Information processing device, information processing method, and information processing system

Abstract

To appropriately check whether content conforms to a rule of a target for which the content should be checked.SOLUTION: An information processing device 1 includes a storage unit 12 that stores descriptive sentences of a plurality of rules in association with the respective rules, an acquisition unit 131 that acquires content which is a check target, an input unit 133 that executes an input process of inputting the content and descriptive sentences stored in the storage unit 12 in association with one or more target rules of the plurality of rules for which the content should be checked, to a large scale language model, and an output unit 134 that outputs a check result based on a sentence outputted from the large scale language model in response to the input process.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to an information processing device, an information processing method, and an information processing system for checking content. [Background technology]

[0002] 2. Description of the Related Art Conventionally, there is known an information determination device that checks whether the written content of a text sentence complies with standards such as laws, regulations, and the like, and outputs the results (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6621095 Summary of the Invention [Problem to be solved by the invention]

[0004] There are many different types of rules. Conventional techniques have been unable to determine whether a content, such as a text sentence or an image, is suitable for a given rule.

[0005] The present invention has been made in consideration of these points, and has an object to make it possible to appropriately check whether or not content complies with the standards for which the content should be checked. [Means for solving the problem]

[0006] A first aspect of the information processing device of the present invention is an information processing device for checking whether content complies with a specified norm, and includes: a memory unit that stores an explanatory text of the content of the norm in association with each of a plurality of norms; an acquisition unit that acquires the content to be checked; an input unit that executes an input process to input the content and the explanatory text stored in the memory unit in association with one or more norms among the plurality of norms for which the content is to be checked, into a large-scale language model; and an output unit that outputs a check result based on a sentence output by the large-scale language model in response to the input process.

[0007] The memory unit may store the explanatory text in association with a product category, the acquisition unit may further acquire product category information in association with the content and indicating a category of the product indicated by the content, and the input unit may input the explanatory text stored in the memory unit in association with a product category corresponding to the product category information to the large-scale language model.

[0008] The information processing device may further include an identification unit that identifies a category of the product indicated by the content, the memory unit may store the explanatory text in association with the product category, and the input unit may input the explanatory text stored in the memory unit in association with the product category identified by the identification unit to the large-scale language model.

[0009] The storage unit may store, as the explanation text, at least one of one or more appropriate sentence examples conforming to a standard and one or more inappropriate sentence examples not conforming to a standard, in association with the product category.

[0010] The acquisition unit may further acquire designation information for designating one or more usage norms among the plurality of norms against which the content is to be checked, and the input unit may input the explanatory text stored in the memory unit in association with the norm indicated by the designation information to the large-scale language model.

[0011] The acquisition unit may further acquire designation information for specifying one or more usage rules among the multiple rules against which the content is to be checked, the input unit may input the explanatory text stored in the memory unit in association with the one or more usage rules into the large-scale language model, and the output unit may output the check result including a reason for judgment, associated with each of the one or more usage rules.

[0012] The memory unit may store information for identifying one of the plurality of large-scale language models in association with each of the plurality of norms, and the input unit may input the explanatory text stored in the memory unit in association with the one or more norms to the large-scale language model associated with the one or more norms in the memory unit.

[0013] The information processing device may further have an identification unit that identifies, from the plurality of norms, a usage norm against which the content is to be checked by analyzing the content, and the input unit may input, to the large-scale language model, the explanatory text stored in the memory unit in association with the usage norm identified by the identification unit.

[0014] The input unit may further perform an additional input process to identify additional information regarding the content to be added based on the content output by the large-scale language model after the large-scale language model has output in response to the input process, and input the identified additional information into the large-scale language model, and the output unit may output a check result based on the sentence output by the large-scale language model in response to the additional input process.

[0015] An information processing method of a second aspect of the present invention is an information processing method executed by a computer for checking whether content complies with a specified standard, comprising the steps of: acquiring the content to be checked; reading from a memory unit that stores an explanatory text of the content of a standard associated with each of a plurality of standards, the explanatory text stored in the memory unit in association with one or more standards among the plurality of standards for which the content is to be checked; executing an input process to input the content and the explanatory text read from the memory unit into a large-scale language model; and outputting a check result based on a sentence output by the large-scale language model in response to the input process.

