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

Abstract

The process of generating general-purpose information based on text. [Solution] The information processing device according to the present invention comprises an acquisition unit, a vector conversion unit, and an extraction unit. The acquisition unit acquires character vectors converted from characters. The vector conversion unit converts the character vectors into user vectors. The extraction unit extracts target users corresponding to the user vectors.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, and an information processing program.

Background Art

[0002] Conventionally, various techniques for receiving character input and generating information have been disclosed. For example, Patent Document 1 discloses a method for determining recommended products based on language elements extracted by natural language analysis or the like.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the prior art, the types of information to be generated are determined. For example, in the case of Patent Document 1, the information generated based on characters is information on recommended products. However, in order to flexibly respond to user requests, it is preferable to generate more general-purpose information.

[0005] The present application has been made in view of the above, and an object thereof is to provide an information processing apparatus, an information processing method, and an information processing program that generate general-purpose information based on characters.

Means for Solving the Problems

[0006] [[ID=4〗]] The information processing apparatus according to the present application includes an acquisition unit that acquires a character vector converted from characters, a vector conversion unit that converts the character vector into a user vector, and an extraction unit that extracts a target user corresponding to the user vector.

Effects of the Invention

[0007] According to one embodiment, general-purpose information can be generated based on characters. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 is an overview diagram illustrating the information processing according to the embodiment. [Figure 2] Figure 2 shows an example of how the results of extracting target users according to the embodiment are used. [Figure 3] Figure 3 is a block diagram showing an example configuration of an information processing apparatus according to the embodiment. [Figure 4] Figure 4 is a flowchart showing an example of information processing according to the present invention. [Figure 5] Figure 5 is a hardware configuration diagram showing an example of a computer that implements the functions of the information processing device according to the embodiment. [Modes for carrying out the invention]

[0009] The following describes in detail, with reference to the drawings, embodiments for implementing the information processing device, information processing method, and information processing program according to the present application (hereinafter referred to as "embodiments"). Note that these embodiments do not limit the information processing device, information processing method, and information processing program according to the present application. Furthermore, the same parts are denoted by the same reference numerals in the following embodiments, and redundant descriptions are omitted.

[0010] [Embodiment] [1. Information Processing] First, an example of information processing according to the embodiment will be explained using Figure 1. Figure 1 is an overview diagram of the information processing according to the embodiment.

[0011] The information processing shown in Figure 1 is performed by the information processing device 1. The information processing device 1 is a server device that generates general-purpose information based on characters. Specific examples of server devices include computers such as PCs and blade servers, mainframes, and workstations. Furthermore, the server device may be implemented through cloud computing.

[0012] As shown in Figure 1, the information processing according to the embodiment includes a model creation phase and a policy utilization phase. In the model creation phase, model M1 is trained. In the policy utilization phase, general-purpose information is generated by converting character vectors to user vectors using model M1. Specifically, target users that match any descriptive text entered by user U are extracted, and the extracted target user information can be used broadly and generally.

[0013] [1-1. Model Creation] First, various types of information used to create the model are stored in a storage device (step S11). The storage device may be a storage unit provided by the information processing device 1, or it may be a data server different from the information processing device 1. The information stored in the storage device will be described below.

[0014] For example, a storage device may record user information for multiple target users. Examples of user information include identifying information (such as a user ID) that identifies the target user. Other examples of user information include information about the target user's attributes, such as gender, age, place of residence, demographics, psychographics, geographics, and behavioral attributes. Furthermore, examples of user information may include segments or personas (profiles) to which the target user belongs in the field of marketing.

[0015] User information can be collected, for example, from the user information of web services. Although there is no particular limitation on the type of web service, as an example, it can include Internet connection, search services, SNS (Social Networking Service), e-commerce (EC: Electronic Commerce), electronic payment, online games, online banking, online trading, accommodation and ticket reservations, video and music distribution, news, maps, route search, route guidance, route information, operation information, weather forecasts, etc. Alternatively, the user information may be information generated for virtual target users by manual input operations or known learning data generation techniques.

[0016] Furthermore, the storage device stores a description text of the user group. The user group is a group consisting of a plurality of target users and is set to have some meaning.

[0017] The user group can be set, for example, by extracting target users with similar specific attributes from among a plurality of target users whose user information is recorded in the storage device. For example, a user group can be set for each of a plurality of target users included in a specific age group or a plurality of target users living in a specific region.

