Information processing device and information processing method

Abstract

An information processing device (10) comprises: a draft text reception unit (11) that receives a draft text of recommendation information to be distributed to a user; a text extraction unit (13) that extracts, on the basis of attribute information of the user and the draft text, a plurality of texts of the recommendation information, which suit the preference of the user, from a storage unit having the distributed recommendation information stored therein; and an instruction information generation unit (15) that presents, to a generative AI, the draft text received by the draft text reception unit (11) and the texts which suit the preference of the user and which have been extracted by the text extraction unit (13), and outputs instruction information for causing the generative AI to revise the draft text into text having the content that suits the preference of the user.

Description

Information processing device and information processing method

[0001] One aspect of this disclosure relates to an information processing device and an information processing method.

[0002] Systems that deliver recommended content based on user preference information are known. For example, Patent Document 1 discloses a recommendation device that delivers advertisements tailored to a user's hobbies and preferences based on preference information stored in a database.

[0003] Japanese Patent Publication No. 2004-94384

[0004] However, even if the above-mentioned conventional recommendation system delivers advertisements (recommended information) tailored to the user's hobbies and preferences, users do not necessarily view or read the delivered advertisements. Therefore, there was a risk that the promotional effect would not be fully realized.

[0005] One aspect of this disclosure aims to provide an information processing device and an information processing method that can enhance the promotional effect when delivering recommendation information to users.

[0006] An information processing device relating to one aspect of this disclosure includes: a text draft receiving unit that receives text drafts of recommendation information to be distributed to a user; a text extraction unit that extracts multiple recommendation information texts that match the user's preferences from a storage unit that stores information on previously distributed recommendation information, based on the user's attribute information and the text drafts; and an instruction information generation unit that presents the text drafts received by the text draft receiving unit and the texts that match the user's preferences extracted by the text extraction unit to a generating AI, and outputs instruction information to cause the generating AI to revise the text drafts into text that matches the user's preferences.

[0007] According to one aspect of this disclosure, it is possible to enhance the promotional effect when delivering recommendation information to users.

[0008] Figure 1 is a diagram showing an example of the functional configuration of a system including an information processing device according to one embodiment. Figure 2 is a diagram showing an example of a distribution information database according to one embodiment. Figure 3(A) is a diagram illustrating a method for estimating the predicted open rate. Figure 3(B) is a diagram showing the predicted open rate estimated by the open rate estimation unit. Figure 4 is a flowchart showing an example of processing by the information processing device. Figure 5(A) is an example of a prompt generated by the instruction information generation unit. Figure 5(B) is an example of a corrected text output by providing the prompt shown in Figure 5(A) to the generation AI model. Figure 6 is a diagram showing an example of the hardware configuration of a computer used in the information processing device according to one embodiment.

[0009] The embodiments described herein will be described in detail below with reference to the drawings. In the description of the drawings, the same elements will be denoted by the same reference numerals, and redundant descriptions will be omitted. Furthermore, the embodiments described herein in the following description are specific examples of the disclosure and are not limited to these embodiments unless otherwise stated to limit the disclosure.

[0010] Figure 1 is a diagram showing an example of the configuration of a system 1 including an information processing device 10 according to one embodiment. As shown in Figure 1, the information processing device 10 is configured to communicate via a network with a plurality of user terminals 5 used by each of a plurality of users. The network is any communication network, such as a mobile communication network or the Internet.

[0011] The information processing device 10 performs processing to cause the generation AI model 30 to generate recommendation information for a predetermined product or service for the user. More specifically, the information processing device 10 generates and outputs prompts (instruction information) to instruct the generation AI model 30 to generate recommendation information that matches the user's preferences. The recommendation information may include text, images (still images or videos), audio, or two or more of these types of information. In this embodiment, the information processing device 10 generates and outputs prompts to output text that can enhance the promotional effect for the user to whom the information is distributed.

[0012] User terminal 5 is a terminal used by a user who is the target of receiving recommendation information for a predetermined product or service. User terminal 5 can take any form, such as a personal computer, smartphone, tablet, feature phone, server device, or game device. Although only one user terminal 5 is shown in Figure 1, the information processing device 10 may be configured to send recommendation information for each of multiple users to each user's user terminal 5. In this case, the system 1 shown in Figure 1 may include multiple user terminals 5 used by each of the multiple users.

[0013] The information processing device 10 includes a generation AI model 30, a storage unit 20 that stores an attribute information database 21, a psychological attribute database 25, and a distribution information database 27, and a psychological attribute estimation unit 23, an open rate estimation unit 29, a draft document reception unit 11, a document extraction unit 13, an instruction information generation unit 15, and a distribution unit 17 as functional blocks.

[0014] The generation AI model 30 is a model that, in response to the input of a prompt (instruction information) output by the information processing device 10, generates content (in this embodiment, text included in recommendation information) according to any or a combination of the instruction, context, question, and output format indicated by the prompt, and returns the content as response information.