[0016] An information processing system of a third aspect of the present invention comprises an information terminal that accepts designation of content to be checked, and an information processing device for checking whether the content complies with a predetermined standard, wherein the information terminal has a communication unit that transmits the content to the information processing device, and a display unit that displays the check result received from the information processing device, and the information processing device has a memory unit that stores an explanatory text of the content of the standard in association with each of a plurality of standards, an acquisition unit that acquires the content to be checked, an input unit that executes an input process to input the content and the explanatory text stored in the memory unit in association with one or more standards among the plurality of standards for which the content is to be checked, into a large-scale language model, and an output unit that outputs the check result based on a sentence output by the large-scale language model in response to the input process to the information terminal. Effect of the Invention

[0017] According to the present invention, it is possible to appropriately check whether or not content conforms to the standards for which the content should be checked. [Brief description of the drawings]

[0018] [Figure 1] 1 is a diagram for explaining an overview of an information processing system S. [Diagram 2] FIG. 11 is a diagram illustrating an example of a relationship between a product category and a reference category. [Diagram 3] 1 is a diagram illustrating a configuration of an information processing device 1. [Figure 4] FIG. 2 is a diagram showing an example of a norm explanatory sentence stored in the storage unit 12. [Diagram 5] 13A to 13C are diagrams showing examples of content and norm explanatory sentences input by the input unit 133 to the LLM unit 3, and examples of judgment sentences output by the LLM unit 3. [Figure 6] 13A to 13C are diagrams showing examples of content and norm explanatory sentences input by the input unit 133 to the LLM unit 3, and examples of judgment sentences output by the LLM unit 3. [Figure 7] 4 is a flowchart showing a process flow of the information processing device 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] [Outline of Information Processing System S] FIG. 1 is a diagram for explaining the outline of an information processing system S. The information processing system S is a system for checking whether content including product descriptions conforms to laws, regulations, administrative guidance, and norms such as voluntary industry standards. The content includes at least one of text and images (including video), and is used as an advertisement for the product to consumers of the product. The content is, for example, text or images that are planned to be included on the product's packaging, product catalog, or homepage. The content may be audio data corresponding to the sound of the text being read aloud.

[0020] The information processing system S includes an information processing device 1, an information terminal 2, and an LLM (Large Language Model) unit 3. The information processing device 1 can transmit and receive data between the information terminal 2 and the LLM unit 3 via a network N. The information processing device 1 is a computer, such as a server, that checks whether or not the content to be checked complies with a predetermined standard.

[0021] Although details will be described later, the information processing device 1 stores an explanatory text for each of a plurality of norms (hereinafter, sometimes referred to as a "norm explanatory text"). The norm explanatory text may be the original text of the norm (e.g., a legal provision), a summary of the norm's content, example sentences that conform to the norm, or example sentences that do not conform to the norm, etc. The norms that must be observed differ for each product category. Therefore, the information processing device 1 may store an explanatory text for each of a plurality of norms in association with the product category.

[0022] Fig. 2 is a diagram showing an example of the relationship between product categories and normative categories. In Fig. 2, medical drugs, over-the-counter drugs, medical devices, quasi-drugs, cosmetics, and health foods are exemplified as product categories to be checked by the information processing device 1, but the information processing device 1 may check contents related to other product categories.

[0023] In FIG. 2, legal regulations, administrative guidance, self-regulation, and expert knowledge are shown as examples of norm categories, but the information processing device 1 may store explanatory text of norms in other categories and check content targeting norms in other categories.

[0024] Returning to Fig. 1, the information terminal 2 is a terminal used by a user who wishes to check content, and is, for example, a personal computer or a smartphone. The information terminal 2 accepts input of the content to be checked. The information terminal 2 has a communication unit that transmits the input content to the information processing device 1, a display unit (for example, a display or a light-emitting diode) that displays the check result received from the information processing device 1, and a CPU (Central Processing Unit) that controls the communication unit and the display unit.