[0018] Another example is that the user group can be set by extracting target users who have performed a specific action. For example, a user group can be set for each of a plurality of target users who use online shopping with a frequency above a threshold or a plurality of target users who have traveled to a specific region.

[0019] Furthermore, explanatory texts are set for each user group. The explanatory text of a user group is a description of the characteristics of the user group in natural language. For example, for a user group set for target users living in a specific region, a character string "people living in ~" is set. Also, for example, for a user group set for target users who use online shopping with a frequency above a threshold, a character string "people who often use online shopping" is set.

[0020] The explanatory text of a user group can be set manually by a person who has referred to the user information of the target users included in the user group, for example. Alternatively, the explanatory text of the user group may be set first, and the user group may be set so as to match the explanatory text. For example, the explanatory text "people who often use online shopping" is set first, and the user group and its explanatory text are set by extracting target users who use online shopping with a frequency above a threshold.

[0021] One target user may be included in multiple user groups. For example, one target user may be included in the user group of "people who often use online shopping" and also in the user group of "people who enjoy making travel and vacation plans".

[0022] Also, multiple user groups may be set hierarchically so that the size of the set changes. For example, a user group of a parent category "people who like shopping" is set, and user groups of child categories such as "people who often use online shopping" and "people who like high-end brands and tend to purchase high-end products" may be further set. At this time, the parent category is the union of the child categories.

[0023] Furthermore, user groups may be defined based on combinations of categories. For example, there may be a user group for "people who like shopping," a user group for "people who enjoy planning trips and vacations," and a user group for "people who like shopping AND enjoy planning trips and vacations."

[0024] Furthermore, multiple similar categories may be merged as appropriate. For example, two user groups could be created for two categories, "people who like sports" and "people who like fitness," or one user group could be created for a single category, "people who like either sports or fitness."

[0025] Next, the information processing device 1 performs a conversion process from natural language to vectors (step S12) and a conversion process from user information to vectors (step S13). The order in which steps S12 and S13 are performed is arbitrary, and they may be performed in parallel.

[0026] First, let's explain the process in step S12. In step S12, the information processing device 1 performs a natural language to vector conversion process on the user group's description to generate a character vector A1. Character vector A1 is an example of a character vector for training. For example, the information processing device 1 can use an Embedding model or the like to convert the user group's description into a character vector A1 with an arbitrary number of dimensions.

[0027] The conversion of natural language to vectors is a well-known technique used in various natural language-based generation technologies. For example, well-known natural language processing models such as GPT (Generative Pre-trained Transformer) can perform various generation processes upon receiving natural language input, but in all cases, the first step is conversion to vectors. The information processing device 1 may use such a well-known natural language processing model to perform the conversion of natural language to vectors. Note that GPT is just one example, and other large-scale language models (LLMs) may also be used.

[0028] Next, the process in step S13 will be explained. In step S13, the information processing device 1 performs a conversion process from user information to a vector for each user group to generate a user vector B1. User vector B1 is an example of a user vector for learning.

[0029] Specifically, the information processing device 1 first converts user information into a user vector for each target user included in the user group. For example, the user vectors are calculated such that the cosine similarity is high among target users with similar tendencies and low among target users with different tendencies. Furthermore, the information processing device 1 generates a user vector B1 by averaging the user vectors calculated for each target user included in the user group. There are no particular restrictions on the number of dimensions of the user vector B1, but it may be different from the number of dimensions of the character vector A1.

[0030] Then, the information processing device 1 creates model M1 by performing machine learning using the training data (step S14). The training data consists of pairs of character vectors A1 and user vectors B1, and is created for each set user group. In other words, the information processing device 1 can create a large amount of training data by setting a large number of user groups.

[0031] The specific configuration of Model M1 is not particularly limited and can be implemented using known AI (Artificial Intelligence) technologies such as neural networks. For example, the information processing device 1 creates Model M1, which converts a character vector to a user vector, by adjusting the parameters of the neural network so that when a character vector A1 is input, information similar to the paired user vector B1 is output. Model M1 is an example of a conversion model that is functionally configured to convert an input character vector into a user vector.