[0015] A prompt is information indicating an instruction or question input to the generating AI model 30 in an interactive system such as a command-line interface (CLI). A prompt can include various types of input information. Input information may include, for example, data files with filenames containing predetermined extensions, such as text data, image data, application-related data, audio data, video data, and still image data. Application-related data refers to data such as document data, table data, and graph data that can be processed by a default application program.

[0016] The generative AI model 30 may consist of, for example, a Large Language Model (LLM) and a user interface (UI) for interacting with the user. The generative AI model 30 may also be an interactive AI model capable of text chat or voice chat with the user. Examples of such generative AI models 30 include ChatGPT, GPT-3.5, GPT-4V, and PaLM2. However, the generative AI model 30 may be an AI model other than the large language model described above.

[0017] The storage unit 20 stores an attribute information database 21 (hereinafter referred to as "attribute information DB 21"), a psychological attribute database 25 (hereinafter referred to as "psychological attribute DB 25"), and a distribution information database 27 (hereinafter referred to as "distribution information DB 27"). The storage unit 20 that stores the above databases may consist of one hardware device (storage) such as an HDD (Hard Disk Drive) or SSD (Solid State Drive), or two or more hardware devices. Alternatively, the storage unit 20 that stores the above databases may consist of hardware devices built into multiple network-connected server devices.

[0018] The attribute information DB21 consists of information about the user's attributes. Attribute information may include, for example, the user's age, gender, occupation, personality, location information, user subscription status and usage status of the product or service being promoted, and information about the installation status and usage status of applications. Attribute information may be information entered (registered) by the user themselves, information obtained from various sensors on the user terminal 5, or information estimated based on the user's behavioral history (for example, the usage history of the user terminal 5).

[0019] The Psychological Attributes DB25 consists of information about each user's psychological attributes estimated from the user's attribute information (hereinafter simply referred to as "estimated psychological attributes"). Examples of estimated psychological attributes include BigFive, HLC (Health Locus of Control), time discount rates (mean time discount rate, magnitude effect, hyperbolic discount, sign effect), behavioral inhibition system / behavioral approach system scales (so-called BIS (behavioral inhibition system) / BAS (behavioral approach system)), empathy scales, interpersonal reactivity index (IRI (Interpersonal Reactivity Index)), self-awareness scales, etc.

[0020] In the following explanation, the five traits of the Big Five theory (extraversion, conscientiousness, agreeableness, openness, and emotional instability) are described as estimates of psychological attributes. Estimates of psychological attributes are shown, for example, in Figure 2, etc. [O ｎ , C ｎ , E ｎ , A ｎ , N ｎ It can be expressed as follows: n represents the username. In other words, the estimated psychological attributes of a user differ from user to user. Each of the five trait factors of the Big Five theory is represented by a score, such as the z-value or standard score of a standard normal distribution. The further the score of each factor is from the mean, the greater the degree to which the person fits the corresponding factor.

[0021] The distribution information DB 27 stores information about recommendation information that has already been distributed. The information about distributed recommendation information stored in the distribution information DB 27 includes, for example, the content of the recommendation information, the text included in the recommendation information, information on whether or not various recommendation information has been distributed to the user terminal 5, destination information and distribution time information for the distribution of the recommendation information, information on whether or not the distributed recommendation information has been opened by the user terminal 5, information on the time of opening and the user terminal that opened it, and information on the predicted open rate y of the distributed recommendation information. The predicted open rate y is a value estimated by the open rate estimation unit 29, which will be described later.

[0022] The psychological attribute estimation unit 23 estimates an estimated value of the psychological attribute based on the attribute information of the user. For example, the psychological attribute estimation unit 23 can estimate the estimated value of the psychological attribute for each user using existing techniques from, for example, application usage log data obtained from the user terminal 5 stored in the attribute information DB 21. The psychological attribute estimation unit 23 stores, for example, the scores of each factor for each user in the psychological attribute DB 25 as [O ｎ , C ｎ , E ｎ , A ｎ , N ｎ .

[0023] The opening rate estimation unit 29 estimates, for each user, a predicted opening rate y, which is the probability of opening the distributed recommendation information, based on the text included in the recommendation information distributed by email and the estimated value of the psychological attribute. For example, among the five dimensions of the above Big Five (openness, conscientiousness, extraversion, agreeableness, neuroticism), for conscientiousness, the lower the conscientiousness, the more irresponsible the personality, and thus there is a tendency that it is difficult to open the received message in the first place. However, even if the user has an irresponsible personality with low conscientiousness, if the user opens the message, it can be considered that the user is likely to open the message with interest. Thus, from the estimated value of the psychological attribute, it is possible to infer the psychological factors of the user at the time of message opening (such as whether the user opened the message with interest), and it becomes possible to consider the psychological factors of the user.