[0025] The LLM unit 3 has a large-scale language model. The large-scale language model is a computer language model that is composed of a neural network with many parameters and is created by machine learning using a large amount of unlabeled text data and training data. The LLM unit 3 may have a large-scale visual language model (LVM: Large Vision Model) instead of or in addition to the large-scale language model. The visual language model is created by machine learning using text-attached image data as training data, and the LLM unit 3 may receive image data as input and output text data or image data. The LLM unit 3 is composed of, for example, one or more computers different from the information processing device 1, but the information processing device 1 may have the function of the LLM unit 3. The LLM unit 3 may have a plurality of different large-scale language models trained using different training data.

[0026] Hereinafter, the basic operation of the information processing system S will be described with reference to Fig. 1. When the information terminal 2 accepts an input of content to be checked, it transmits the input content together with request data indicating a request for checking to the information processing device 1 ((1) in Fig. 1). The request data includes, for example, information indicating the category of the product to which the content is related and the standard category of the subject to be checked.

[0027] When the information processing device 1 receives the content together with the request data, the information processing device 1 selects, from among a plurality of norms, one or more norms that correspond to a predetermined norm or a norm specified in the request data. Also, from among a plurality of product categories, the information processing device 1 selects, from among a plurality of product categories, a predetermined product category, a product category specified in the request data, or one or more product categories identified by analyzing the content.

[0028] The information processing device 1 transmits the content received from the information terminal 2 and the norm explanatory text corresponding to the selected one or more norms and one or more product categories to the LLM unit 3 ((2) in FIG. 1). At this time, the information processing device 1 may transmit the content and the norm explanatory text to the LLM unit 3 in a state in which the LLM unit 3 can identify the content to be checked and the norm explanatory text. For example, the information processing device 1 transmits the content and the norm explanatory text to the LLM unit 3 with a flag attached to the content to be checked indicating that it is content and with a flag attached to the norm explanatory text indicating that it is a norm explanatory text.

[0029] The information processing device 1 may transmit the content and the norm explanation sentence in a predetermined order to the LLM unit 3. As an example, the information processing device 1 first transmits the content to the LLM unit 3, and thereafter transmits a plurality of norm explanation sentences to the LLM unit 3, so that the LLM unit 3 can appropriately recognize the content to be checked.

[0030] The LLM unit 3 judges whether the content complies with the norm based on the content and the norm explanation sentence received from the information processing device 1. The LLM unit 3 transmits a judgment statement indicating the judgment result to the information processing device 1 ((3) in FIG. 1). The judgment statement may include the judgment result, the judgment reason, or an example of correction in association with the checked norm.

[0031] When the information processing device 1 receives the judgment statement from the LLM unit 3, it transmits a check result based on the judgment statement to the information terminal 2 ((4) in FIG. 1). The check result may be the same as the judgment statement, or may be a sentence that is a modified version of the judgment statement.

[0032] The information terminal 2 allows a user who wishes to check the content to confirm the check result by displaying the check result received from the information processing device 1. In this way, in the information processing system S, even if there are multiple standards that the content must comply with, it becomes possible to appropriately check whether the content complies with the target standards for which the content should be checked.

[0033] [Configuration of information processing device 1] 3 is a diagram showing a configuration of the information processing device 1. The information processing device 1 has a communication unit 11, a storage unit 12, and a control unit 13. The control unit 13 has an acquisition unit 131, an identification unit 132, an input unit 133, and an output unit 134. The control unit 13 does not necessarily have to have the identification unit 132.

[0034] The communication unit 11 includes a communication interface for transmitting and receiving data to and from the information terminal 2 and the LLM unit 3 via the network N. The communication unit 11 receives content transmitted from the information processing device 1, and inputs the received content to the acquisition unit 131. The communication unit 11 transmits the content and the norm explanatory sentence input from the input unit 133 to the LLM unit 3. When the communication unit 11 receives a judgment sentence from the LLM unit 3, it inputs the received judgment sentence to the input unit 133. In addition, the communication unit 11 transmits the check result input from the output unit 134 to the information terminal 2.

