 206. The CPU 201, ROM 202, and RAM 203 constitute a control unit 200 having a computer configuration. That is, the control unit 200 executes a control process for the edge server 2, which will be described later, by the CPU 201 operating in accordance with the control program unit 2041 stored in the ROM 202 or memory unit 204 and loaded into the RAM 203.

[0033] The reference DB 2042 is a data table or database for managing information related to an organization to which a user who operates the edge device 1 belongs. Fig. 4 is a diagram showing an example of the data configuration of the reference DB 2042. As shown in Fig. 4, the reference DB 2042 stores user attributes, classification codes, and DB information such as names and descriptions corresponding to the classification codes in association with each other. Here, a set of user attributes, classification codes, and DB information corresponds to reference data.

[0034] A classification code is an example of identification information associated with one or more user attributes. For example, if the user attribute "A1" shown in FIG. 4 is identification information corresponding to the accounting department, the user attribute "A1" is included in the classification code "1," and therefore the user attribute "A1," i.e., the accounting department, corresponds to the DB information of the classification code "1." Similarly, if the user attribute "B1" is identification information corresponding to the engineering department, the user attribute "B1" is included in the classification code "2," and therefore the user attribute "B1," i.e., the engineering department, corresponds to the DB information of the classification code "2." Here, the classification code can also be rephrased as identification information for distinguishing user attributes that share a common language rule used in documents.

[0035] The name represents, for example, a keyword or term shared (used) within a group to which the user who operates the edge device 1 belongs. The explanation is, for example, a description that explains the usage of the corresponding name. For example, the explanation corresponding to the name "POS" is "an abbreviation for Point of Sale, a term that manages information on the point at which a product in a store is sold."

[0036] Here, the names and descriptions are examples of language rules in this embodiment. The language rules define the language rules to be used in documents handled by an organization classified by user attributes and classification codes. The names and descriptions are used, for example, when replacing synonyms.

[0037] The names and descriptions are not limited to those written in natural language, but may be stored in a format written in semantic vectors calculated by known natural language processing techniques, etc. The system may be configured to store the names and descriptions in both a format written in natural language and a format written in semantic vectors. It is preferable that the keywords to be registered as names are those used within a group to which the user belongs, and are characteristic of that group (e.g., jargon).

[0038] Returning to FIG. 3, the proofreading LLM2043 is a generation AI that generates text, such as a large language model (LLM). The proofreading LLM2043 corrects text according to user attributes to generate corrected text. Here, the corrected text is text in which expressions such as typos and omissions contained in text extracted by OCR processing have been corrected. The typos and omissions include typos and omissions contained in the original text as well as typos and omissions caused by misreading (misrecognition) due to OCR processing. Note that, although an LLM is used as the generation AI in this embodiment, the generation AI is not limited to an LLM as long as it is capable of generating text.

[0039] The proofreading LLM2043 is constructed using well-known deep learning techniques and is an LLM that is functionalized to output corrected sentences based on the conditions specified in the input text. Here, the conditions may be, for example, reference conditions that serve as references for deriving the output results or constraints that narrow down the output results.

[0040] The proofreading LLM 2043 of this embodiment generates a corrected sentence corresponding to the user attributes of the user operating the edge device 1 in response to input of a query text (hereinafter also referred to as a prompt) generated based on DB information extracted from the reference DB 2042 and an instruction statement instructing the correction of the sentence.

[0041] The PDF file storage unit 2044 is an example of a storage device. The PDF file storage unit 2044 stores PDF files. The PDF file storage unit 2044 preferably has a data structure in which storage areas for PDF files are divided according to user attributes. The PDF file storage unit 2044 stores, according to user attributes, PDF files that have been updated from text extracted by OCR processing to corrected text generated by the proofreading LLM 2043.

[0042] The control unit 200 is also connected to a communication unit 205 via a bus 206. The communication unit 205 is a communication interface such as a LAN I / F, and is connected to a network Na. The communication unit 205 transmits and receives various types of information to and from the edge device 1 via the network Na.

