Information processing device, information processing method, program, and recording medium

Abstract

The present invention structures inputted address information in a more suitable manner. This information processing device comprises: a reception means that receives input of input information including information corresponding to one among first address information defined so that the position of an object is subdivided in stages into predetermined management units, and code information allowing unique identification of the position indicated by the first address information; and a structuring means that, by inputting the input information to a model trained on the basis of machine learning, causes the model to output second address information subjected to structuring so that components for each predetermined management unit constituting the first address information are classified. The model is trained on the basis of machine learning using learning data that includes: information corresponding to a type among third address information that has not been subjected to the structuring, and code information corresponding to the third address information; and fourth address information obtained by subjecting the third address information to the structuring.

Description

Information processing device, information processing method, program, and recording medium

[0001] The present disclosure relates to an information processing device, an information processing method, a program, and a recording medium.

[0002] When providing services, various systems may accept and use address information (e.g., information indicating the location of an object by administrative division such as prefecture, city, ward, town, or village) input from a user, the address information indicating the location of the destination, location, or residence. In such cases, various methods for inputting address information have been adopted, such as a method of individually inputting address information into input boxes provided for each component constituting the address information, such as a division such as prefecture, city, ward, town, or village, or a method of collectively inputting a series of components into a single input box. Furthermore, various technologies for registering and managing address information received as input in a database or the like have been considered. For example, Patent Document 1 discloses an example of a technology that supports a user's input operation when receiving address information input from a user, assuming that the address information is registered and managed as data.

[0003] Japanese Unexamined Patent Publication No. 7-191988

[0004] On the other hand, in situations where address information is registered and managed in a database, etc., it may be preferable to manage the individual components that make up the address information, such as prefectures, cities, wards, towns, and villages, separately. For example, in recent years, it has been planned to transition the data format for foreign remittances to a format that complies with the ISO 20022 standard, and in complying with this standard, there is a demand for structuring address information so that address information can be managed by separating it into components such as prefectures, cities, wards, towns, and villages.

[0005] In view of the above problems, an object of the present invention is to realize structuring of input address information in a more suitable manner.

[0006] The information processing device of the present invention has a receiving means for receiving input of input information including first address information defined so that the location of the target is subdivided in stages by predetermined management units and information regarding the format of the address information, and a structuring means for inputting the input information to a model trained based on machine learning, and causing the model to output second address information structured according to the format of the address information so that the components of each of the predetermined management units that make up the first address information are separated, and the model is trained based on machine learning using training data including third address information defined so that the location of the target is subdivided in stages by the management units and that is not structured, and fourth address information in which the third address information is structured according to the format of the address information.

[0007] According to the present invention, it is possible to realize structuring of input address information in a more suitable manner.

[0008] FIG. 1 is a diagram showing an example of a system configuration of an information processing system. FIG. 2 is a diagram showing an example of an address structure model. FIG. 3 is a diagram showing an example of conditions related to structuring of address information. FIG. 4 is a diagram showing an example of a hardware configuration of an information processing device. FIG. 5 is a diagram showing an example of a functional configuration of the information processing system. FIG. 6 is a diagram showing an example of an application scene of the information processing system. FIG. 7 is a diagram showing an example of a functional configuration of the information processing device. FIG. 8 is a diagram showing an example of an application scene of the information processing system.

[0009] Preferred embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configuration are designated by the same reference numerals, and redundant description will be omitted. In this disclosure, "address information" refers to information for managing the location of an object, such as a destination, location, or residence, by dividing it into predetermined management units, such as administrative divisions representing place names, such as prefectures, cities, wards, towns, and villages, or addresses. While the following description will be based on Japanese place names as a representative example, the same method can be used for administrative divisions, such as states and provinces, in other countries, as will be described later.

[0010] 1 to 3, an overview of an information processing system 1 according to an embodiment of the present disclosure will be described. The information processing system 1 according to this embodiment has a function of receiving input of location information indicating the location of an object such as a destination, a location, or a residence, structuring the location information by dividing it into components for management by gradually subdividing it in predetermined management units, and then outputting the structured address information. To realize this function, the information processing system 1 realizes processing related to the structuring of address information using a learning model that has been trained based on machine learning such as so-called AI.

[0011] For example, Figure 1 is a diagram showing an example of a schematic system configuration of an information processing system 1 according to this embodiment. The information processing system 1 includes a learning device 100, an information processing device 200, and a terminal device 300. The learning device 100 executes various processes related to learning a learning model (hereinafter also referred to as an address structure model) that performs structuring processes on input address information. An address structure model 500 is a schematic representation of the learning model learned by the learning device 100.