[0035] The memory unit 12 has storage media such as a ROM (Read Only Memory), a RAM (Random Access Memory), and an SSD (Solid State Drive). The memory unit 12 stores programs executed by the control unit 13. The memory unit 12 also stores various data used by the control unit 13 to check content. The memory unit 12 stores an explanatory text of the content of a norm in association with each of a plurality of norms. The memory unit 12 may store a plurality of norm explanatory texts in association with a plurality of product categories. The memory unit 12 may store at least one of one or more appropriate sentence examples that conform to the norm and one or more inappropriate sentence examples that do not conform to the norm as explanatory text in association with a product category.

[0036] 4 is a diagram showing examples of norm explanatory sentences stored in the storage unit 12. Norm explanatory sentence E1 is an example of a standard original sentence. Norm explanatory sentence E2 is an example of a sentence that conforms to the norm and an example of a sentence that does not conform to the norm. The norm explanatory sentence is associated with at least one of a product category, a norm category, a content category, and an appropriate category.

[0037] Normative explanation text E1 is based on the "Standards for Appropriate Advertising of Drugs, etc." and is associated with the product categories covered by the "Standards for Appropriate Advertising of Drugs, etc.", which are "medical drugs," "over-the-counter drugs," "medical devices," "quasi-drugs," and "cosmetics." Normative explanation text E1 is also associated with the normative category "legal regulations" and the content category "original text."

[0038] Norm explanatory sentence E2 is an example sentence related to cosmetics, and is therefore associated with the product category "Cosmetics." Norm explanatory sentence E2 is associated with, for example, the norm category "Administrative Guidance." Norm explanatory sentence E2 is a case study sentence, and is therefore associated with the content category "Case Study Sentence." Norm explanatory sentence E2 also includes appropriate and inappropriate sentence examples, and is therefore associated with the appropriate categories "Appropriate" and "Inappropriate." Memory unit 12 stores a large number of norm explanatory sentences structured in this manner.

[0039] However, depending on the type of teacher data used for learning by the large-scale language model, there may be some fields in which the large-scale language model excels and some fields in which it does not excel. Therefore, the storage unit 12 may store information for identifying one of the large-scale language models in association with each of the multiple norms. The storage unit 12 may store information for identifying one of the large-scale language models held by the LLM unit 3 as the large-scale language model to be used in association with each of the multiple product categories.

[0040] The control unit 13 includes, for example, a CPU (Central Processing Unit). The control unit 13 executes a program stored in the storage unit 12, thereby functioning as an acquisition unit 131, a specification unit 132, an input unit 133, and an output unit .

[0041] The acquisition unit 131 acquires the content to be checked via the communication unit 11. The acquisition unit 131 notifies the specification unit 132 and the input unit 133 of the details of the acquired content. The acquisition unit 131 may store the acquired content in the storage unit 12, thereby notifying the specification unit 132 and the input unit 133 of the details of the content.

[0042] The acquisition unit 131 may further acquire product category information indicating a category of a product indicated by the content in association with the content. The acquisition unit 131 acquires the product category information included in the request data received from the information processing device 1 via the communication unit 11, for example. The acquisition unit 131 notifies the identification unit 132 of the acquired product category information.

[0043] The acquisition unit 131 may further acquire designation information for designating one or more usage rules for checking the content among the multiple rules. The acquisition unit 131 acquires the designation information included in the request data received from the information processing device 1 via the communication unit 11, for example. The acquisition unit 131 notifies the input unit 133 of the acquired designation information.

[0044] The identification unit 132 identifies the category of the product indicated by the content. The identification unit 132 identifies the category of the product based on, for example, product category information notified from the acquisition unit 131. The identification unit 132 may identify the category of the product indicated by the content by searching for words included in the content in a word database associated with the product category stored in the storage unit 12. The identification unit 132 may input the content and a text sentence for causing the LLM unit 3 to infer the product category to the LLM unit 3, and identify the product category based on the text sentence output from the LLM unit 3. The identification unit 132 may input the content and a text sentence for causing the LLM unit 3 to infer the use of the content to the LLM unit 3, and identify the use of the content, such as whether it is for packaging or for a press release, based on the text sentence output from the LLM unit 3.