[0043] Next, a functional configuration of the edge device 1 will be described. Fig. 5 is a block diagram showing an example of the functional configuration of the edge device 1. As shown in Fig. 5, the control unit 100 includes, as functional components, a user attribute acquisition unit 1001, an OCR processing unit 1002, a transmission / reception unit 1003, and a display control unit 1004. Note that the functional configuration of the edge device 1 is not limited to this.

[0044] Specifically, the control unit 100 (CPU 101) of the edge device 1 realizes the above-mentioned functional configuration by executing the control program unit 1041 stored in the memory unit 104. Note that in this embodiment, the above-mentioned functional configuration is a software configuration realized by cooperation between the processor and program of the edge device 1, but is not limited to this, and a hardware configuration in which part or all of the functional configuration is realized by a dedicated circuit or the like may also be used.

[0045] The user attribute acquisition unit 1001 acquires user attributes. Specifically, the user attribute acquisition unit 1001 cooperates with the card reader 109 of the edge device 1 to acquire the user attributes of the user who operates the edge device 1.

[0046] The OCR processing unit 1002 extracts text from an image acquired by the edge device 1 by performing OCR processing on the image. Specifically, when the OCR processing unit 1002 receives an instruction to execute OCR processing from a user operating the edge device 1 via the operation unit 107, the OCR processing unit 1002 acquires an image of a document using the reading unit 108. The OCR processing unit 1002 then performs OCR processing on the acquired image to extract text data contained in the image as text (hereinafter also referred to as extracted text). The OCR processing unit 1002 also generates a document file converted into a general-purpose file format based on the extracted text. In this embodiment, the document file converted into a general-purpose file format is a PDF file.

[0047] The transmitting / receiving unit 1003 transmits and receives various information to and from the edge server 2. Specifically, the transmitting / receiving unit 1003 transmits the user attributes acquired by the user attribute acquisition unit 1001 and the PDF file generated by the OCR processing unit 1002 to the edge server 2. Furthermore, upon receiving a storage completion notification from the edge server 2, the transmitting / receiving unit 1003 instructs the display control unit 1004, which will be described later, to display a message indicating that the PDF storage has been completed on the display unit 106 of the edge device 1.

[0048] The display control unit 1004 cooperates with other functional units to display various screens on the display unit 106. Specifically, upon receiving an instruction from the transmitting / receiving unit 1003, the display control unit 1004 displays a message indicating that the PDF has been stored on the display unit 106 of the edge device 1.

[0049] Next, the functional configuration of the edge server 2 will be described. Fig. 6 is a block diagram showing an example of the functional configuration of the edge server 2. As shown in Fig. 6, the control unit 200 has, as its functional components, a reception processing unit 2001, a text extraction unit 2002, a read processing unit 2003, a prompt generation unit 2004, a text acquisition processing unit 2005, a text conversion processing unit 2006, and an output processing unit 2007. Note that the functional configuration of the edge server 2 is not limited to this.

[0050] Specifically, the control unit 200 (CPU 201) of the edge server 2 realizes the above-mentioned functional configuration by executing the control program unit 2041 stored in the memory unit 204. In this embodiment, the above-mentioned functional configuration is a software configuration realized by cooperation between the processor and program of the edge server 2, but is not limited to this, and a hardware configuration in which part or all of the functional configuration is realized by a dedicated circuit or the like may also be used.

[0051] The receiving processor 2001 is an example of a first acquisition unit. The receiving processor 2001 acquires text data and user attributes of a user who uses the text data from the edge device 1. Specifically, the receiving processor 2001 receives (acquires) the PDF file and user attributes transmitted by the edge device 1.

[0052] The text extraction unit 2002 extracts text from the document file converted into a general-purpose file format. Specifically, when the receiving and processing unit 2001 receives (acquires) the PDF file and user attributes, the text extraction unit 2002 extracts sentences from the PDF file. Here, the sentences extracted from the PDF file are called extracted sentences.