[0024] Furthermore, the recipients Z of the recommendation information are determined based on user attributes stored in the attribute information DB21. Examples of recipients Z include the type of OS (Android® or iOS, etc.) and the mobile carrier. As shown in Figure 3(A), for emails that are selected for delivery (Z=1), it is possible to record whether they were opened (Y1=1) or not (Y1=0), but for emails that are not selected for delivery (Z=0), it is not possible to record whether they were opened (Y1=1) or not (Y1=0). For this reason, it has not been possible to consider the recipient bias in the past.

[0025] More specifically, users who open emails may be more likely to have a habit of checking their email frequently or to be interested in specific content. Conversely, users who do not open emails may be more likely to have a habit of not checking their email often or to be uninterested in the content of the email. Therefore, analyzing open rates using only data from users who opened emails (Z=1) can lead to the following distribution bias: The results will only be based on the characteristics of users who open emails, and are therefore less likely to accurately reflect the entire user base (both those who opened and those who did not). In particular, the open rate for users who do open emails may be overestimated, while the actual open rate for the entire user base may be lower.

[0026] In this regard, the open rate estimation unit 29 of this embodiment calculates the predicted open rate y based on estimated psychological attributes, taking into account target bias (such as determining target users based on the OS, mobile carrier, etc.) by using, for example, the Heckit model. As a result, the open rate estimation unit 29 can calculate the predicted open rate y after removing target bias, and can calculate the predicted open rate y for each user and for each sentence of the delivered recommendation information, including users who are not the target audience, as shown in Figure 2.

[0027] For example, we will explain how to correct for target bias using the Heckit model. The estimation model that determines whether or not to open an email is as follows. Note that the estimation model below is a model that has been pre-built by machine learning from the estimation model represented by the following equation (1). Z = Xγ + u ... (1) Here, Z is a variable that indicates whether or not to open an email, X is an explanatory variable (for example, user attributes), γ is a coefficient, and u is an error term.

[0028] The predictive model for the open rate of emails opened by users is as follows. Note that the predictive model below is a model that has been pre-built using machine learning based on the predictive model represented by equation (2) below. Y = Psychological attribute * β + ρλ(Xγ) + e ... (2) Here, Y is the open rate, psychological attribute is the explanatory variable, β is the vector of coefficients corresponding to each psychological attribute, ρλ(Xγ) is the term for correcting selection bias, and e is the error term. * means the dot product of the psychological attribute vector and the coefficient vector. For example, ρ is an image of the correlation coefficient between u and e.

[0029] Then, using the coefficient β estimated by this Heckit model, the predicted open rate y for users who did not receive the email (Z=0) and users who did receive it (Z=1) is calculated. This makes it possible to construct a delivery information DB27 that takes into account selection bias (target bias) for users who did not open the email. More specifically, in this embodiment, after estimating the psychological attributes of the users, selection bias by the target audience is taken into account, and by learning from training data that sets the estimated values ​​of psychological attributes and the words in which the user responded (opened) the email, a delivery information DB27 that can take into account the inner thoughts of the users can be constructed.

[0030] The draft text receiving unit 11 receives draft text for recommendation information to be distributed to users. The draft text for recommendation information distributed to users includes not only text about the content of the product or service, but also text of a message from the provider of the product or service. This message is used, for example, as the title when the recommendation information is distributed via email or other means.

[0031] The document proposal reception unit 11 receives, for example, a document proposal of recommendation information to be distributed to the user from an input device such as a keyboard. The document proposal reception unit 11 receives, for example, a document proposal such as "Join XX service!" from the distributor of the recommendation information (the user of the information processing device 10) via a keyboard or the like.

[0032] The document extraction unit 13 extracts a plurality of documents of recommendation information suitable for the user's preference from the distribution information DB 27 in the storage unit 20 in which the documents of the distributed recommendation information are stored, based on the user's attribute information and the document proposal received by the document proposal reception unit 11. The document extraction unit 13 extracts a document suitable for the user's preference based on the document proposal and the estimated value of the psychological attribute for each user stored in the psychological attribute DB 25 in the storage unit 20.

[0033] More specifically, the document extraction unit 13 calculates the similarity between the document proposal vectorized by a method such as embedding based on the estimated value of the psychological attribute and the document vectorized by the same method, and extracts a predetermined number of documents (for example, 100 documents) in the order of the calculated similarity from high to low. For the determination of the similarity, for example, the determination based on the cosine similarity is used.

[0034] The document extraction unit 13 further extracts a plurality of documents of the recommendation information from the distribution information DB 27 in consideration of the predicted opening rate y included in the distribution information DB 27. Specifically, the document extraction unit 13 extracts the documents of the recommendation information weighted based on the predicted opening rate y. The document extraction unit 13 extracts a larger number of documents of the recommendation information as the predicted opening rate y is higher. For example, the document extraction unit 13 extracts the documents of the recommendation information such that the number of documents of the recommendation information with a predicted opening rate y of 1.0 is twice the number of documents of the recommendation information with a predicted opening rate y of 0.5. As a result, for example, when the document extraction unit 13 extracts the top 100 documents of the recommendation information based on the cosine similarity (or when extracting the documents of the recommendation information with a cosine similarity above a predetermined threshold), more documents with a high predicted opening rate y are extracted, and it becomes difficult to extract documents with a low predicted opening rate y.