[0012] Here, with reference to FIG. 2 , an example of an address structure model learned by the learning device 100 will be described. The learning model 510 shown in FIG. 2 is an example of the address structure model 500 shown in FIG. 1 . Address information input as a series of information (e.g., address information input in plain text) is input, and the learning model 510 classifies the address information into components such as prefectures, cities, wards, towns, villages, block numbers, and so on, and links these components to the corresponding component types to structure and output the address information so that it can be individually referenced. For example, in the example shown in FIG. 2 , address information input as "Toranomon, Minato Ward, Tokyo..." is classified into components based on the administrative divisions of the target country, such as administrative districts, block numbers, and block numbers that make up the address information, such as "Tokyo," "Minato Ward," "Toranomon," and so on. As a specific example, information indicated as "Tokyo" is classified and managed as information indicating a "prefecture." Information indicated as "Minato Ward" is classified and managed as information indicating a "ward" within a city, ward, town, or village. Similarly, information indicated as "Toranomon" is classified and managed as information indicating a "town" within a city, ward, town, or village. In this way, the input address information is divided into components of a predetermined management unit, and then structured into data in a predetermined format (for example, XML, etc.) and managed. By structuring the address information in this way, it becomes possible to mechanically extract the components contained in the address information for each individual management unit.

[0013] In this disclosure, address information before being structured by an address structure model (e.g., the learning model 510 shown in FIG. 2 ) (i.e., address information that serves as input to the address structure model) is also referred to as unstructured address information for convenience. Furthermore, address information structured by the address structure model is also referred to as structured address information for convenience. Note that unstructured address information is not limited to the plain text address information exemplified in FIG. 2 , but may also include address information that does not satisfy predefined structuring conditions, such as address information in which only some information (e.g., prefecture, city, ward, town, or village) is classified.

[0014] Here, an example of conditions for structuring address information will be described with reference to FIG. 3 . FIG. 3 shows an example of conditions for structuring address information defined in the ISO 20022 standard published by the Payments Market Practice Group. The example shown in FIG. 3 shows the correspondence between the types of components that can make up address information managed by the standard and the types of components that can make up address information managed in each country. The address structure model according to this embodiment may, for example, apply a structuring process to input unstructured address information in accordance with the structuring conditions illustrated in FIG. 3 and output the resulting structured address information.

[0015] The address structure model according to this embodiment can be constructed based on supervised learning using, as learning data, unstructured address information and structured address information that indicates the correct answer obtained by structuring the unstructured address information. The mechanism for learning the address structure model will be described in detail later.

[0016] Referring again to FIG. 1 , as described above, the address structure model 500 learned by the learning device 100 is applied as a learning model used by the information processing device 200 to structure address information. The information processing device 200 accepts input of address information (unstructured address information), generates structured address information that is structured so as to separate the address information into components for each predetermined management unit, and outputs the structured address information to a predetermined output destination. Specifically, the address structure model 500 applied to the information processing device 200 is responsible for processing related to the structuring of address information. The information processing device 200 inputs the unstructured address information received as input to the address structure model 500, causing the address structure model 500 to output structured address information obtained by structuring the unstructured address information. The information processing device 200 then outputs the structured address information to a predetermined output destination.

[0017] The terminal device 300 serves as an input / output interface when a user uses the information processing device 200. As a specific example, the terminal device 300 may accept input of address information (unstructured address information) to be structurized and transmit the address information to the information processing device 200 as the target for structuration. Furthermore, the terminal device 300 may receive structured address information as a result of structuring the address information transmitted as input from the information processing device 200 and present the structured address information to the user. As another example, the terminal device 300 may extract components for each predetermined management unit contained in the structured address information received from the information processing device 200 and automatically input the components into an input field of a predetermined UI. Applying such control makes it possible to realize, for example, a control in which address information is input in plain text from a user and then each component constituting the address information is automatically input into an input amount field individually provided for each type of component.

[0018] The overview of the information processing system 1 according to an embodiment of the present disclosure has been described above with reference to FIGS.

[0019] <Hardware Configuration> An example of the hardware configuration of an information processing device 900 that can be used as a device (e.g., the learning device 100, the information processing device 200, and the terminal device 300) constituting the information processing system according to this embodiment will be described with reference to FIG. 4 . As shown in FIG. 4 , the information processing device 900 according to this embodiment includes a CPU (Central Processing Unit) 910, a ROM (Read Only Memory) 920, and a RAM (Random Access Memory) 930. The information processing device 900 also includes an auxiliary storage device 940 and a network I / F 970. The information processing device 900 may also include at least one of an output device 950 and an input device 960. The CPU 910 , ROM 920 , RAM 930 , auxiliary storage device 940 , output device 950 , input device 960 , and network I / F 970 are interconnected via a bus 980 .

[0020] The CPU 910 is a central processing unit that controls various operations of the information processing device 900. For example, the CPU 910 may control the operation of the entire information processing device 900. The ROM 920 stores control programs, boot programs, and the like that can be executed by the CPU 910. The RAM 930 is the main storage memory of the CPU 910, and is used as a work area or a temporary storage area for expanding various programs.

[0021] The auxiliary storage device 940 stores various data and various programs. The auxiliary storage device 940 is realized by a storage device that can temporarily or permanently store various data, such as a non-volatile memory such as a hard disk drive (HDD) or a solid state drive (SSD).

[0022] The output device 950 is a device that outputs various types of information and is used to present various types of information to a user. For example, the output device 950 is realized by a display device such as a display. In this case, the output device 950 presents information to a user by displaying various types of display information. As another example, the output device 950 may be realized by an audio output device that outputs sounds such as voices and electronic sounds. In this case, the output device 950 presents information to a user by outputting sounds such as voices and electronic sounds. Furthermore, the device used as the output device 950 may be changed as appropriate depending on the medium used to present information to a user.