[0053] The reading processing unit 2003 is an example of a reading means. The reading processing unit 2003 reads, from the reference DB 2042, a linguistic rule corresponding to the user attribute acquired by the receiving processing unit 2001. Specifically, the reading processing unit 2003 refers to the reference DB 2042 of the edge server 2, and identifies a classification code corresponding to the user attribute received (acquired) by the receiving processing unit 2001. Next, the reading processing unit 2003 reads, from the reference DB 2042, DB information corresponding to the identified classification code, which has a similarity to a term or usage of the term included in the extracted sentence that is equal to or greater than a predetermined threshold.

[0054] Here, the linguistic rules are explanatory text that explains terms and their usage. The DB information of the reference DB 2042 is an example of the linguistic rules. Note that any technology can be used to extract elements that correspond to the linguistic rules from a sentence, and known technologies such as natural language processing and morphological analysis can be used.

[0055] As a method for calculating the similarity, for example, calculation may be made based on the matching rate of words between the extracted sentence extracted by the text extraction unit 2002 and the DB information corresponding to the classification code identified by the read processing unit 2003. Alternatively, for example, calculation may be made by converting the extracted sentence extracted by the text extraction unit 2002 into a semantic vector using a known natural language processing technique or the like, and calculating the cosine similarity between the extracted sentence extracted by the text extraction unit 2002 and the DB information that has been converted into a semantic vector.

[0056] 4, if the user attribute is "A1" and the extracted sentence contains "term A," the read processing unit 2003 refers to the reference DB 2042 and identifies that the classification code corresponding to the user attribute "A1" is "1." Next, since "term A" included in the extracted sentence is included in the DB information corresponding to the classification code "1" (the similarity is equal to or greater than the threshold), the read processing unit 2003 reads out the DB information corresponding to "term A" from the reference DB 2042, i.e., the name "term A" and the explanation "The target area of ​​term A is...".

[0057] 4, if the user attribute is "A1" and the extracted sentence contains "term C," the read processing unit 2003 refers to the reference DB 2042 and identifies that the classification code corresponding to the user attribute "A1" is "1." In this case, since "term C" contained in the extracted sentence is not included in the DB information for classification code "1" (since the similarity is not equal to or greater than the threshold), the read processing unit 2003 does not read DB information for "term C" from the reference DB 2042.

[0058] If a term included in the extracted sentence is not included in the DB information corresponding to the classification code, the read processing unit 2003 may refer to DB information for another classification code to determine whether the term included in the extracted sentence exists. For example, if the user attribute is "A1" and the extracted sentence includes "Term C," the read processing unit 2003 determines whether "Term C" exists by referring to DB information for classification code "2" because "Term C" does not exist in DB information for classification code "1." In the reference DB 2042 shown in FIG. 4, "Term C" exists in DB information for classification code "2." Therefore, the read processing unit 2003 reads the explanation for "Term C" and searches DB information for classification code "1" for an explanation similar to the content of the explanation. If an explanation with a similarity equal to or greater than the threshold exists in DB information for classification code "1," the read processing unit 2003 reads the corresponding DB information as DB information corresponding to "Term C." For example, if the similarity between the explanation of "Term C" and the explanation of "Term B" is above a threshold, in other words, if a synonym for "Term C" is "Term B," the read processing unit 2003 reads out the DB information for "Term B" in relation to "Term C."

[0059] Furthermore, if a term included in an extracted sentence is not included in the DB information corresponding to the classification code, the reading processing unit 2003 may identify the usage of the term included in the extracted sentence by referring to dictionary data (not shown) in which the usage of each term is stored. In this case, the reading processing unit 2003 may search for DB information of synonyms from DB information corresponding to the user attributes based on the identified usage, and read out the corresponding DB information as DB information corresponding to the term included in the extracted sentence. The dictionary data may be stored in the memory unit 204 or in an external device.