[0032] [1-2. Measure usage] The created model M1 is used in response to user U's input. First, user U inputs an arbitrary description through terminal 3 (step S21). Terminal 3 can be any device that can accept the input of an explanation. For example, terminal 3 can be a smartphone or a PC. The arbitrary description is entered in natural language.

[0033] As an example, let's consider a case where User U is a marketing representative for a company that sells car air fresheners. In this case, User U might enter a description such as, "A user who owns a car and has recently become concerned about the smell inside their car." In other words, User U enters a description of the user of their company's product. To give another example, User U might enter a description such as, "Improves the smell inside a car." In other words, User U enters a description of their company's product itself.

[0034] The information processing device 1 generates a character vector A2 by performing a natural language-to-vector conversion process on the explanatory text input by user U (step S22). For example, the information processing device 1 can perform the conversion process in step S22 using an Embedding model that converts characters to vectors, or a natural language processing model such as GPT, similar to step S12. Note that character vector A2 is an example of a character vector converted from characters. Furthermore, the information processing device 1 converts character vector A2 into a user vector B2 by inputting it into model M1 (step S23).

[0035] Furthermore, the information processing device 1 extracts target users corresponding to user vector B2 (step S24). For example, the information processing device 1 converts the user information of each target user into a user vector and compares it with user vector B2, and extracts target users whose cosine similarity is above a threshold as target users corresponding to user vector B2.

[0036] As explained above, the series of processes shown in Figure 1 allows for the extraction of target users based on any description entered by user U. The extracted target users themselves are useful information. For example, the market size can be estimated from the number of extracted target users. Furthermore, if there is a bias in the attributes of the extracted target users, this can be reflected in marketing strategies. Therefore, the information processing device 1 may notify user U of the extracted target users directly. For example, user U can check the notification about the extracted target users via the display of terminal 3.

[0037] Furthermore, the information processing device 1 may use the extracted target users as reference information to provide services to user U (step S25). For example, the information processing device 1 provides services to support user U's advertising-related work.

[0038] An example of a service that utilizes the extracted target users as reference information is explained using Figure 2. Figure 2 is a diagram illustrating an example of how the extracted target users are used according to the embodiment. Figure 2 is merely an example, and the use of the extracted target users is not limited to advertising.

[0039] For example, User U is a marketing professional who delivers advertisements online using the advertising platform shown in Figure 2. In this case, it is possible to improve advertising efficiency by using the extracted target users as reference information. As an example, as shown in Figure 2, advertisements can be delivered to the device 4 owned by the extracted target users. In other words, according to the series of processes shown in Figures 1 and 2, targeted advertising can be carried out based on any description entered by User U.

[0040] As described above, according to the information processing of this embodiment, target users are extracted based on the description entered by user U. The extracted target users are useful information in themselves, and are also general-purpose information that can be used as reference information for providing various services.

[0041] [2. Information Processing Devices] Next, an example of the configuration of the information processing device 1 according to the embodiment will be described using Figure 3. Figure 3 is a block diagram showing an example of the configuration of the information processing device 1 according to the embodiment. As shown in Figure 3, the information processing device 1 comprises a communication unit 11, a storage unit 12, and a control unit 13. The information processing device 1 may also have an input unit (for example, a keyboard or mouse) that accepts various operations from an administrator or other user of the information processing device 1, and a display unit (for example, a liquid crystal display) for displaying various information.

[0042] The communication unit 11 is implemented, for example, by a NIC (Network Interface Card). The communication unit 11 is connected to a communication network such as 4G (4th Generation) or 5G (5th Generation) by wire or wireless connection, and transmits and receives information with the various devices shown in Figure 1 via the communication network.

[0043] The memory unit 12 is implemented by, for example, semiconductor memory elements such as RAM (Random Access Memory) and flash memory, or by storage devices such as hard disks and optical discs.

[0044] The control unit 13 is a controller and is realized by executing various programs (corresponding to an example of an information processing program) stored in the memory device inside the information processing device 1 using RAM as a working area, for example, by a CPU (Central Processing Unit) or MPU (Micro Processing Unit). The control unit 13 may be realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array). The control unit 13 includes an acquisition unit 131, a model creation unit 132, a vector conversion unit 133, an extraction unit 134, and an output unit 135.