[0035] The instruction information generation unit 15 presents the draft text received by the draft text receiving unit and the text that matches the user's preferences extracted by the text extraction unit 13 to the generation AI model 30, and outputs instruction information to the generation AI model 30 to revise the draft text into text that matches the user's preferences. The instruction information generation unit 15 outputs instruction information to revise the text so that the more similar the texts are, the more it reflects that preference. The instruction information generation unit 15 may provide instruction information that offers only one revision suggestion, or it may provide instruction information that offers two or more revision suggestions. If the instruction information generation unit 15 offers two or more revision suggestions, it may present the instruction information to display them in the recommended order.

[0036] The distribution unit 17 automatically distributes the revised draft text output from the generation AI model 30, based on the prompt generated by the instruction information generation unit 15, to the recipient as recommendation information text. If there are two or more revised drafts output from the generation AI model 30, the distributor of the recommendation information (the user of the information processing device 10) may select and distribute the appropriate revised drafts.

[0037] The functional blocks described above in the information processing device 10 are intended to function within the information processing device 10, but are not limited to this. For example, some of the functional blocks of the information processing device 10 may function in a computer device different from the information processing device 10, within a computer device connected to the information processing device 10 via a network, while appropriately sending and receiving information with the information processing device 10. Furthermore, the functional blocks in the information processing device 10 do not have to be configured as shown in Figure 1, some of the functional blocks of the information processing device 10 may be omitted, multiple functional blocks may be integrated into one functional block, or one functional block may be decomposed into multiple functional blocks.

[0038] Next, mainly referring to FIGS. 4 and 5, an example of the processing executed by the above-described information processing apparatus 10, that is, an example of a method for generating a prompt for instructing the generation AI model 30 to generate recommendation information according to the user's preferences will be described. FIG. 4 is a flowchart showing an example of processing by the information processing apparatus. FIG. 5(A) is an example of a prompt generated by the instruction information generation unit 15. FIG. 5(B) is an example of a corrected sentence output when the prompt shown in FIG. 5(A) is input (given) to the generation AI model.

[0039] As shown in FIG. 4, the sentence draft reception unit 11 receives a sentence draft of the recommendation information to be distributed to the user (step S1). The sentence draft reception unit 11 receives, for example, a sentence "Join the XX service" as shown in area A2 of FIG. 5(A). The sentence draft reception unit 11 is input via a keyboard or the like by the distributor of the recommendation information or the like.

[0040] Next, the sentence extraction unit 13 obtains an estimated value ([O ｎ , C ｎ , E ｎ , A ｎ , N ｎ ) of the psychological attribute of the distribution target user from the psychological attribute DB 25 (step S2). The sentence extraction unit 13 extracts the sentences of the recommendation information included in the distribution information DB 27 based on the estimated value ([O ｎ , C ｎ , E ｎ , A ｎ , N ｎ ) of the psychological attribute of the distribution target user and the sentence draft received in step S1 (step S3).

[0041] As shown in FIG. 5(B), the sentence extraction unit 13 extracts, for example, 100 sentences including sentences such as "It is very beneficial to join the XX service!" and "You will suffer losses if you do not join the XX service."

[0042] Next, the instruction information generation unit 15 presents the draft sentence "Let's subscribe to the XX service" received by the draft sentence reception unit 11, and 100 sentences extracted by the sentence extraction unit 13, including "You'll save money if you subscribe to the XX service!" and "You'll lose out if you don't subscribe to the XX service." to the generating AI model 30. More specifically, the instruction information generation unit 15 generates a prompt as shown in Figure 5(A) (step S4). The prompt generated by the instruction information generation unit 15, for example, has "Please output three sentences corrected according to the example" written in the initial area A0, 100 sentences extracted by the sentence extraction unit 13, including "You'll save money if you subscribe to the XX service!" and "You'll lose out if you don't subscribe to the XX service." written in area A1 below area A0, and the draft sentence "Let's subscribe to the XX service" obtained by the sentence extraction unit 13 is written in area A2 below area A1.

[0043] Next, the instruction information generation unit 15 inputs the prompt generated in step S4 to the generation AI model 30 (step S5). As a result, the generation AI model 30 outputs a revised version of the draft sentence "Join the XX service" acquired by the text extraction unit 13. The distribution unit 17 receives the revised version "Join the XX service now and you'll get lots of benefits." output from the generation AI model 30 (step S6). The distribution unit 17 then displays a screen in which the revised version "Join the XX service now and you'll get lots of benefits." is displayed in area A3 below area A2, as shown in Figure 5(B). Users of the information processing device 10 can understand the revised version provided by the generation AI model 30 from a screen such as the one shown in Figure 5(B) displayed on a monitor.