[0060] The prompt generation unit 2004 is an example of a generation means. The prompt generation unit 2004 generates a prompt including an instruction sentence that instructs the user to correct the extracted sentence based on the linguistic rules read by the read processing unit 2003. Specifically, the prompt generation unit 2004 generates a prompt including an instruction sentence that instructs the user to correct the extracted sentence based on the user attributes received (acquired) by the receiving processing unit 2001 and on the DB information read by the read processing unit 2003 from the reference DB 2042.

[0061] Here, the prompt may be configured to switch from a plurality of types of templates depending on user attributes, etc. In addition, it is preferable to include specific instructions in the prompt, such as "only correct typos" or "replace the parts to be corrected with DB information read from the reference DB 2042," as details of the instruction.

[0062] The text acquisition processor 2005 is an example of a second acquisition means. The text acquisition processor 2005 corrects the text in accordance with the instructions in the prompt, inputs the extracted text and the prompt generated by the prompt generator 2004 to a proofreading LLM 2043 that is functionally configured to generate corrected text, and acquires the corrected text generated by the proofreading LLM 2043. Specifically, the text acquisition processor 2005 corrects the extracted text in accordance with the instructions included in the prompt, inputs the extracted text and the prompt generated by the prompt generator 2004 to the proofreading LLM 2043 that is functionally configured to generate corrected text, and acquires the corrected text generated by the proofreading LLM 2043.

[0063] The text conversion processing unit 2006 is an example of a conversion means. The text conversion processing unit 2006 converts the corrected text into a general-purpose file format that includes the corrected text. Specifically, the text conversion processing unit 2006 converts the corrected text acquired by the text acquisition processing unit 2005 into a PDF file.

[0064] The output processing unit 2007 determines the output destination according to the user attributes, and outputs (stores) the PDF file to the determined output destination. Specifically, the output processing unit 2007 determines the output destination according to the user attributes. Here, the output destination can be set arbitrarily. For example, the output processing unit 2007 stores the PDF file in the PDF file storage unit 2044. Also, for example, the output processing unit 2007 transmits a storage completion notification to the edge device 1.

[0065] Next, the control processing of the information processing system S will be described. FIG. 7 is a sequence diagram showing an example of the control processing of the information processing system S according to the embodiment. In the sequence diagram shown in FIG. 7, the edge device 1 transmits user attributes and a PDF file to the edge server 2. Then, the edge server 2 generates a prompt including an instruction sentence instructing correction of the extracted sentence based on DB information corresponding to the user attributes read from the reference DB 2042. Next, the edge server 2 inputs the prompt to the proofreading LLM 2043 and acquires the corrected sentence generated by the proofreading LLM 2043. Next, the edge server 2 outputs the acquired corrected sentence to an output destination according to the user attributes. This shows an example of processing.

[0066] First, the user attribute acquisition unit 1001 of the edge device 1 cooperates with the card reader 109 of the edge device 1 to acquire the user attributes of the user operating the edge device 1 (step S101). Next, when the OCR processing unit 1002 of the edge device 1 receives an instruction to execute OCR processing from the user operating the edge device 1 via the operation unit 107, it acquires an image of a document using the reading unit 108. Then, the OCR processing unit 1002 executes OCR processing on the acquired image to extract character data contained in the image as a sentence (step S102).

[0067] Next, the OCR processing unit 1002 generates a PDF file based on the extracted sentence (step S103). Next, the transmitting / receiving unit 1003 of the edge device 1 transmits the user attributes acquired by the user attribute acquisition unit 1001 and the PDF file generated by the OCR processing unit 1002 to the edge server 2 (step S104).

[0068] Meanwhile, the receiving processor 2001 of the edge server 2 receives (acquires) the user attributes and the PDF file from the edge device 1 (step S105). Next, when the receiving processor 2001 receives (acquires) the PDF file and the user attributes, the text extractor 2002 extracts text from the PDF file (step S106).