[0045] The acquisition unit 131 acquires various types of information in each step of model creation and policy utilization shown in Figure 1. For example, in step S12, the acquisition unit 131 acquires character vector A1 by performing a natural language to vector conversion process on the description of the user group. Also, for example, in step S13, the acquisition unit 131 acquires user vector B1 by performing a vector conversion process and averaging on the user information of each target user included in the user group. Also, for example, in step S21, the acquisition unit 131 receives input of a description from user U. Also, for example, in step S22, the acquisition unit 131 acquires character vector A2 by performing a natural language to vector conversion process on the description entered by user U.

[0046] Up to this point, it has been explained that the conversion process from natural language to vectors is performed by the acquisition unit 131, but this process may be performed by an external device different from the information processing device 1. In this case, for example, in step S12, the external device performs the conversion process from natural language to vectors for the user group's description, thereby generating character vector A1. Similarly, in step S22, the external device performs the conversion process from natural language to vectors for the description entered by user U, thereby generating character vector A2. The acquisition unit 131 then acquires the character vectors A1 and A2 generated by the external device via the communication unit 11 or the communication network. That is, the acquisition unit 131 may acquire the character vectors converted from characters by performing the conversion process from natural language to vectors, or it may acquire them from an external device.

[0047] Similarly, the process of converting user information to a vector may be performed by an external device different from the information processing device 1. In this case, for example, in step S13, the external device performs the conversion process to a vector and averaging for the user information of each target user included in the user group, thereby generating the user vector B1. The acquisition unit 131 then acquires the user vector B1 generated by the external device via the communication unit 11 or the communication network.

[0048] The model creation unit 132 executes step S14 shown in Figure 1 to create a model M1 that converts character vectors into user vectors. The created model M1 is stored, for example, in the storage unit 12.

[0049] The creation process for model M1 may be performed by an external device. In this case, the information processing device 1 can acquire the created model M1 via the communication unit 11 or a communication network and store it in the storage unit 12. In this case, the information processing device 1 does not need to have a model creation unit 132.

[0050] The vector conversion unit 133 executes step S23 shown in Figure 1 to convert the character vector A2 into the user vector B2. The extraction unit 134 then executes step S24 shown in Figure 1 to extract the target user corresponding to the user vector B2.

[0051] The output unit 135 performs output processing based on the target users extracted by the extraction unit 134. For example, the output unit 135 notifies user U of the target users extracted by the extraction unit 134. Alternatively, for example, the output unit 135 uses the target users extracted by the extraction unit 134 as reference information to provide services to user U.

[0052] [3. Processing Flow] Next, the processing procedure executed by the information processing device 1 according to the embodiment will be described using Figure 4. Figure 4 is a flowchart showing an example of information processing according to the embodiment.

[0053] First, the information processing device 1 obtains a model M1 that converts character vectors into user vectors (step S101). As described above, the information processing device 1 may obtain the model M1 by executing step S14 shown in Figure 1, or it may obtain a model M1 created by an external device.

[0054] Next, the information processing device 1 receives an arbitrary descriptive text in natural language from user U (step S102) and performs a conversion process from natural language to vectors (step S103). Next, the information processing device 1 inputs the character vector generated in step S103 into model M1 and converts it into a user vector (step S104), and extracts the target user corresponding to the user vector (step S105). The information processing device 1 also uses the extracted target user as reference information to provide services to user U (step S106). Although not shown in Figure 4, the information processing device 1 may, instead of or in addition to the process in step S106, notify user U of the extracted target user.

[0055] Here, the information processing device 1 may decide whether or not to provide feedback to user U regarding the extracted target user (step S107).

[0056] If feedback is provided (step S107 affirmative), the information processing device 1 performs a conversion process from user information to vectors for the extracted target users and generates user vectors (step S108). For example, the information processing device 1 performs a conversion process from user information to vectors for each target user and generates user vectors by averaging the multiple converted vectors.

[0057] Next, the information processing device 1 inputs the user vector into model M1 and converts it into a character vector (step S109). When this step is performed for the purpose of providing feedback, model M1 is configured to perform bidirectional conversion between character vectors and user vectors. For example, in step S14 of Figure 1, the information processing device 1 can create a model M1 that performs bidirectional conversion between character vectors and user vectors by adjusting the parameters of the neural network so that when character vector A1 is input, information similar to the paired user vector B1 is output, and when user vector B1 is input, information similar to the paired character vector A1 is output.