[0023] The input device 960 is used to receive various instructions from the user. In this embodiment, the input device 960 includes input devices such as a mouse, a keyboard, and a touch panel. As another example, the input device 960 may include a sound collection device such as a microphone, and collect voices uttered by the user. In this case, various analysis processes such as acoustic analysis and natural language processing are performed on the collected voices, and the contents of the voices are recognized as instructions from the user. Furthermore, the device applied as the input device 960 may be changed as appropriate depending on the method for recognizing instructions from the user. Furthermore, multiple types of devices may be applied as the input device 960.

[0024] The network I / F 970 is used for communication with external devices via a network. Note that the device used as the network I / F 970 may be changed as appropriate depending on the type of communication path and the communication method used.

[0025] The CPU 910 deploys a program stored in the ROM 920 or the auxiliary storage device 940 to the RAM 930 and executes this program, thereby realizing the functional configuration and processing of the information processing system 1 shown in Figures 5, 7, and 8.

[0026] An example of the hardware configuration of the information processing device 900 that can be used as a device constituting the information processing system 1 according to this embodiment has been described above with reference to FIG.

[0027] <Functional configuration and processing> Referring to Figure 5, an example of the functional configuration and processing of the information processing system 1 according to this embodiment will be described, focusing in particular on the parts related to learning the address structure model 510 and the parts related to the operation of the address structure model 510 (the parts related to structuring address information).

[0028] First, an example of the functional configuration and processing of the learning device 100 will be described, focusing particularly on the parts related to learning the address structure model 510. The learning device 100 includes a learning unit 109 and learning data D101. The learning unit 109 learns (constructs) the address structure model 510 based on supervised learning using the learning data D101 described below. The learning data D101 represents learning data that the learning unit 109 applies to learning the address structure model 520, and includes input data D102 and teacher data D103. The input data D102 corresponds to unstructured address information applied as input to the address structure model 510. The teacher data D103 corresponds to structured address information that indicates the correct answer resulting from structuring the unstructured address information represented as the input data D102.

[0029] The configuration and processing of the learning unit 109 will now be described in more detail. The learning unit 109 includes an address structure model 510 to be learned, and a determination unit 105. The address structure model 510 serves as a structuring processing unit that performs structuring on input data D102 as described with reference to FIG. 2 and outputs the resulting structured address information as output data D104. The determination unit 105 compares the output data D104 (structured address information) output by the address structure model 510 with the training data D103 (structured address information) associated with the input data D102 from which the output data D104 was generated, and feeds the result back to the address structure model 510. Based on this feedback from the determination unit 105, parameters such as processing weights that the address structure model 510 applies to the input are updated, thereby learning the address structure model 510.

[0030] The type of learning model applied as the address structure model 510 is not particularly limited as long as it is a model that is learned based on supervised learning. Furthermore, parameters to be updated based on feedback from the determination unit 105 may be changed as appropriate depending on the type of learning model applied as the address structure model 510. Since an existing learning model can be applied to the address structure model 510, a detailed description thereof will be omitted. The address structure model 510 learned by the learning unit 109 is then applied to the structuring processing unit 209 of the information processing device 200. Furthermore, when learning the address structure model 510, unstructured address information (input data D102) corresponding to the input of the address structure model 510 corresponds to an example of third address information, and structured address information (output data D104) corresponding to the output corresponds to an example of fourth address information.

[0031] Next, an example of the functional configuration and processing of the information processing device 200 will be described, focusing on the portion related to the operation of the address structure model 510—in other words, the portion that accepts input of unstructured address information, performs structuring processing on the unstructured address information, and outputs the resulting structured address information. The information processing device 200 includes a structuring processing unit 209. The structuring processing unit 209 accepts address information to be structurized (unstructured address information) as input data D201, performs structuring on the address information as described with reference to FIG. 2 , and outputs the resulting structured address information as output data D202. Specifically, the actual processing related to structuring address information is realized by the address structure model 510 trained by the learning device 100. The structuring processing unit 209 inputs input data D201 (unstructured address information) to the address structure model 510, and outputs structured address information obtained by structuring the address information indicated by the input data D201 as output data D202. The structuring processing unit 209 then outputs the output data D202 to a predetermined output destination (for example, the terminal device 300 shown in FIG. 1 ). Note that when the address structure model 510 is operated, the unstructured address information (input data D201) that corresponds to the input of the address structure model 510 corresponds to an example of first address information, and the structured address information (output data D202) that corresponds to the output corresponds to an example of second address information.

[0032] By applying the above configuration and processing, even when address information to be managed is input as unstructured address information such as plain text, the address information can be divided into components of a predetermined management unit and structured. As a result, even in a situation where address information input from a user is accepted in order to manage address information in compliance with the ISO 20022 standard, the user can enter the address information using a simpler input method without having to be concerned about the address information format specified in the standard.

[0033] <Modification> Next, a modification of the information processing system according to the present embodiment will be described. In the above-described embodiment, an example of a technology for implementing a function of structuring address information received as input as unstructured address information so that components of the address information are classified into predetermined management units by using a learning model (address structure model) trained based on machine learning was described. In this modification, an example of an aspect in which input and output are changed according to the application based on the technology of the above-described embodiment will be described together with a method for implementing the same.