[0069] Next, the reading processing unit 2003 of the edge server 2 refers to the reference DB 2042 and identifies the classification code corresponding to the user attribute received (acquired) by the receiving processing unit 2001. Next, the reading processing unit 2003 reads from the reference DB 2042, among the DB information corresponding to the identified classification code, DB information whose similarity to the term or the usage of the term contained in the extracted sentence is equal to or greater than a predetermined threshold (step S107).

[0070] Next, the prompt generating unit 2004 of the edge server 2 generates a prompt including an instruction sentence to correct the extracted sentence, based on the DB information read by the read processing unit 2003 from the reference DB 2042 (step S108).

[0071] Next, the text acquisition processor 2005 of the edge server 2 modifies the extracted sentence according to the instruction included in the prompt, and inputs the extracted sentence and the prompt generated by the prompt generator 2004 to the proofreading LLM 2043, which is functionally configured to generate a corrected sentence (step S109). Next, the text acquisition processor 2005 acquires the corrected sentence generated by the proofreading LLM 2043 (step S110).

[0072] Next, the text conversion processing unit 2006 of the edge server 2 converts the corrected text acquired by the text acquisition processing unit 2005 into a PDF file (step S111). Next, the output processing unit 2007 of the edge server 2 determines an output destination according to the user attributes, and outputs (stores) the PDF file in the determined output destination (in this example, the PDF file storage unit 2044) (step S112). Next, the output processing unit 2007 transmits a storage completion notification to the edge device 1 (step S113).

[0073] Next, when the transmitting / receiving unit 1003 receives a storage completion notification from the edge server 2, it instructs the display control unit 1004 to display a message indicating that the PDF has been stored on the display unit 106 of the edge device 1 (step S114).

[0074] As described above, in the information processing system S of this embodiment, the edge device 1 transmits the user attributes and the PDF file to the edge server 2. Then, the edge server 2 generates a prompt including an instruction sentence instructing the user to correct the extracted sentence based on the DB information corresponding to the user attributes read from the reference DB 2042. Next, the edge server 2 inputs the prompt to the proofreading LLM 2043 and acquires the corrected sentence generated by the proofreading LLM 2043. Next, the edge server 2 outputs the acquired corrected sentence to an output destination according to the user attributes.

[0075] As a result, the information processing system S of this embodiment can acquire corrected text obtained by correcting the text extracted by OCR processing according to the attributes of the user operating the edge device 1. Furthermore, the information processing system S of this embodiment can output (store) the PDF file containing the corrected text at a different output destination according to the attributes of the user operating the edge device 1. Therefore, text can be corrected according to the user attributes, thereby improving convenience.

[0076] The above-described embodiment can be modified as needed by partially changing the configuration or functions of each of the above-described devices. Therefore, several modifications of the above-described embodiment will be described below as other embodiments. The following mainly focuses on differences from the above-described embodiment, and detailed descriptions of commonalities with the content already described will be omitted. The modifications described below may be implemented individually or in appropriate combination.

[0077] (Variation 1) In the above-described embodiment, the method by which the edge device 1 acquires user attributes is such that the user attribute acquisition unit 1001 of the edge device 1 cooperates with the card reader 109, but this is not limited to this. Alternatively, the method may involve capturing an image of a medium containing information that identifies an individual, such as an employee ID card, using an imaging unit (not shown), inputting user attributes by a user operating the edge device 1 via the operation unit 107, or reading user attributes from text extracted by OCR processing.

[0078] (Variation 2) In the above-described embodiment, the edge server 2 outputs the PDF file to the PDF file storage unit 2044 in accordance with the user attributes. However, this is not limiting, and the output destination of the PDF file may be specified by input from the operation unit 107 of the edge device 1. In this case, the transceiver unit 1003 of the edge device 1 transmits the specified output destination along with the user attributes and the PDF file to the edge server 2. Furthermore, the output processing unit 2007 of the edge server 2 outputs the PDF file including the corrected text to the specified output destination. Furthermore, the edge server 2 may transmit the PDF file including the corrected text to an external server, for example, a cloud server (not shown).