[0058] Next, the information processing device 1 performs a conversion process from vector to natural language on the character vector converted in step S109 and generates an explanatory text (step S110). The generated explanatory text corresponds to an explanatory text that describes the characteristics of the target user (user group) extracted in step S105 using natural characters.

[0059] Furthermore, the conversion process from vectors to natural language is a well-known technique used in natural language processing models. For example, in natural language processing models that perform translation or dialogue, the input natural language is converted into a vector, various processes are performed on it, and the processed vector is converted back into natural language for output. Similarly, the information processing device 1 can perform the conversion process from vectors to natural language.

[0060] Then, the information processing device 1 presents the explanatory text converted in step S110 to user U (step S111). This allows user U to confirm whether the target users they intended have been extracted.

[0061] The information processing device 1 accepts a selection from user U regarding whether or not to revise the description (step S112). If the user U chooses to revise the description (step S112 affirmative), the device proceeds back to step S102. At this time, user U can make adjustments to the description, such as adding conditions to prevent the extraction of unintended users with undesirable attributes, or changing the wording to prevent the exclusion of users being extracted, based on the description presented in step S111, and then re-enter the description.

[0062] Furthermore, when the system proceeds to step S102 again, user U may enter a new description or modify the previously entered description. In other words, when the system proceeds to step S102 again, the information processing device 1 accepts the re-entry or modification of the description.

[0063] The feedback processing described above can be carried out through dialogue between User U and Information Processing Device 1. For example, Information Processing Device 1 can ask whether the extracted target user matches the user's image, accept a request to revise the description if it does not, and then ask again whether the extracted target user matches the user's image.

[0064] On the other hand, if no correction is made (step S112 is denied), or if feedback is not required (step S107 is denied), the information processing device 1 terminates the process.

[0065] It is preferable that the description entered in step S102 is well-organized to provide a clear and concise definition, but it is acceptable for it to contain ambiguous or abstract parts. For example, the description entered in step S102 may be an uncalibrated text written freely by user U, or a fragmented text written before the target user profile to be extracted is finalized. Modern LLMs can extract features from texts containing ambiguous or abstract parts, and even if an unorganized description is entered in step S102, there is a possibility that the target user desired by user U can be extracted. Furthermore, if the target user desired by user U is not extracted, the description can be adjusted through the feedback process described above so that the desired target user is extracted.

[0066] [4. Variations] In known natural language processing models, input characters are converted into vectors, and then various processes are performed on the vectors. For example, in the case of GPT, weighting is performed by a process called attention. The various character vectors in the embodiment may or may not reflect processes such as attention.

[0067] For example, if character vectors A1 and A2 in Figure 1 are highlighted, the system will extract target users by giving more weight to important words in the description entered by user U. If character vectors A1 and A2 are not highlighted, the system will extract target users by strictly following the description entered by user U. The highlighting of highlighting may be toggled as appropriate by user U.

[0068] Furthermore, although it was explained that user U inputs explanatory text in steps S21 in Figure 1 and S102 in Figure 4, the characters that user U inputs do not have to be sentences. For example, user U may input multiple words instead of a sentence.

[0069] Furthermore, Figure 4 illustrates an example in which the series of processes in steps S107 to S111 are performed before accepting revisions to the explanatory text; however, steps S107 to S111 may be omitted. That is, the information processing device 1 may accept revisions to the explanatory text from user U based on the notification about the target user extracted in step S105 and the results of the service provided in step S106.

[0070] For example, when User U checks the notification about the target users extracted in step S105, it is conceivable that the number of extracted target users is less than User U expected. One possible reason for this is that the entered description contains unnecessary conditions, excluding some of the target users being extracted. If User U then proceeds back to step S102, they can modify the wording to ensure that the target users they want to extract are not excluded, and then re-enter the description.

[0071] Another possible scenario is that the actual market size is smaller than User U anticipated. If User U then proceeds back to step S102, they can consider strategies to expand the market size.

[0072] For example, user U first enters a description such as, "A user who owns a car and has recently become concerned about the smell inside their car." If the number of target users extracted is less than expected, one possible modification would be to change "owns a car" to "has opportunities to drive a car." This would allow for the extraction of additional people who do not own a car but have opportunities to drive one, such as family members or friends of car owners, or rental car users. Another possible modification would be to change "has opportunities to drive a car" to "has opportunities to ride a train." This would allow for an estimation of how much market size could be expanded if the company were to enter the market for train air fresheners.