[0034] (Variant 1) As variant 1 of the information processing system according to this embodiment, an example of the configuration and processing for structuring address information (unstructured address information) received as input to conform to the address information format of a country that serves as the standard for managing address information will be described.

[0035] For example, Figure 6 shows an example of an application scenario of the information processing system according to this modification. The example shown in Figure 6 schematically illustrates a situation in which address information for both the remitter and the remittance recipient in a so-called remittance system is structured based on the ISO 20022 standard. More specifically, in the example shown in Figure 6, Japanese address information is specified as unstructured address information for the remitter's address, and Chinese address information is specified as unstructured address information for the remittance recipient's address. The address structure model 520 according to this embodiment accepts each of these address information as input, separates each address information into components according to the format of each country, and then structures and outputs the components.

[0036] Next, referring to Fig. 7, an example of the functional configuration and processing of the information processing system according to this modification will be described assuming the usage scenario illustrated in Fig. 6, focusing particularly on the parts related to learning the address structure model 520 and the parts related to the operation of the address structure model 520. The learning device 120 and the information processing device 220 shown in Fig. 7 correspond to the learning device 100 and the information processing device 200 in the above-described embodiment.

[0037] First, an example of the functional configuration and processing of the learning device 120 will be described, focusing particularly on the parts related to learning the address structure model 520. In this modification, the address structure model 520 receives unstructured address information in a country-specific format and a specification of the format to be used when structuring the address information as input, structures the unstructured address information in the specified format, and outputs structured address information as a result. The learning device 120 learns (constructs) this address structure model 520 based on supervised learning.

[0038] The learning device 120 includes a learning unit 129 and learning data D121. The learning unit 129 learns (constructs) the address structure model 520 based on supervised learning using the learning data D121 described below. The learning data D121 indicates learning data that the learning device 120 applies to learning the address structure model 520, and includes first input data D122, second input data D123, and teacher data D124. The first input data D122 corresponds to unstructured address information in a country-specific format among the input data to the address structure model 520. The second input data D123 corresponds to the specification of a format to be applied when structuring address information among the input data to the address structure model 520. The teacher data D124 corresponds to structured address information indicating the correct answer resulting from structuring the unstructured address information indicated as the first input data D122 based on the format specified as the second input data D123. In this modification, training data D124 is associated with a combination of the first input data D122 and the second input data D123. That is, even if the content of the unstructured address information applied as the first input data D122 is the same, if the information indicating the format applied as the second input data is different, different training data D124 is associated depending on the format.

[0039] The configuration and processing of the learning unit 129 will now be described in more detail. The learning unit 129 includes an address structure model 520 to be learned and a determination unit 126. The address structure model 520 serves as a structuring processing unit that structures the first input data D122 based on a format specified as the second input data D123 and outputs the resulting structured address information as output data D125. The determination unit 126 compares the output data D125 (structured address information) output by the address structure model 520 with the training data D124 (structured address information) associated with the first input data D122 and the second input data D123 from which the output data D125 was generated, and feeds the results back to the address structure model 520. Based on the feedback from the determination unit 126, parameters such as processing weights applied by the address structure model 520 to the input are updated, thereby learning the address structure model 520. As in the above-described embodiment, the type of learning model applied as the address structure model 520 is not particularly limited as long as it is a model that is learned based on supervised learning.

[0040] Next, an example of the functional configuration and processing of the information processing device 220 will be described, focusing on the portion related to the operation of the address structure model 520—in other words, the portion that accepts input of unstructured address information and a specified format, processes the unstructured address information according to the specified format, and outputs the resulting structured address information. The information processing device 200 includes a structuring processing unit 229. The structuring processing unit 229 accepts as input first input data D221 indicating address information to be structurized (unstructured address information by format) and second input data D222 indicating a specified format to be applied when structuring the address information. The structuring processing unit 209 then structures the address information indicated by the first input data D221 based on the format specified as the second input data D222, and outputs the resulting structured address information as output data D223. Specifically, the actual processing related to structuring address information is realized by the address structure model 520 trained by the learning device 120. The structuring processing unit 229 inputs first input data D221 (unstructured address information by format) and second input data D222 (specified format) to the address structure model 520, and outputs, as output data D223, structured address information obtained by structuring the address information indicated by the first input data D221 according to the format indicated by the second input data D222. The structuring processing unit 229 then outputs the output data D223 to a predetermined output destination (for example, the terminal device 300 shown in FIG. 1 ).

[0041] By applying the above configuration and processing, as in the above-described embodiment, even when address information to be managed is input as unstructured address information such as plain text, it is possible to divide the address information into components of a predetermined management unit and structure the address information. Furthermore, in this modification, by specifying information indicating a format to be applied when structuring the address information (for example, specifying a country), it is possible to structure the input address information according to the specified format.

[0042] (Variation 2) As Variation 2 of the information processing system according to this embodiment, an example of the configuration and processing will be described in which the function of structuring address information (unstructured address information) received as input to conform to the address information format of a country that is the standard for managing address information is realized in a manner different from that of Variation 1. Specifically, in this variation, an example will be described in which unstructured address information is structured into structured address information using the address structure model described in the above embodiment with reference to FIG. 2, and then the structured address information is converted into structured address information conforming to the format of the target country using another learning model.

