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

Abstract

This enables the estimation of engagement that reflects multiple actions. [Solution] The information processing device according to the present invention is characterized by comprising: a generation unit that generates a user feature vector by vectorizing user features including user behavior based on multiple behavior logs; a collection unit that collects multiple action probabilities related to a product; a learning unit that trains a single model with a set of user feature vectors and multiple action probabilities; a calculation unit that inputs the user feature vector to the single model and compresses it to one dimension at the intermediate output to calculate an engagement score that reflects multiple action probabilities; and an estimation unit that performs engagement estimation using the engagement score.

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] There is disclosed a technique for predicting a conversion rate of advertisement content when the advertisement content is distributed on a distribution surface, using a prediction model that predicts the conversion rate of the advertisement content with input information including advertisement information, user information, and distribution surface information as an input (see Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the above prior art, the conversion rate is predicted using a prediction model that predicts the conversion rate of advertisement content. Although there is a method of using the conversion probability as the level of engagement, in that method, it is necessary to distribute advertisements in advance until conversion results are accumulated. Also, the effects of advertisements whose purpose is not to increase conversions, such as advertisements that appeal to recognition, cannot be measured. Therefore, there is a need for a method that utilizes multiple actions, such as actions that reflect recognition and interest, and actions that can be acquired before advertisement distribution, for estimating engagement.

[0005] The present application has been made in view of the above, and an object thereof is to enable estimation of engagement that reflects multiple actions.

Means for Solving the Problems

[0006] The information processing device according to the present invention is characterized by comprising: a generation unit that generates a user feature vector by vectorizing user features including user behavior based on multiple behavior logs; a collection unit that collects multiple action probabilities related to a product; a learning unit that trains a single model with the combination of the user feature vector and the multiple action probabilities; a calculation unit that inputs the user feature vector to the single model and compresses it to one dimension at the intermediate output to calculate an engagement score that reflects the multiple action probabilities; and an estimation unit that performs engagement estimation using the engagement score. [Effects of the Invention]

[0007] According to one embodiment, it is possible to estimate engagement that reflects multiple actions. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 is an explanatory diagram showing an overview of the information processing system according to the embodiment. [Figure 2] Figure 2 is an explanatory diagram illustrating the overview of the purchase funnel. [Figure 3] Figure 3 is an explanatory diagram illustrating the overview of the engagement score. [Figure 4] Figure 4 is an explanatory diagram illustrating the overview of the multiple-action consideration model. [Figure 5] Figure 5 is an explanatory diagram illustrating the overview of the potential estimation for each funnel. [Figure 6] Figure 6 is an explanatory diagram illustrating the overview of how advertising effectiveness is visualized. [Figure 7] Figure 7 is an explanatory diagram illustrating the overview of brand preference evaluation compared to competitors. [Figure 8] Figure 8 shows an example of the configuration of a terminal device according to an embodiment. [Figure 9] Figure 9 shows an example of the configuration of a server device according to this embodiment. [Figure 10] Figure 10 is a flowchart showing the processing procedure according to the embodiment. [Figure 11] Figure 11 shows an example of a hardware configuration. [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] [1. Overview of the Information Processing System] First, with reference to Figure 1, an overview of the information processing system according to the embodiment will be described. Figure 1 is an explanatory diagram showing an overview of the information processing system according to the embodiment. As shown in Figure 1, the information processing system 1 according to the embodiment includes a terminal device 10 and a server device 100. The terminal device 10 and the server device 100 are connected to each other via a network N, either by wired or wireless means, enabling communication between them. This allows the terminal device 10 to cooperate with the server device 100. The network N is, for example, a LAN (Local Area Network), a WAN (Wide Area Network), or the Internet.

[0011] Terminal device 10 is an information processing device used by user U. For example, terminal device 10 may be a smart device such as a smartphone or tablet, a PC (Personal Computer) such as a desktop or notebook (laptop), a mobile phone such as a feature phone, a PDA (Personal Digital Assistant), a game console or AV equipment with communication functions, an information appliance or digital appliance, a car navigation system, a wearable device such as a smartwatch, head-mounted display, or smart glasses. Alternatively, terminal device 10 may be a house or building compatible with the Internet of Things (IoT), a car, a home appliance, an electronic device, etc.

[0012] In this embodiment, the terminal device 10 is a smart device such as a smartphone or tablet used by user U, and is a mobile terminal device that can communicate with any server device via wireless communication networks such as LTE (Long Term Evolution), 4G (4th Generation), 5G (5th Generation), Bluetooth (registered trademark), or wireless LAN. The terminal device 10 also has a screen such as a liquid crystal display with touch panel functionality, and accepts various operations on displayed data such as content from user U using a finger or stylus, such as tapping, sliding, and scrolling. Operations performed on the area of ​​the screen where content is displayed may also be considered as operations on the content. Furthermore, the terminal device 10 may be an information processing device such as a desktop PC or notebook PC, not just a smart device.

[0013] The server device 100 is, for example, a computer such as a PC or blade server, or a mainframe or workstation. The server device 100 may also be implemented through cloud computing.

[0014] In this embodiment, the server device 100 is an information processing device that cooperates with the terminal device 10 of each user U and provides the terminal device 10 of each user U with API (Application Programming Interface) services for various applications (hereinafter referred to as apps) and the like, as well as various data, and is realized by a computer, a cloud system, or the like.

[0015] Further, the server device 100 may be an information processing device that provides some online service to the terminal device 10 of each user U. For example, as an online service, the server device 100 may provide services such as Internet connection, search service, chat service, dialogue service by voice, image, video, etc., SNS (Social Networking Service), e-commerce (EC: Electronic Commerce), electronic payment, online game, online banking, online trading, accommodation and ticket reservation, video and music distribution, news, map, route search, route guidance, route information, operation information, weather forecast, etc. Actually, the server device 100 may cooperate with various servers that provide the above-mentioned online services and mediate the online services, or be in charge of the processing of the online services.