[0073] [5. Effects] The information processing device 1 according to this embodiment includes an acquisition unit 131, a vector conversion unit 133, and an extraction unit 134. The acquisition unit 131 acquires character vectors converted from characters. The vector conversion unit 133 converts the character vectors into user vectors. The extraction unit 134 extracts target users corresponding to the user vectors.

[0074] Furthermore, the vector conversion unit 133 converts character vectors into user vectors by inputting them into a conversion model.

[0075] Furthermore, the information processing device 1 according to the embodiment further comprises a model creation unit 132. The model creation unit 132 creates a transformation model by performing machine learning using training data consisting of a training character vector converted from the descriptive text of the user group and a training user vector obtained by averaging user vectors converted from the user information of each of the multiple target users included in the user group.

[0076] Furthermore, the information processing device 1 according to this embodiment further includes an output unit 135. The output unit 135 notifies the user U who entered the characters of the extracted target user.

[0077] Furthermore, the output unit 135 uses the extracted target users as reference information to provide services to user U who entered text.

[0078] Furthermore, the output unit 135 performs a conversion process on the extracted user information of the target user into a vector, averages the converted vector to generate a user vector, converts the generated user vector into a character vector, and performs a conversion process on the character vector into natural language to generate an explanatory text which is then presented to the user U who entered the text. In addition, the acquisition unit 131 accepts re-entry or correction of text from user U.

[0079] Furthermore, the acquisition unit 131 obtains character vectors by performing a conversion process from natural language to vectors.

[0080] Through any or a combination of the above-described processes, the information processing device according to the present invention can generate general-purpose information based on characters.

[0081] [6. Hardware Configuration] Furthermore, the information processing device 1 according to the above embodiment is realized by a computer 1000 having a configuration such as that shown in Figure 5. Figure 5 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing device according to the embodiment. The computer 1000 is connected to an output device 1010 and an input device 1020, and has a configuration in which an arithmetic unit 1030, a primary storage device 1040, a secondary storage device 1050, an output I / F (Interface) 1060, an input I / F 1070, and a network I / F 1080 are connected by a bus 1090.

[0082] The arithmetic unit 1030 operates based on programs stored in the primary storage device 1040 and the secondary storage device 1050, as well as programs read from the input device 1020, and performs various processes. The arithmetic unit 1030 can be implemented using, for example, a CPU, MPU, ASIC, FPGA, etc.

[0083] The primary storage device 1040 is a memory device, such as RAM, that temporarily stores data used by the arithmetic unit 1030 for various calculations. The secondary storage device 1050 is a storage device in which data used by the arithmetic unit 1030 for various calculations and various databases are registered, and can be implemented using ROM (Read Only Memory), HDD (Hard Disk Drive), SSD (Solid State Drive), flash memory, etc. The secondary storage device 1050 may be internal storage or external storage. The secondary storage device 1050 may also be a removable storage medium such as USB (Universal Serial Bus) memory or SD (Secure Digital) memory card. The secondary storage device 1050 may also be cloud storage (online storage), NAS (Network Attached Storage), file server, etc.

[0084] The output I / F 1060 is an interface for transmitting information to be output to output devices 1010, such as displays, projectors, and printers, and is implemented using connectors of standards such as USB (Universal Serial Bus), DVI (Digital Visual Interface), and HDMI (High Definition Multimedia Interface). The input I / F 1070 is an interface for receiving information from various input devices 1020, such as mice, keyboards, keypads, buttons, and scanners, and is implemented using, for example, USB.

[0085] Furthermore, the output interface 1060 and input interface 1070 may be wirelessly connected to the output device 1010 and input device 1020, respectively. In other words, the output device 1010 and input device 1020 may be wireless devices.

[0086] Furthermore, the output device 1010 and the input device 1020 may be integrated as a touch panel. In this case, the output I / F 1060 and the input I / F 1070 may also be integrated as an input / output I / F.

[0087] The input device 1020 may also be a device that reads information from, for example, an optical recording medium such as a CD (Compact Disc), DVD (Digital Versatile Disc), or PD (Phase Change Rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory.