[0079] (Variation 3) In the above embodiment, the edge server 2 stores the PDF file in the PDF file storage unit 2044 in accordance with the user attributes and then transmits a storage completion notification to the edge device 1. However, the output processing unit 2007 of the edge server 2 may acquire a URL link (hereinafter also referred to as a link) of the stored PDF file when storing the PDF file in the PDF file storage unit 2044 in accordance with the user attributes, and transmit the acquired link together with the storage completion notification to the edge device 1. Furthermore, when the transmission / reception unit 1003 of the edge device 1 receives the link and the storage completion notification from the edge server 2, it may instruct the display control unit 1004 to display the link and a message indicating that the PDF storage has been completed on the display unit 106.

[0080] (Variation 4) In the above embodiment, the OCR processing unit 1002 of the edge device 1 generates a PDF file based on the extracted text and transmits it to the edge server 2, but this is not limiting, and the extracted text may be transmitted to the edge server 2. In this case, the transmitting / receiving unit 1003 of the edge device 1 transmits the text (text data) extracted by the OCR processing unit 1002 and the user attributes to the edge server 2. In this case, the text extraction unit 2002 of the edge server 2 may extract the text from the text data transmitted from the edge device 1.

[0081] When this modification is adopted, the above-described steps S103 and S105 may be skipped. Also, a document file may be generated by converting the text extracted by the text extraction unit 2002 of the edge server 2 into a general-purpose file format.

[0082] (Variation 5) In the above embodiment, the text conversion processing unit 2006 of the edge server 2 converts the corrected text acquired by the text acquisition processing unit 2005 into a PDF file, and the output processing unit 2007 stores the PDF file containing the corrected text in the PDF file storage unit 2044, but this is not limited to this, and the output processing unit 2007 may also be configured to send the corrected text (text data) to the edge server 2 along with a storage completion notification.

[0083] (Variation 6) In the above-described embodiment, the edge server 2 is configured to be installed in an office or the like, but this is not limiting, and the edge server 2 may be a cloud server installed on a cloud.

[0084] (Variation 7) In the above embodiment, the linguistic rules are explanatory text (names and explanations in the reference DB 2042) that explain terms and their usage, but the linguistic rules are not limited to this and may also include rules for written expression such as style and grammar. This allows the format of the entire document to be modified according to user attributes, thereby further improving convenience.

[0085] (Variation 8) In the above embodiment, the read processing unit 2003 of the edge server 2 identifies the classification code corresponding to the user attributes received (acquired) by the receiving processing unit 2001 of the edge server 2, but this is not limiting.

[0086] For example, the user attribute acquisition unit 1001 of the edge device 1 acquires a user ID that uniquely identifies a user and transmits the user ID to the edge server 2. Next, upon receiving (acquiring) the user ID, the reception processing unit 2001 refers to a table (not shown) in which user IDs and user attributes are associated and stored in the memory unit 204 of the edge server 2, and identifies the user attribute corresponding to the received (acquired) user ID. Then, the read processing unit 2003 may be configured to identify a classification code corresponding to the user attribute identified by the reception processing unit 2001.

[0087] (Variation 9) In the above-described embodiment, a PDF file generated based on text extracted by the OCR processing unit 1002 of the edge device 1 is sent to the edge server 2, but this is not limited to this, and text data (such as a PDF file) created in advance as electronic data may also be sent to the edge server 2.

[0088] For example, the edge device 1 may transmit a document (text data) created by a user to the edge server 2. In this case, the text extraction unit 2002 of the edge server 2 may extract sentences from the document (text data) received (acquired) by the receiving processing unit 2001.

[0089] The programs executed by the information processing system S of the embodiment and the modified example may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Also, the programs executed by the information processing system S of the embodiment and the modified example may be provided or distributed via a network such as the Internet.