[0045] The identification unit 132 may identify a usage rule to be checked for the content from among the multiple rules. For example, the identification unit 132 analyzes the content to identify a noun indicating a product category included in the content, and identifies the usage rule to be checked by referring to a rule category stored in the storage unit 12 in association with the identified noun.

[0046] As an example, when the content contains the noun "medicine," the identification unit 132 identifies a norm associated with a medical drug or an over-the-counter drug and stored in the storage unit 12 as the usage norm. When the content contains the noun "skin," the identification unit 132 identifies a norm associated with a quasi-drug or a cosmetic and stored in the storage unit 12 as the usage norm.

[0047] The input unit 133 executes an input process for inputting the content and a norm explanatory sentence stored in the storage unit 12 in association with one or more norms that are the targets for checking the content among the multiple norms, to the LLM unit 3. After executing the input process, the input unit 133 acquires a judgment statement transmitted from the LLM unit 3. The input unit 133 notifies the output unit 134 of the acquired judgment statement.

[0048] The output unit 134 acquires the determination statement from the input unit 133, and outputs a check result based on the determination statement output by the LLM unit 3 in response to the input processing. The output unit 134 transmits the check result to the information terminal 2 via the communication unit 11, for example, but may also transmit the check result to a terminal other than the information terminal 2 or output it to a printer.

[0049] The output unit 134 may output the check result including the judgment reason in association with each of one or more usage rules. The judgment reason is, for example, the reason why the content does not comply with the rule, which is included in the judgment statement received from the LLM unit 3. By outputting such a check result by the output unit 134, the user who requested the content to be checked can understand the reason why the content does not comply with the rule, making it easier to correct the content.

[0050] The output unit 134 may output the judgment statement output by the LLM unit 3 as the check result as is, or may process the judgment statement before outputting the check result. For example, if the content of the judgment statement does not match the format specified by the user of the information terminal 2, the output unit 134 creates a check result in which the format is processed. As an example, the output unit 134 creates a check result in which the content of the judgment statement is added as a comment to a file including the content. The output unit 134 may output a check result created by merging the judgment statements output by the LLM unit 3 multiple times into one.

[0051] The output unit 134 may output a proposal for correcting the portion that does not conform to the standard based on the check result received from the LLM unit 3. For example, when the output unit 134 obtains a judgment statement from the LLM unit 3 that the portion "become slim" does not conform to the standard, the output unit 134 outputs a proposal for correcting the portion to "become able to replenish nutrients while exercising" based on the judgment reason included in the judgment statement. By the output unit 134 outputting such a proposal for correction, it becomes easier for the user who requested the content check to appropriately correct the content.

[0052] The operations of the input unit 133 and the output unit 134 will now be described in detail. The input unit 133 inputs, for example, together with the content, a norm explanatory sentence stored in the storage unit 12 in association with a product category corresponding to the product category information acquired by the acquisition unit 131 in association with the content, to the LLM unit 3. The input unit 133 may input, together with the content, to the LLM unit 3, a norm explanatory sentence stored in the storage unit 12 in association with a product category identified by the identification unit 132 based on the content.

[0053] The input unit 133 may input the norm explanation sentence stored in the storage unit 12 in association with the norm indicated by the designation information acquired by the acquisition unit 131 to the LLM unit 3. In this case, the input unit 133 selects one or more norms indicated by the designation information from among a plurality of norms corresponding to the content, and inputs one or more norm explanation sentences corresponding to the selected one or more norms to the LLM unit 3. By the input unit 133 operating in this manner, the accuracy of determining whether the content complies with the norm that the user wants to check whether the content is appropriate is improved.