[0088] The network interface 1080 receives data from other devices via network N and sends it to the computing unit 1030, and also transmits data generated by the computing unit 1030 to other devices via network N.

[0089] The arithmetic unit 1030 controls the output device 1010 and the input device 1020 via the output interface 1060 and the input interface 1070. For example, the arithmetic unit 1030 loads a program from the input device 1020 or the secondary storage device 1050 onto the primary storage device 1040 and executes the loaded program.

[0090] For example, when computer 1000 functions as a server device 100, the arithmetic unit 1030 of computer 1000 realizes the functions of the control unit 130 by executing a program loaded onto the primary storage device 1040. Alternatively, the arithmetic unit 1030 of computer 1000 may load a program obtained from another device via the network interface 1080 onto the primary storage device 1040 and execute the loaded program. Furthermore, the arithmetic unit 1030 of computer 1000 may cooperate with other devices via the network interface 1080 and call and use program functions, data, etc., from other programs on other devices.

[0091] [7. Other] Although embodiments of the present invention have been described above, the present invention is not limited by the content of these embodiments. Furthermore, the aforementioned components include those that can be easily conceived by those skilled in the art, those that are substantially the same, and those that fall within the so-called equivalent range. Moreover, the aforementioned components can be combined as appropriate. Furthermore, various omissions, substitutions, or modifications of the components can be made without departing from the gist of the embodiments described above.

[0092] Furthermore, among the processes described in the above embodiments, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically by known methods. In addition, the processing procedures, specific names, and information including various data and parameters shown in the above document and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the information shown.

[0093] Furthermore, the components of each illustrated device are functionally conceptual and do not necessarily need to be physically configured as shown. In other words, the specific forms of distribution and integration of each device are not limited to those shown, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads and usage conditions.

[0094] For example, the information processing device described above may be implemented using multiple server computers, and depending on the function, it may be implemented by calling external platforms, etc., via APIs (Application Programming Interfaces) or network computing, allowing for flexible configuration changes.

[0095] Furthermore, the terms "section, module, unit" mentioned above can be replaced with "means" or "circuit," etc. For example, the acquisition unit can be replaced with acquisition means or acquisition circuit. [Explanation of Symbols]

[0096] 1. Information Processing Device 11 Communications Department 12 Storage section 13 Control Unit 131 Acquisition Department 132 Model Creation Department 133 Vector Transformation Unit 134 Extraction part 135 Output section 3 terminals 4 terminals

Claims

1. An acquisition unit that obtains a character vector converted from a character, A vector conversion unit that converts the aforementioned character vector into a user vector, An extraction unit that extracts target users corresponding to the user vector, An information processing device characterized by comprising:

2. The vector conversion unit converts the character vector into the user vector by inputting it into the conversion model. The information processing apparatus according to feature 1.

3. The system further includes a model creation unit that creates the conversion model by performing machine learning using training data that includes a training character vector converted from a user group description and a training user vector obtained by averaging user vectors converted from the user information of each of the multiple target users included in the user group. The information processing apparatus according to feature 2.

4. The system further includes an output unit that notifies the user who entered the characters of the target user extracted by the extraction unit. The information processing apparatus according to feature 1.

5. The system further comprises an output unit that uses the target users extracted by the extraction unit as reference information to provide services to the users who entered the characters. The information processing apparatus according to feature 1.

6. The system further includes an output unit that performs a vector conversion process on the extracted user information of the target user, averages the resulting vectors to generate a user vector, converts the generated user vector into a character vector, and then performs a natural language conversion process on the character vector to generate an explanatory text which is then presented to the user who entered the characters. The acquisition unit receives requests from the user to re-enter or modify the characters. The information processing apparatus according to feature 1.

7. The acquisition unit acquires the character vector by performing a conversion process from natural language to vector. The information processing apparatus according to feature 1.

8. A method of information processing performed by a computer, The acquisition process involves obtaining a character vector converted from a character, A vector conversion step that converts the aforementioned character vector into a user vector, An extraction step for extracting target users corresponding to the user vector, An information processing method characterized by including

9. Procedure for obtaining a character vector converted from a character, A vector conversion procedure for converting the aforementioned character vector into a user vector, Extraction procedure for extracting target users corresponding to the user vector mentioned above An information processing program characterized by causing a computer to execute it.

Images