[0043] Fig. 8 is a diagram showing an example of the functional configuration of an information processing device according to this modification. The information processing device 240 shown in Fig. 8 corresponds to the information processing device 220 in the above-mentioned modification 1. Furthermore, the first input data D241 and the second input data D243 shown in Fig. 8 correspond to the first input data D221 and the second input data D222 in the above-mentioned modification 1. The information processing device 240 according to this modification includes a structuring processing unit 249. The structuring processing unit 249 plays a role similar to that of the structuring processing unit 229 in the information processing device 220 according to the above-mentioned modification 1.

[0044] As shown in FIG. 8 , the structuring processing unit 249 includes an address structure model 530 that structures address information and an address conversion model 535 that converts the format of the address information structured by the address structure model 530. A learning model substantially similar to that of the address structure model 510 described with reference to FIG. 2 in the above-described embodiment is applied to the address structure model 530. Under these assumptions, the address structure model 530 receives first input data D241 (unstructured address information) as input, structures the first input data D241 as described with reference to FIG. 2, and outputs the resulting structured address information as first output data D242. The address conversion model 535 receives first output data D242 (structured address information) and second input data D243 (format) as input, converts the format of the structured address information indicated by the first output data D242 into the format specified as the second input data D243, and outputs the resulting structured address information as second output data D244. The address conversion model 535 may be learned (constructed) based on supervised learning in which the structured address information before format conversion and the format to be converted are used as input data, and the structured address information indicating the result of converting the format of the structured address information is used as training data. In the example shown in Fig. 8, the address structure model 530 corresponds to an example of the first model, and the address conversion model 535 corresponds to an example of the second model.

[0045] By applying the above configuration and processing, similar to the above-mentioned variant example 1, by specifying information indicating the format to be applied when structuring address information (for example, specifying the country), it becomes possible to structure the input address information according to that format.

[0046] (Variation 3) Variation 3 of the information processing system according to the present embodiment will be described below. In this variation, an example will be described in which code information capable of uniquely identifying a target location, such as a postal code, is accepted as input as unstructured address information to be input to an address structure model, and the address structure model is caused to perform the above-described structuring of the address information indicated by the code information, and the resulting structured address information is output.

[0047] For example, FIG. 9 shows an example of an application scenario of the information processing system according to this modification. The address structure model 540 shown in FIG. 9 accepts, as input data, a postal code corresponding to unstructured address information and information indicating the country where the postal code is managed, and performs the above-described structuring on the address information corresponding to the location indicated by the postal code in the country, thereby outputting the resulting structured address information. In the example shown in FIG. 9 , the address structure model 540 is trained (constructed) so that, when outputting structured address information, address information structured according to the U.S. format is output. The address structure model 540 is trained (constructed) based on supervised learning, as in the above-described embodiment and other modifications. Specifically, while the above-described embodiment and other modifications use address information such as plain text as the unstructured address information serving as input data, this modification simply uses code information such as a postal code instead of the address information. Furthermore, if the address structure model 540 is constructed to be able to accept code information for multiple countries as input, information on the country to which the code information belongs can be applied in addition to the code information as input data. Furthermore, the format of the structured address information to be applied as training data may be similarly adjusted to match the format of the structured address information to be output by the address structure model 540 .

[0048] This modification can also be combined with the above-described modification 1 or modification 2. In that case, the address structure model can be learned (constructed) by replacing the unstructured address information applied as input data to the address structure model in modification 1 or modification 2 with code information such as a postal code and information on the country to which the code information belongs, as in this modification.

[0049] By applying the above configuration and processing, for example, in a situation where the target address information can be uniquely identified by code information such as a postal code, the user can simply input the code information to have the information processing system output the corresponding structured address information.

[0050] <Conclusion> As described above, the information processing device according to this embodiment accepts input information including information corresponding to either one of the following types: first address information, which is defined so that the location of the target is subdivided in stages by predetermined management units, or code information capable of uniquely identifying the location indicated by the first address information. The information processing device then inputs the input information to a model trained based on machine learning, causing the model to output second address information, which is structured so that the components constituting the first address information are separated by the predetermined management units. The model is also trained (constructed) by a learning device according to this embodiment. Specifically, the learning device trains the model based on machine learning using training data including information corresponding to either one of the following types: third address information, which is defined so that the location of the target is subdivided in stages by the management units, or code information corresponding to the third address information, and fourth address information, which is structured so that the components constituting the third address information are separated by the management units. By applying the above configuration, it is possible to more preferably structure the address information accepted as input. This means that, for example, even in a situation where address information is to be input from a user in order to manage address information in compliance with the ISO 20022 standard, the user can enter the address information using a simpler input method without having to be aware of the address information format specified in the standard.

[0051] The above-described embodiment is merely an example and does not necessarily limit the configuration or processing of the present invention. Various modifications and alterations may be made without departing from the technical spirit of the present invention. For example, while the address structure model has been described as being constructed based on supervised learning, any method may be used. Furthermore, the application of the above-described embodiment is not particularly limited. In other words, the technology disclosed herein can be applied to any system that uses address information to provide products or services, provided that structuring is applied to the management of the address information. Examples of such systems include not only the remittance system described above, but also systems that provide delivery services such as courier services. The present invention also includes a program that implements the functions of the above-described embodiment and a computer-readable recording medium storing the program.