[0054] By the input unit 133 operating in this manner, the accuracy of determining whether the content complies with the norms that should be observed by advertisements for products introduced in the content, daily reports from medical representatives of medical drugs, materials used by doctors in lectures, etc. is improved. Also, since the input unit 133 inputs norm explanatory text corresponding to the norms for which the conformity of the content should be checked to the LLM unit 3 and does not input unnecessary explanatory text of other norms to the LLM unit 3, the accuracy of determining whether the content complies with the norms for which the conformity of the content should be checked is improved. Also, since the input unit 133 does not input unnecessary explanatory text to the LLM unit 3, unnecessary inference calculations are reduced, making it possible to reduce costs and shorten the determination time.

[0055] The input unit 133 may input, together with the content and the norm explanatory sentence, a text sentence indicating a product category corresponding to the product category information or a product category identified by the identification unit 132 to the LLM unit 3. By the input unit 133 notifying the LLM unit 3 of the product category in this manner, it becomes easier for the LLM unit 3 to output a judgment sentence according to the product category.

[0056] Before inputting the content to the LLM unit 3 or after inputting the content to the LLM unit 3, the input unit 133 may sequentially input a plurality of norm explanatory sentences stored in the storage unit 12 in association with one or more usage norms to the LLM unit 3. In this case, the input unit 133 inputs information indicating the relationship between the plurality of norm explanatory sentences and the content to the LLM unit 3 so that it becomes clear which content the plurality of norm explanatory sentences are used to check.

[0057] As an example, the input unit 133 inputs content to the LLM unit 3, inputs the number of norm explanations to be used for checking the content to the LLM unit 3, and then sequentially inputs the input number of norm explanations to the LLM unit 3. The input unit 133 may attach common identification information to the content and the multiple norm explanations to be used for checking the content. By operating the input unit 133 in this manner, the LLM unit 3 can recognize the multiple norm explanations to be used for checking the content, and the LLM unit 3 can perform a check using a large number of norm explanations.

[0058] The input unit 133 may input the explanatory sentence stored in the storage unit 12 in association with one or more norms to a large-scale language model in the storage unit 12 associated with one or more norms identified by the identification unit 132 as targets for checking the content, among the multiple large-scale language models held by the LLM unit 3. The input unit 133 may input the norm explanatory sentence stored in the storage unit 12 in association with one or more norms to a large-scale language model in the storage unit 12 associated with one or more product categories related to the content, among the multiple large-scale language models. By operating the input unit 133 in this manner, the accuracy of determining the suitability of the content is improved when norms or product categories suitable for each of the multiple large-scale language models are different.

[0059] To check one piece of content, the input unit 133 may repeatedly input a text sentence to the LLM unit 3 multiple times. Specifically, after the LLM unit 3 outputs in response to the input process, the input unit 133 may further execute additional input processing to identify additional information related to the content to be added based on the content output by the LLM unit 3 and input the identified additional information to the LLM unit 3. In this case, the output unit 134 outputs a check result based on the sentence output by the LLM unit 3 in response to the additional input processing.

[0060] As an example, assume that the input unit 133 inputs the content and the norm explanatory sentence to the LLM unit 3, and then receives a determination statement from the LLM unit 3 indicating that the content may not conform to the norm, but there is no problem if there is evidence. In response to this, the input unit 133 may input, for example, a text sentence indicating the contents of the existing evidence to the LLM unit 3. In response to this, when the input unit 133 receives a further determination statement from the LLM unit 3 (for example, a determination statement saying "If that evidence is available, the content conforms to the norm"), the output unit 134 outputs, together with the content, a check result indicating that the content conforms to the norm.

[0061] The input unit 133 may create a proposed revision for the content based on the judgment reason received from the LLM unit 3, and input the created proposed revision to the LLM unit 3. On the other hand, when the input unit 133 receives a judgment statement from the LLM unit 3 that the revised content complies with the norm, the input unit 133 instructs the output unit 134 to output the check result. When the input unit 133 receives a judgment statement from the LLM unit 3 that the revised content also does not comply with the norm, the input unit 133 creates a further proposed revision based on the judgment reason, and inputs the proposed revision to the LLM unit 3. The input unit 133 repeats the process of correcting the content until it receives a judgment statement from the LLM unit 3 that the revised content complies with the norm, and the output unit 134 can present the user with a proposed revision for the content that complies with the norm.