[0044] The distribution unit 17 distributes recommendation information, along with the suggested revisions provided by the generation AI model 30, to users stored in the attribute information DB 21 (step S7). If the generation AI provides multiple suggested revisions, the distributor of the recommendation information selects a revision and distributes the recommendation information to users stored in the attribute information DB 21. The text of the recommendation information distributed by the distribution unit 17 is stored in the distribution information DB 27 as part of the distribution history.

[0045] Next, the effects and benefits of the information processing device 10 according to the above embodiment will be described.

[0046] According to the information processing device 10 and information processing method of the above embodiment, first, multiple sentences that match the user's preferences are extracted from the sentences of recommendation information that have been distributed in the past, and based on the extracted sentences, prompts are generated for the generating AI to revise the draft sentences of the recommendation information prepared by, for example, the service or product provider. In the information processing device 10 and information processing method of the above embodiment, the generating AI can be made to revise the draft sentences that match the user's preferences using the prompts generated in this way. This makes it possible to enhance the promotional effect when distributing recommendation information to users.

[0047] The information processing device 10 further includes a psychological attribute estimation unit 23 that estimates psychological attribute values ​​based on the user's attribute information, and the text extraction unit 13 may extract text that suits the user's preferences based on the draft text and the estimated psychological attribute values. This makes it possible to enhance the promotional effect more effectively according to the user's personality (taking into account the user's inner self).

[0048] Furthermore, the text extraction unit 13 may calculate the similarity between the vectorized text drafts based on estimated psychological attributes and the texts vectorized based on estimated psychological attributes, and extract a predetermined number of texts in order of the calculated similarity. In this case as well, texts that are close to the user's preferences can be extracted according to the user's personality.

[0049] Furthermore, the instruction information generation unit 15 may output instruction information to modify sentences that are more similar to each other, so that the similarity is reflected in the sentences. In this case, instruction information will be output to the generating AI so that the sentences become closer to the user's preferences.

[0050] Furthermore, the information processing device 10 includes an open rate estimation unit 29 that estimates a predicted open rate y for each user, which is the probability that the user will open the recommendation information delivered by email, based on the text contained in the recommendation information delivered by email and the user's psychological attributes. The text extraction unit 13 may extract the text of the recommendation information based on an extraction rule that prioritizes the extraction of text from recommendation information with a higher predicted open rate y. In this case, instruction information will be output that allows the user to revise the draft text based on the text that is most likely to be opened by the user when the recommendation information is delivered.

[0051] Furthermore, the recommendation information may include messages directed at the user. The text draft receiving unit may receive draft messages, and the text extraction unit may extract multiple message texts that match the user's preferences. In this case, if the recommendation information includes messages directed at the user, those messages can be made into text that matches the user's preferences.

[0052] The apparatus and method of this disclosure may have the following configurations.

[0053] [1] An information processing device comprising: a text draft receiving unit that receives text drafts of recommendation information to be distributed to a user; a text extraction unit that extracts multiple texts of recommendation information that match the user's preferences from a storage unit that stores information on previously distributed recommendation information, based on the user's attribute information and the text drafts; and an instruction information generation unit that presents the text drafts received by the text draft receiving unit and the texts that match the user's preferences extracted by the text extraction unit to a generating AI, and outputs instruction information to cause the generating AI to revise the text drafts into text that matches the user's preferences.

[0054] [2] The information processing apparatus according to [1], further comprising a psychological attribute estimation unit that estimates psychological attribute values ​​based on the user's attribute information, wherein the text extraction unit extracts text that is in line with the user's preferences based on the draft text and the estimated psychological attribute values.

[0055] [3] The information processing apparatus according to [2] above, wherein the text extraction unit calculates the similarity between the draft text vectorized based on the estimated values ​​of the psychological attributes and the text vectorized based on the estimated values ​​of the psychological attributes, and extracts a predetermined number of the texts in descending order of the calculated similarity.

[0056] [4] The information processing apparatus according to [3] above, wherein the instruction information generation unit outputs instruction information for modifying sentences that are more similar to each other so that the tendency is reflected in the sentence.

[0057] [5] The information processing apparatus according to any one of [1] to [4] above, wherein the storage unit records whether or not the delivered recommendation information has been opened at the recipient, and the text extraction unit extracts text that matches the user's preferences from the recommendation information that has been recorded as having been opened.

[0058] [6] An information processing device according to any one of [2] to [4] above, further comprising an open rate estimation unit that estimates a predicted open rate for each user, which is the probability of opening the delivered recommendation information, based on the text contained in the recommendation information delivered by email and the estimated values ​​of the psychological attributes, and the text extraction unit extracts the text of the recommendation information based on an extraction rule that prioritizes the extraction of text of the recommendation information that has a higher predicted open rate.

[0059] [7] The information processing apparatus according to any one of [1] to [6] above, wherein the recommendation information includes a message directed to the user or a title of the recommendation information, the draft text receiving unit receives draft text of the message or the title, and the text extraction unit extracts multiple texts of the message or the title that match the user's preferences.