[0090] The programs executed by each device in the above-described embodiments are provided in a state where they are pre-installed in a ROM, a storage unit, etc. The programs executed by each device in the above-described embodiments may be provided by being recorded in an installable or executable file format on a computer-readable recording medium such as a CD-ROM, a flexible disk (FD), a CD-R, or a digital versatile disk (DVD).

[0091] Furthermore, the programs executed by each device in the above-described embodiments may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Also, the programs executed by each device in the above-described embodiments may be provided or distributed via a network such as the Internet.

[0092] These novel embodiments and their modifications may be embodied in various other forms, and various omissions, substitutions, changes, and combinations may be made without departing from the spirit of the invention. These embodiments and their modifications are included within the scope and spirit of the inventions, and are also included in the inventions described in the claims and their equivalents. [Explanation of symbols]

[0093] S Information Processing System 1. Edge Devices 2 Edge Server 1001 User attribute acquisition unit 1002 OCR processing section 1003 Transmitter / Receiver 1004 Display control unit 2001 Receiving Processing Unit 2002 Text Extraction Unit 2003 Read processing unit 2004 Prompt Generation Unit 2005 Text Acquisition Processing Unit 2006 Text conversion processing section 2007 Output Processing Unit [Prior art documents] [Patent documents]

[0094] [Patent Document 1] Patent Publication No. 2021-039694

Claims

1. a first acquisition means for acquiring text data and user attributes of a user who uses the text data; a reading means for reading a language rule corresponding to the user attribute acquired by the first acquiring means from reference data that defines a language rule for a language to be used depending on a type of user attribute; a generating means for generating query information including an instruction sentence for instructing correction of text included in the text data based on the linguistic rules read by the reading means; a second acquisition means for inputting the text data and the inquiry information generated by the generation means into a generation AI that is configured to modify the text in accordance with instructions in the inquiry information and generate a modified text, and acquiring the modified text generated by the generation AI; an output means for outputting the corrected sentence acquired by the second acquisition means; An information processing device comprising:

2. The reference data includes, as the language rules, terms and explanatory text explaining the usage of the terms; the reading means reads from the reference data, among the linguistic rules corresponding to the user attributes, linguistic rules whose similarity to a term or a usage of the term included in the text data is equal to or greater than a threshold value; The information processing device according to claim 1 .

3. the output means determines an output destination of the corrected sentence in accordance with the user attribute acquired by the first acquisition means, and outputs the corrected sentence to the determined output destination. The information processing device according to claim 1 .

4. further comprising a conversion means for converting the corrected text into a general-purpose file format; the output means outputs the corrected sentence converted by the conversion means.

4. The information processing device according to claim 1.

5. An information processing method executed by an information processing device, a first acquisition step of acquiring text data and user attributes of a user who uses the text data; a reading step of reading, from reference data that defines language rules for languages ​​used according to types of user attributes, language rules corresponding to the user attributes acquired in the first acquiring step; a generating step of generating query information including an instruction sentence instructing correction of text included in the text data based on the linguistic rules read in the reading step; a second acquisition step of inputting the text data and the inquiry information generated by the generation step into a generation AI that is functionally configured to modify the text in accordance with the instructions in the inquiry information and generate a modified text, and acquiring the modified text generated by the generation AI; an output step of outputting the corrected sentence acquired in the second acquisition step; An information processing method including:

6. The computer of the information processing device, a first acquisition means for acquiring text data and user attributes of a user who uses the text data; a reading means for reading a language rule corresponding to the user attribute acquired by the first acquiring means from reference data that defines a language rule for a language to be used depending on a type of user attribute; a generating means for generating query information including an instruction sentence for instructing correction of text included in the text data based on the linguistic rules read by the reading means; a second acquisition means for inputting the text data and the inquiry information generated by the generation means into a generation AI that is configured to modify the text in accordance with instructions in the inquiry information and generate a modified text, and acquiring the modified text generated by the generation AI; an output means for outputting the corrected sentence acquired by the second acquisition means; A program to function as a

Images