[0062] [Examples of content, normative statements and judgment statements] 5 and 6 are diagrams showing examples of content and norm explanatory sentences input by the input unit 133 to the LLM unit 3, and examples of judgment sentences output by the LLM unit 3. FIG.

[0063] 5, the content input by the user includes the expression "It's tea." The identification unit 132 then refers to a database showing the relationship between words and product categories stored in the storage unit 12, and identifies that the product to which the content is related is "health food." The input unit 133 inputs a norm explanatory sentence related to health food to the LLM unit 3.

[0064] The judgment statement output by the LLM unit 3 indicates the judgment result indicating that the content does not comply with the standard and the reason for the judgment. The LLM unit 3 outputs a judgment statement corresponding to health foods based on the words, expressions, nuances, and overall intention contained in the content and the standard explanation statement. In addition, since the content contains pharmaceutical expressions, the LLM unit 3 also outputs a judgment statement for when the product category is pharmaceuticals, etc., as reference information.

[0065] 6, the content input by the user includes expressions related to hair and ingredients. The identification unit 132 then refers to a database indicating the relationship between words stored in the storage unit 12 and product categories, thereby identifying that the product to which the content relates is cosmetics. The input unit 133 inputs a norm explanation sentence related to cosmetics to the LLM unit 3. The judgment sentence output by the LLM unit 3 indicates the judgment result as well as the norm on which the judgment is based.

[0066] [Processing flow in information processing device 1] 7 is a flowchart showing the flow of processing of the information processing device 1. The flowchart shown in FIG.

[0067] The acquisition unit 131 acquires content inputted in the information terminal 2 (S11). The identification unit 132, for example, analyzes words contained in the inputted content to identify a product category to which the content is related (S12). The identification unit 132 also identifies a norm stored in the storage unit 12 in association with the identified product category as a usage norm to be checked (S13).

[0068] Next, the input unit 133 selects one or more norm explanation sentences corresponding to the usage norm identified by the identification unit 132 from the multiple norm explanation sentences stored in the storage unit 12 (S14). The input unit 133 inputs the selected content and the norm explanation sentence to the LLM unit 3 (S15).

[0069] When the input unit 133 acquires the determination statement from the LLM unit 3 (S16), it judges whether or not the acquired determination statement needs to be corrected (S17). When the input unit 133 judges that the determination statement does not need to be corrected (NO in S17), it notifies the output unit 134 of the determination statement. The output unit 134 outputs the check result including the determination statement (S18).

[0070] When the input unit 133 determines that the judgment sentence needs to be corrected (YES in S17), it inputs additional information required to correct the judgment sentence to the LLM unit 3 (S19). Thereafter, the input unit 133 repeats the processes of S16, S17, and S19 until the judgment sentence does not need to be corrected. Through the above procedure, the information processing device 1 can notify the user of the check result in which it has been appropriately determined whether or not the content input by the user at the information terminal 2 complies with the criteria required for checking the content. Furthermore, when the input unit 133 determines that the judgment sentence needs to be corrected, the output unit 134 can notify the user of a high-quality check result by inputting additional information to the LLM unit 3 until the judgment sentence does not need to be corrected.

[0071] [Effects of information processing device 1] As described above, the information processing device 1 has an input unit 133 that executes an input process of inputting to the LLM unit 3 the content input by the user to the information terminal 2 and a norm explanatory sentence stored in the storage unit 12 in association with one or more norms for which the content is to be checked among a plurality of norms, and an output unit 134 that outputs a check result based on the sentence output by the LLM unit 3 in response to the input process. With such a configuration, the information processing device 1 can appropriately check whether or not the content complies with the norm for which the content should be checked.