[0052] The following configuration also falls within the technical scope of the present disclosure: (1) An information processing device comprising: a receiving means for receiving input of input information including information corresponding to one of the following types: first address information defined so that the location of an object is subdivided in stages by predetermined management units, and code information capable of uniquely identifying the location indicated by the first address information; and a structuring means for inputting the input information to a model trained based on machine learning, thereby causing the model to output second address information structured so that components of the first address information are separated for each of the predetermined management units, wherein the model is trained based on machine learning using training data including information corresponding to the type of third address information defined so that the location of an object is subdivided in stages by the management units, and the unstructured third address information and the code information capable of uniquely identifying the location indicated by the third address information, and fourth address information obtained by structuring the third address information. (2) The information processing device according to (1), wherein the receiving means receives input of information relating to a format of address information as the input information, and the structuring means inputs the input information including information relating to the format to the model, thereby causing the model to output the second address information structured according to the format. (3) The information processing device according to (2), wherein the format is a format defined for each country, and the receiving means receives, as the information relating to the format, designation of a country corresponding to the format to be applied to the second address information.(4) The information processing device according to (1), wherein the accepting means accepts input of information regarding a format of address information as the input information, the structuring means inputs the first address information included in the input information or code information corresponding to the first address information to a first model that is the model, thereby causing the first model to output the second address information, and inputs the second address information output by the first model and information regarding the format included in the input information to a second model constructed based on machine learning, thereby causing the second model to output the second address information structured according to the format, and the second model is trained based on machine learning using training data including the fourth address information, information regarding the format, and the fourth address information structured according to the format. (5) The information processing device according to (1), wherein the accepting means accepts input of a postal code corresponding to the first address information as the code information included in the input information, and the model receives the input of the postal code corresponding to the first address information and outputs the second address information structured based on the address information corresponding to the postal code. (6) The information processing device according to any one of (1) to (5), wherein the management unit is a unit based on an administrative division of a country that serves as a reference. (7) The information processing device further includes a learning unit that learns a model based on machine learning using learning data including information corresponding to one of types of third address information defined so that a location of an object is subdivided in stages by predetermined management units and code information corresponding to the third address information, and fourth address information structured for the third address information so that components of the third address information are divided for each of the predetermined management units, and the model learned by the learning unit receives input information including information corresponding to the type of first address information defined so that a location of an object is subdivided in stages by the management units and code information that can uniquely identify a location indicated by the first address information, and outputs second address information structured for the first address information.(8) The information processing device according to (7), wherein the learning means inputs information corresponding to any one of the third address information and code information corresponding to the third address information to the model, and learns the model based on supervised learning using the fourth address information as training data. (9) The information processing device according to (7), wherein the learning data includes information related to a format of address information, and the fourth address information is address information obtained by structuring the third address information according to the format, and the learning means inputs information corresponding to any one of the third address information and code information corresponding to the third address information, and information related to the format to the model, and learns the model based on supervised learning using the fourth address information as training data. (10) An information processing method executed by an information processing device, comprising: a receiving step of receiving input of input information including information corresponding to one of the following types: first address information defined so that the location of an object is subdivided in stages by predetermined management units; and code information that can uniquely identify the location indicated by the first address information; and a structuring step of inputting the input information to a model trained based on machine learning, thereby causing the model to output second address information structured so that components of the first address information are separated for each of the predetermined management units, wherein the model is trained based on machine learning using training data including information corresponding to the type of third address information defined so that the location of the object is subdivided in stages by the management units, and the unstructured third address information and the code information that can uniquely identify the location indicated by the third address information, and fourth address information structured from the third address information.(11) An information processing method executed by an information processing device, comprising: a learning step of learning a model based on machine learning using learning data including information corresponding to one of the following types: third address information defined so that the location of the target is subdivided in stages by predetermined management units; and code information corresponding to the third address information; and fourth address information structured for the third address information so that the components of the third address information are divided for each of the predetermined management units; wherein the model learned in the learning step receives input information including information corresponding to the type of first address information defined so that the location of the target is subdivided in stages by the management units; and code information that can uniquely identify the location indicated by the first address information, and outputs second address information structured for the first address information. (12) A program that causes a computer to execute the following steps: a receiving step of receiving input information including information corresponding to one of the following types: first address information defined so that the location of the target is subdivided in stages by predetermined management units, and code information that can uniquely identify the location indicated by the first address information; and a structuring step of inputting the input information to a model trained based on machine learning, thereby causing the model to output second address information that is structured so that components of the first address information are separated for each of the predetermined management units; wherein the model is trained based on machine learning using training data that includes information corresponding to the type of third address information defined so that the location of the target is subdivided in stages by the management units, and unstructured third address information, and code information that can uniquely identify the location indicated by the third address information, and fourth address information in which the third address information is structured.(13) A program that causes a computer to execute a learning step of learning a model based on machine learning using learning data including information corresponding to one of the following types: third address information defined so that the location of the target is subdivided in stages by predetermined management units; and code information corresponding to the third address information; and fourth address information structured for the third address information so that the components of the third address information are divided for each of the predetermined management units; and the model learned in the learning step receives input information including information corresponding to the type of first address information defined so that the location of the target is subdivided in stages by the management units; and code information that can uniquely identify the location indicated by the first address information, and outputs second address information structured for the first address information.