[0060] [8] Information processing method comprising: a text draft reception step of receiving text drafts of recommendation information to be distributed to a user; a text extraction step of extracting multiple texts of recommendation information that match the user's preferences from a storage unit that stores distributed recommendation information, based on the user's attribute information and the text drafts; and a text draft generation step of presenting the text drafts received in the text draft reception step and the texts that match the user's preferences extracted in the text extraction step to a generating AI, and outputting instruction information to the generating AI to modify the text drafts into text that matches the user's preferences.

[0061] Although one embodiment has been described above, one aspect of this disclosure is not limited to the above embodiment. Various modifications are possible without departing from the spirit of this disclosure. (Modifications)

[0062] In the above embodiment, an example was given in which the distribution information DB27 stored in the storage unit 20 is associated with and stored the predicted open rate y for the distributed recommendation information. However, instead of the predicted open rate y, it may be recorded whether or not the distributed recommendation information was opened at the recipient. Even in this case, the text extraction unit 13 will prioritize extracting draft texts of recommendation information that has already been opened, thereby increasing the possibility of extracting text that is close to the user's preferences.

[0063] In the above embodiments and modifications, the distribution information DB27 stores the open status of the distributed recommendation information (predicted open rate y or whether or not it was opened), and the text extraction unit 13 is described as extracting text based on these open statuses, but it is not limited to this. For example, the text extraction unit 13 may extract text based on information that the distributed product or service has been purchased or the service has been subscribed to.

[0064] In the above embodiment and modified example, the open rate estimation unit 29 calculates the predicted open rate y based on the estimated values ​​of each user's psychological attributes stored in the psychological attribute DB 25. However, the predicted open rate y may also be calculated based on factors other than the estimated values ​​of psychological attributes. For example, instead of or in addition to the estimated values ​​of psychological attributes, the predicted open rate y may be calculated based on one or more pieces of user attribute information stored in the attribute information DB 21 described above. Furthermore, the open rate estimation unit 29 may also calculate the predicted open rate y based on various factors stored in databases other than those provided in the above embodiment and modified example, such as information about the user's app usage history, location information, contract information, and healthcare information.

[0065] For example, the distribution information DB27 may store "information related to service subscriptions," such as information indicating whether a user has subscribed to a service, and information regarding the subscribed user's name, service name, and subscription procedure time. In this case as well, the service subscription rate can be calculated in the same way as the method for calculating the predicted open rate y described above. The text extraction unit 13 may extract text based on such a service subscription rate.

[0066] The information processing device 10 of the above embodiment only needs to include at least a document draft receiving unit 11, a document extraction unit 13, and an instruction information generation unit 15, and the other components mentioned above may be configured by an external server device.

[0067] The block diagrams used in the description of the above embodiments show functional units. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using a single device that is physically or logically coupled, or it may be realized using two or more physically or logically separated devices that are directly or indirectly connected (for example, using wired, wireless, etc.). A functional block may also be realized by combining software with the single device or the multiple devices described above.

[0068] Functions include, but are not limited to, judgment, decision, judgment, calculation, calculation, processing, derivation, investigation, exploration, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, assumption, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating (mapping), and assigning. For example, a functional block (configuration part) that enables transmission is called a transmitting unit or transmitter. In all cases, as mentioned above, the method of implementation is not particularly limited.

[0069] For example, the information processing device 10 in one embodiment of the present disclosure may function as a computer that performs prompt generation processing as described in the present disclosure. Figure 6 is a diagram showing an example of the hardware configuration of the information processing device 10 according to one embodiment of the present disclosure. The above-described information processing device 10 may be physically configured as a computer device including a processor 1001, memory 1002, storage 1003, communication device 1004, input device 1005, output device 1006, bus 1007, etc.

[0070] In the following explanation, the term "device" can be replaced with "circuit," "device," "unit," etc. The hardware configuration of the information processing device 10 may include one or more of the devices shown in the figure, or it may be configured without some of the devices.

[0071] Each function in the information processing device 10 is realized by loading predetermined software (programs) onto hardware such as the processor 1001 and memory 1002, which allows the processor 1001 to perform calculations, control communication by the communication device 1004, and control at least one of data reading and writing in the memory 1002 and storage 1003.

[0072] The processor 1001 controls the entire computer, for example, by running the operating system. The processor 1001 may be composed of a central processing unit (CPU) that includes interfaces with peripheral devices, control devices, arithmetic units, registers, etc. For example, the psychological attribute estimation unit 23, open rate estimation unit 29, document draft reception unit 11, document extraction unit 13, instruction information generation unit 15, and distribution unit 17 described above may be implemented including the processor 1001.

[0073] Furthermore, the processor 1001 reads programs (program code), software modules, data, etc., from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes accordingly. The program used is one that causes the computer to execute at least a part of the operations described in the above embodiment. For example, the psychological attribute estimation unit 23, the open rate estimation unit 29, the draft document reception unit 11, the document extraction unit 13, the instruction information generation unit 15, and the distribution unit 17 may be implemented by a control program stored in the memory 1002 and running on the processor 1001, and other functional blocks may be implemented similarly. The above-described various processes have been explained as being executed by one processor 1001, but they may be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may also be transmitted from a network via a telecommunications line.