[0072] Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments, and various modifications and changes are possible within the scope of the gist of the present invention. For example, all or part of the device can be configured by distributing or integrating functionally or physically in any unit. In addition, new embodiments resulting from any combination of multiple embodiments are also included in the embodiments of the present invention. The effect of the new embodiment resulting from the combination combines the effect of the original embodiment. [Explanation of symbols]

[0073] 1. Information processing device 2. Information terminal 3 LLM Department 11 Communications Department 12 Storage section 13 Control section 131 Acquisition Department 132 Specific part 133 Input section 134 Output section

Claims

1. An information processing device for checking whether or not content conforms to a predetermined standard, a storage unit that stores an explanatory text of the content of each of the plurality of norms in association with each of the plurality of norms; An acquisition unit that acquires the content to be checked; an input unit that executes an input process to input the content and the explanatory sentence stored in the storage unit in association with one or more norms that are targets for checking the content among the plurality of norms, into a large-scale language model; an output unit that outputs a check result based on a sentence output by the large-scale language model in response to the input process; An information processing device having the above configuration.

2. The storage unit stores the description in association with a product category, The acquisition unit further acquires, in association with the content, product category information indicating a category of a product indicated by the content; the input unit inputs the description stored in the storage unit in association with the product category corresponding to the product category information to the large-scale language model; The information processing device according to claim 1 .

3. The content management system further includes an identification unit that identifies a category of a product indicated by the content, The storage unit stores the description in association with a product category, the input unit inputs the description stored in the storage unit in association with the product category identified by the identification unit into the large-scale language model; The information processing device according to claim 1 .

4. the storage unit stores, as the description, at least one of one or more appropriate sentence examples conforming to a standard and one or more inappropriate sentence examples not conforming to the standard, in association with the product category; 4. The information processing device according to claim 2 or 3.

5. The acquisition unit further acquires designation information for designating the one or more usage rules that are to be checked for the content among the plurality of rules, The input unit inputs the explanatory sentence stored in the storage unit in association with the norm indicated by the specification information into the large-scale language model. The information processing device according to claim 1 .

6. The acquisition unit further acquires designation information for designating one or more usage rules that are to be checked for the content from among the plurality of rules; The input unit inputs the explanatory sentence stored in the storage unit in association with the one or more usage rules into the large-scale language model; The output unit outputs the check result including a reason for judgment in association with each of the one or more rules of use. The information processing device according to claim 1 .

7. the storage unit stores information for identifying any one of the plurality of large-scale language models in association with each of the plurality of norms; The input unit inputs the explanatory sentence stored in the storage unit in association with the one or more models into the large-scale language model associated with the one or more models in the storage unit. The information processing device according to claim 1 .

8. The method further includes an identification unit that identifies a usage rule that is a target for checking the content among the plurality of rules by analyzing the content, the input unit inputs the explanatory sentence stored in the storage unit in association with the usage rule identified by the identification unit into the large-scale language model; The information processing device according to claim 1 .

9. the input unit further performs an additional input process of identifying additional information related to the content to be added based on the content output by the large-scale language model after the large-scale language model outputs the content in response to the input process, and inputting the identified additional information into the large-scale language model; the output unit outputs a check result based on the sentence output by the large-scale language model in response to the additional input process. The information processing device according to claim 1 .

10. 1. A computer-implemented information processing method for checking whether content complies with predetermined criteria, comprising the steps of: obtaining the content to be checked; a step of reading, from a storage unit that stores an explanatory text of the content of a rule in association with each of a plurality of rules, the explanatory text stored in the storage unit in association with one or more rules that are targets for checking the content among the plurality of rules; executing an input process for inputting the content and the explanatory text read from the storage unit into a large-scale language model; outputting a check result based on the sentence output by the large-scale language model in response to the input process; An information processing method.

11. An information terminal that accepts designation of content to be checked, and an information processing device that checks whether the content complies with a predetermined standard, The information terminal includes: a communication unit that transmits the content to the information processing device; a display unit that displays the check result received from the information processing device; having The information processing device includes: a storage unit that stores an explanatory text of the content of each of the plurality of norms in association with each of the plurality of norms; An acquisition unit that acquires the content to be checked; an input unit that executes an input process to input the content and the explanatory sentence stored in the storage unit in association with one or more norms that are targets for checking the content among the plurality of norms, into a large-scale language model; an output unit that outputs the check result based on the sentence output by the large-scale language model in response to the input process to the information terminal; An information processing system having:

Images