[0053] 1 Information processing system 100, 120 Learning device 105, 126 Determination unit 109, 129 Learning unit 200, 220, 240 Information processing device 209, 229, 249 Structuring processing unit 300 Terminal device 500, 510, 520, 530 Address structure model

Claims

1. An information processing device comprising: a receiving means for receiving input of input information including first address information defined so that the location of an object is subdivided in stages by predetermined management units and information regarding the format of the address information; and a structuring means for inputting the input information into a model trained based on machine learning, and causing the model to output second address information structured according to the format of the address information so that the components of the first address information are separated for each of the predetermined management units, wherein the model is trained based on machine learning using training data including third address information defined so that the location of an object is subdivided in stages by the management units and which is not structured, and fourth address information obtained by structuring the third address information according to the format of the address information.

2. The information processing device according to claim 1, wherein the format is a format defined for each country, and the receiving means receives, as information regarding the format of the address information, a designation of a country corresponding to the format of the address information to be applied to the second address information.

3. The information processing device according to claim 1 or 2, wherein the management unit is a unit based on the administrative division of a country that is a reference.

4. An information processing device having a learning means for learning a model based on machine learning using learning data including third address information specified so that the location of the target is subdivided in stages by predetermined management units, information indicating the format of the address information, and fourth address information structured in accordance with the format of the address information so that the components of the third address information are separated for each of the predetermined management units, wherein the model trained by the learning means receives input information including first address information specified so that the location of the target is subdivided in stages by the management units and information indicating the format of the address information, and outputs second address information structured in accordance with the format of the address information for the first address information.

5. An information processing device as described in claim 4, wherein the learning means learns the model based on supervised learning using the third address information and information indicating the format of the address information as input to the model and fourth address information structured according to the format of the address information as training data.

6. An information processing method executed by an information processing device, comprising: a receiving step of receiving input information including first address information defined so that the location of a target is subdivided in stages by predetermined management units and information related to the format of the address information; and a structuring step of inputting the input information into a model trained based on machine learning, causing the model to output second address information structured according to the format of the address information so that the components of the first address information are separated for each of the predetermined management units, wherein the model is trained based on machine learning using training data including third address information defined so that the location of a target is subdivided in stages by the management units and which is not structured, and fourth address information obtained by structuring the third address information according to the format of the address information.

7. An information processing method executed by an information processing device, comprising: a learning step of learning a model based on machine learning using learning data including: third address information defined so that the location of an object is subdivided in stages by predetermined management units; information indicating the format of the address information; and fourth address information structured in accordance with the format of the address information so that the components of the third address information are separated for each of the predetermined management units; wherein the model learned in the learning step receives input information including first address information defined so that the location of an object is subdivided in stages by the management units and information indicating the format of the address information, and outputs second address information structured in accordance with the format of the address information for the first address information.

8. A program that causes a computer to execute the following steps: a receiving step of receiving input information including first address information defined so that the location of an object is subdivided in stages by predetermined management units, and information regarding the format of the address information; and a structuring step of inputting the input information into a model trained based on machine learning, thereby causing the model to output second address information that has been structured according to the format of the address information so that the components of each of the predetermined management units that make up the first address information are separated; wherein the model is trained based on machine learning using training data that includes third address information defined so that the location of an object is subdivided in stages by the management units, and that has not been structured, and fourth address information in which the third address information has been structured according to the format of the address information.

9. A program that causes a computer to execute a learning step of learning a model based on machine learning using learning data including third address information defined so that the location of an object is subdivided in stages by predetermined management units, information indicating the format of the address information, and fourth address information structured in accordance with the format of the address information so that the components of the third address information are separated for each of the predetermined management units, and the model learned in the learning step receives input information including first address information defined so that the location of an object is subdivided in stages by the management units and information indicating the format of the address information, and outputs second address information structured in accordance with the format of the address information for the first address information.

10. A recording medium having a program recorded thereon, which causes a computer to execute the following steps: a receiving step of receiving input information including first address information defined so that the location of the target is subdivided in stages by predetermined management units and information regarding the format of the address information; and a structuring step of inputting the input information into a model trained based on machine learning, thereby causing the model to output second address information structured according to the format of the address information so that the components of the first address information are separated for each of the predetermined management units; wherein the model is trained based on machine learning using training data including third address information defined so that the location of the target is subdivided in stages by the management units and which is not structured, and fourth address information obtained by structuring the third address information according to the format of the address information.

11. A recording medium having a program recorded thereon, which causes a computer to execute a learning step of learning a model based on machine learning using learning data including third address information specified so that the location of an object is subdivided in stages by predetermined management units, information indicating the format of the address information, and fourth address information structured in accordance with the format of the address information so that the components of the third address information are separated for each of the predetermined management units, and the model learned in the learning step receives input information including first address information specified so that the location of an object is subdivided in stages by the management units and information indicating the format of the address information, and outputs second address information structured in accordance with the format of the address information for the first address information.