[0074] The program executed by the processor 1001 may include: a text draft reception step that receives text drafts of recommendation information to be distributed to the user; a text extraction step that extracts multiple recommendation information texts that match the user's preferences from a storage unit that stores distributed recommendation information, based on the user's attribute information and the text drafts; and a text draft generation step that presents the text drafts received in the text draft reception step and the texts that match the user's preferences extracted in the text extraction step to a generating AI, and outputs instruction information to the generating AI to modify the text drafts into texts that match the user's preferences.

[0075] The memory 1002 is a computer-readable recording medium and may consist of at least one of the following: ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. The memory 1002 may also be called a register, cache, main memory, etc. The memory 1002 can store executable programs (program code), software modules, etc., for carrying out the prompt generation method according to the embodiment of this disclosure.

[0076] The storage 1003 is a computer-readable recording medium and may consist of at least one of the following: an optical disc such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital multipurpose disk, a Blu-ray® disk), a smart card, flash memory (e.g., a card, a stick, a key drive), a floppy® disk, a magnetic strip, etc. The storage 1003 may also be called an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, a server, or other suitable medium that includes at least one of the memory 1002 and the storage 1003. For example, the above-mentioned storage unit 20 may be implemented including the storage 1003.

[0077] The communication device 1004 is hardware (transceiver / receiver device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as a network device, network controller, network card, communication module, etc. The communication device 1004 may be configured to include a high-frequency switch, duplexer, filter, frequency synthesizer, etc., in order to implement at least one of frequency division duplex (FDD) and time division duplex (TDD). For example, the psychological attribute estimation unit 23, open rate estimation unit 29, draft document reception unit 11, document extraction unit 13, instruction information generation unit 15, and distribution unit 17 described above may be implemented including the communication device 1004.

[0078] The input device 1005 is an input device that accepts input from an external source (e.g., a keyboard, mouse, microphone, switch, button, sensor, etc.). The output device 1006 is an output device that outputs to an external source (e.g., a display, speaker, LED lamp, etc.). The input device 1005 and the output device 1006 may be configured as an integrated unit (e.g., a touch panel).

[0079] Furthermore, each device, such as the processor 1001 and memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or different buses may be configured for each device.

[0080] Furthermore, the information processing device 10 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array), and some or all of each functional block may be realized by such hardware. For example, the processor 1001 may be implemented using at least one of these hardware components.

[0081] The notification of information is not limited to the manner / embodiments described herein and may be carried out by other means.

[0082] Each aspect / embodiment described in this disclosure may be applied to at least one of the following systems: LTE (Long Term Evolution), LTE-A (LTE-Advanced), SUPER 3G, IMT-Advanced, 4G (4th generation mobile communication system), 5G (5th generation mobile communication system), FRA (Future Radio Access), NR (new Radio), W-CDMA®, GSM®, CDMA2000, UMB (Ultra Mobile Broadband), IEEE 802.11 (Wi-Fi®), IEEE 802.16 (WiMAX®), IEEE 802.20, UWB (Ultra-WideBand), Bluetooth®, and other appropriate systems, as well as next-generation systems extended based thereon. Furthermore, multiple systems may be applied in combination (for example, a combination of at least one of LTE and LTE-A with 5G).

[0083] The processing procedures, sequences, flowcharts, etc., of each aspect / embodiment described in this disclosure may be reordered, provided they are consistent. For example, the methods described in this disclosure present various step elements using exemplary order and are not limited to the specific order presented.

[0084] Input and output information may be stored in a specific location (e.g., memory) or managed using a management table. Input and output information may be overwritten, updated, or appended to. Output information may be deleted. Input information may be transmitted to other devices.

[0085] The determination may be made by a value represented by one bit (0 or 1), by a boolean value (true or false), or by a numerical comparison (for example, a comparison with a predetermined value).

[0086] Each aspect / embodiment described in this disclosure may be used individually, in combination, or switched between as needed during implementation. Furthermore, notification of specific information (e.g., notification that "X is") is not limited to explicit notification, but may also be implicit (e.g., by not providing such notification).

[0087] Although the present disclosure has been described in detail above, it will be clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the intent and scope of the present disclosure as defined by the claims. Therefore, the descriptions in the present disclosure are illustrative and not intended to be restrictive in any way.

[0088] Software should be broadly interpreted to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, execution threads, procedures, functions, etc., whether they are called software, firmware, middleware, microcode, hardware description languages, or by any other name.

[0089] Furthermore, software, instructions, information, etc., may be transmitted and received via a transmission medium. For example, if software is transmitted from a website, server, or other remote source using at least one of wired technology (such as coaxial cable, fiber optic cable, twisted pair, or digital subscriber line (DSL)) and wireless technology (such as infrared or microwave), then at least one of these wired and wireless technologies is included in the definition of a transmission medium.