12. An information processing device comprising: a receiving means for receiving input of input information including code information that can uniquely identify a location indicated by first address information defined so that the location of the target is subdivided in stages by predetermined management units; and a structuring means for inputting the input information to a model trained based on machine learning, and causing the model to output second address information that is structured so that the components of the first address information are separated for each of the predetermined management units, wherein the model is trained based on machine learning using training data that includes code information that can uniquely identify a location indicated by unstructured third address information defined so that the location of the target is subdivided in stages by the management units, and fourth address information that has been structured based on the third address information.

13. The information processing device according to claim 12, wherein the receiving means receives input of information relating to the format of address information as the input information, and the structuring means inputs the first address information contained in the input information or code information corresponding to the first address information into a first model that is the model, thereby causing the first model to output the second address information, and inputs the second address information output by the first model and information relating to the format contained in the input information into a second model constructed based on machine learning, thereby causing the second model to output the second address information structured according to the format, and the second model is trained based on machine learning using training data including the fourth address information, information relating to the format, and the fourth address information structured according to the format.

14. The information processing device described in claim 13, wherein the receiving means receives input of a postal code corresponding to the first address information as the code information included in the input information, and the model receives input of a postal code corresponding to the first address information and outputs the second address information structured for the address information corresponding to the postal code.

15. An information processing device having a learning means for learning a model based on machine learning using learning data including code information corresponding to third address information defined so that the location of an object is subdivided in stages by predetermined management units, and fourth address information structured for the third address information so that the components of the third address information are divided into the predetermined management units, wherein the model trained by the learning means receives input information including code information that can uniquely identify the location indicated by first address information defined so that the location of an object is subdivided in stages by the management units, and outputs second address information structured for the first address information.

16. The information processing device described in claim 15, wherein the learning data includes information regarding the format of address information, the fourth address information is address information that has been structured to match the format of the third address information, and the learning means inputs information corresponding to any one of the types of the third address information and code information corresponding to the third address information, and information regarding the format, into the model, and learns the model based on supervised learning using the fourth address information as training data.

17. An information processing method executed by an information processing device, comprising: a receiving means for receiving input of input information including code information capable of uniquely identifying a location indicated by first address information defined so that the location of the target is subdivided in stages by predetermined management units; and a structuring means for inputting the input information to a model trained based on machine learning, and causing the model to output second address information structured so that components of the first address information are separated for each of the predetermined management units, wherein the model is trained based on machine learning using training data including code information capable of uniquely identifying a location indicated by unstructured third address information defined so that the location of the target is subdivided in stages by the management units, and fourth address information structured from the third address information.

18. An information processing method executed by an information processing device, comprising: a learning step of learning a model based on machine learning using learning data including code information corresponding to third address information defined so that the location of an object is subdivided in stages by predetermined management units; and fourth address information structured for the third address information so that the components constituting the third address information are separated for each of the predetermined management units; wherein the model trained in the learning step receives input information including code information that can uniquely identify the location indicated by first address information defined so that the location of an object is subdivided in stages by the management units, and outputs second address information structured for the first address information.

19. A program that causes a computer to execute the following: accepting means for accepting input of input information including code information that can uniquely identify a location indicated by first address information that is defined so that the location of the target is subdivided in stages by predetermined management units; and structuring means for inputting the input information to a model trained based on machine learning, and causing the model to output second address information that is structured so that components of the first address information are separated for each of the predetermined management units; wherein the model is trained based on machine learning using training data that includes code information that can uniquely identify a location indicated by unstructured third address information that is defined so that the location of the target is subdivided in stages by the management units, and fourth address information that is structured based on the third address information.

20. A program that causes a computer to execute a learning step of learning a model based on machine learning using learning data including code information corresponding to third address information defined so that the location of an object is subdivided in stages by predetermined management units, and fourth address information structured for the third address information so that the components of the third address information are divided into the predetermined management units, and the model learned in the learning step receives input information including code information that can uniquely identify the location indicated by first address information defined so that the location of an object is subdivided in stages by the management units, and outputs second address information structured for the first address information.

21. A recording medium having a program recorded thereon, which causes a computer to execute: a receiving means for receiving input of input information including code information capable of uniquely identifying a location indicated by first address information, which is defined so that the location of the target is subdivided in stages by predetermined management units; and a structuring means for inputting the input information to a model trained based on machine learning, and causing the model to output second address information structured so that components of the first address information are separated for each of the predetermined management units; wherein the model is trained based on machine learning using training data including code information capable of uniquely identifying a location indicated by unstructured third address information, which is defined so that the location of the target is subdivided in stages by the management units, and fourth address information obtained by structuring the third address information.

22. A recording medium having a program recorded thereon, which causes a computer to execute a learning step of learning a model based on machine learning using learning data including code information corresponding to third address information defined so that the location of an object is subdivided in stages by predetermined management units, and fourth address information structured for the third address information so that the components of the third address information are divided into the predetermined management units, and the model learned in the learning step receives input information including code information that can uniquely identify the location indicated by first address information defined so that the location of an object is subdivided in stages by the management units, and outputs second address information structured for the first address information.

Images