[0090] The information, signals, etc. described in this disclosure may be represented using any of the various different technologies. For example, the data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.

[0091] In addition, terms used in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meaning.

[0092] The terms “system” and “network” as used in this disclosure are interchangeable.

[0093] The terms “determining” and “decision” as used in this disclosure may encompass a wide variety of actions. “Determining” may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, searching, inquiry (e.g., searching in tables, databases, or other data structures), ascertaining, etc. “Determining” may also include receiving (e.g., receiving information), transmitting (e.g., sending information), input, output, accessing (e.g., accessing data in memory), etc. “Determining” may also include resolving, selecting, choosing, establishing, comparing, etc. In other words, "judgment" and "decision" can include considering that some action has been "judged" or "decided." Also, "judgment (decision)" can be reinterpreted as "assuming," "expecting," "considering," etc.

[0094] The terms “connected,” “coupled,” and any variations thereof mean any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are “connected” or “coupled” with each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, “connection” may be reinterpreted as “access.” As used in this disclosure, two elements may be considered to be “connected” or “coupled” with each other using at least one of one or more wires, cables, and printed electrical connections, and, in some non-limiting and non-exclusive examples, electromagnetic energy having wavelengths in the radio frequency domain, microwave domain, and optical (both visible and invisible) domain.

[0095] In this disclosure, the phrase "based on" does not mean "based solely on" unless otherwise specified. In other words, the phrase "based on" means both "based solely on" and "based at least on."

[0096] Where the terms “include,” “including,” and variations thereof are used in this disclosure, these terms are intended to be inclusive, as is the term “comprising.” Furthermore, the term “or” as used in this disclosure is not intended to mean exclusive OR.

[0097] In this disclosure, if articles are added by translation, such as a, an, and the in English, this disclosure may include the fact that the noun following these articles is plural.

[0098] In this disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "combine" may be interpreted similarly to "different."

[0099] 5...User terminal, 10...Information processing device, 11...Document draft receiving unit, 13...Document extraction unit, 15...Instruction information generation unit, 17...Distribution unit, 20...Storage unit, 21...Attribute information database, 23...Psychological attribute estimation unit, 25...Psychological attribute database, 27...Distribution information database, 29...Open rate estimation unit, 30...Generating AI model.

Claims

1. An information processing device comprising: a text draft receiving unit that receives text drafts of recommendation information to be distributed to a user; a text extraction unit that extracts multiple texts of recommendation information that match the user's preferences from a storage unit that stores information on previously distributed recommendation information, based on the user's attribute information and the text drafts; and an instruction information generation unit that presents the text drafts received by the text draft receiving unit and the texts that match the user's preferences extracted by the text extraction unit to a generating AI, and outputs instruction information to cause the generating AI to revise the text drafts into text that matches the user's preferences.

2. The information processing apparatus according to claim 1, further comprising a psychological attribute estimation unit that estimates psychological attribute values ​​based on the user's attribute information, wherein the text extraction unit extracts text that is in line with the user's preferences based on the draft text and the estimated psychological attribute values.

3. The information processing apparatus according to claim 2, wherein the text extraction unit calculates the similarity between the draft text vectorized based on the estimated values ​​of the psychological attributes and the text vectorized based on the estimated values ​​of the psychological attributes, and extracts a predetermined number of the texts in descending order of the calculated similarity.

4. The information processing apparatus according to claim 3, wherein the instruction information generation unit outputs instruction information for modifying sentences that are more similar to each other, so that the similarity of the sentences is reflected in the sentence.

5. The information processing apparatus according to claim 1, wherein the storage unit records whether or not the delivered recommendation information has been opened at the recipient, and the text extraction unit extracts text that matches the user's preferences from the recommendation information that has been recorded as having been opened.

6. The information processing apparatus according to claim 2, further comprising an open rate estimation unit that estimates a predicted open rate for each user, which is the probability of opening the delivered recommendation information, based on the text contained in the recommendation information delivered by email and the estimated values ​​of the psychological attributes, and the text extraction unit extracts the text of the recommendation information based on an extraction rule that prioritizes the extraction of text of the recommendation information that has a higher predicted open rate.

7. The information processing apparatus according to claim 1, wherein the recommendation information includes a message directed to the user, the text draft receiving unit receives a draft of the message, and the text extraction unit extracts a plurality of message texts that match the user's preferences.

8. An information processing method comprising: a text draft reception step for receiving draft text for recommendation information to be distributed to a user; a text extraction step for extracting multiple recommendation information texts that match the user's preferences from a storage unit that stores previously distributed recommendation information, based on the user's attribute information and the text draft; and a text draft generation step for presenting the text draft received in the text draft reception step and the texts that match the user's preferences extracted in the text extraction step to a generating AI, and outputting instruction information to the generating AI to modify the text draft into text that matches the user's preferences.

Images