[0016] Note that the server device 100 can acquire user information about the user U. For example, as user information, the server device 100 acquires information (attribute information) about attributes of the user U such as the gender, age, and residential area of the user U. Further, the server device 100 can acquire information about attributes such as the demographics (demographic attributes), psychographics (psychological attributes), geographics (geographical attributes), and behavioral (behavioral attributes) of the user U. Also, the server device 100 may acquire, as user information, segments and personas (person images) to which the user U belongs in the field of marketing. Then, the server device 100 stores and manages information (attribute information) about the attributes of the user U together with identification information (such as user ID) indicating the user U.

[0017] Further, the server device 100 acquires various types of history information (log data) indicating the actions of the user U from the terminal device 10 of the user U or from various servers or the like based on the user ID or the like. For example, the server device 100 acquires a location history, which is a history of the location and time of the user U, from the terminal device 10. Also, the server device 100 acquires a search history, which is a history of the search queries input by the user U, from a search server (search engine). Further, the server device 100 acquires a browsing history, which is a history of the content browsed by the user U, from a content server. Also, the server device 100 acquires a purchase history (settlement history), which is a history of the user U's product purchases and settlement processes, from an e-commerce server or a settlement processing server. Further, the server device 100 may acquire a listing history or a sales history, which is a history of the user U's listings on the marketplace, from an e-commerce server or a settlement processing server. Also, the server device 100 acquires a posting history, which is a history of the user U's posts, from a posting server or an SNS server that provides a word-of-mouth posting service. Note that each of the above various servers or the like may be the server device 100 itself. That is, the server device 100 may function as each of the above various servers or the like.

[0018] Also, the number of each device included in the information processing system 1 shown in FIG. 1 is not limited to that shown. For example, in FIG. 1, for simplicity of illustration, only one terminal device 10 is shown, but this is merely an example and is not limited, and two or more may be provided.

[0019] 〔2. Engagement Estimation〕 The following explains engagement estimation with reference to Figures 1 to 7. Figure 2 is an explanatory diagram showing an overview of the purchase funnel. As shown in Figure 2, a general marketing concept is to guide users through the purchase funnel, which diagrams the purchasing behavior of prospective customers into four stages: "awareness," "interest," "comparison and consideration," and "purchase," and to gradually transition them through the funnel stages until conversion (CV). Knowing the user's funnel stage allows for various measures to encourage stage transitions, such as delivering ads tailored to the user, customer analysis, and verifying the effectiveness of ads. For example, one possible measure is to deliver creative content that highlights the product's advantages and encourages consideration to the interested stage.

[0020] Furthermore, as shown in Figure 2, there is a method that uses the conversion rate (CVR) as a measure of engagement (response to advertising). For example, it is presumed that the awareness segment, which has a low CVR, has low engagement, while the purchase segment, which has a high CVR, has high engagement. However, this has the following disadvantages. For example, it is necessary to deliver ads in advance until conversion data is accumulated. Also, it is not possible to measure the effectiveness of ads that do not aim to increase conversions, such as ads that promote awareness of products, services, or events.

[0021] Therefore, in this embodiment, the server device 100 utilizes multiple behavioral logs for estimation, such as behaviors that reflect cognition and interest, and behaviors that can be obtained before ad delivery.

[0022] [2-1. Model considering multiple actions] The server device 100 assumes that users have an unobservable latent value called "level of engagement" with a product, and estimates that the higher the engagement, the more likely various actions related to the product are to occur.

[0023] The server device 100 learns multiple action probabilities (action rates) such as site visit probability (site traffic probability), search probability, click probability, and conversion probability, and in the process calculates an engagement score that indicates the "level of engagement," thereby enabling the estimation of engagement that reflects multiple actions.

[0024] Refer to Figure 3 to explain the engagement score. Figure 3 is an explanatory diagram that shows an overview of the engagement score. For example, as shown in Figure 3, an engagement score of "0.2" corresponds to "Search: Medium," "Site Visits: Small," "Clicks: Small," and "Conversions: Small." Conversely, if "Search: Medium," "Site Visits: Small," "Clicks: Small," and "Conversions: Small," the engagement score will be "0.2." Also, an engagement score of "0.8" corresponds to "Search: Large," "Site Visits: Large," "Clicks: Large," and "Conversions: Medium." Conversely, if "Search: Large," "Site Visits: Large," "Clicks: Large," and "Conversions: Medium," the engagement score will be "0.8."

[0025] Refer to Figure 4 to explain the multi-action consideration model. Figure 4 is an explanatory diagram showing an overview of the multi-action consideration model. As shown in Figure 4, the server device 100 uses a deep neural network (DNN) model as the multi-action consideration model and outputs a single score value for each user as an intermediate output of the model. In the example in Figure 4, the DNN model is represented as DNN Layers. Note that the DNN model is just one example. In practice, a similar model may be used.

[0026] The score calculation logic follows these steps: (A1) The DNN model calculates the score. (A2) Apply a nonlinear transformation to the score to calculate the action probability.

[0027] The above procedure is for the inference stage. In reality, there is a training procedure that precedes the inference stage, which involves "training the model and the parameters (variables) of the nonlinear transformation unit."

[0028] The learning procedure is as follows: (B1) Prepare training data using user features as input and observed values ​​of whether or not various actions are performed as the correct labels. (B2) User features are input into the entire model, including DNN layers and nonlinear transformations, and the action probability is estimated. (B3) The parameters of the DNN layers and the nonlinear transformation are trained so that the error between the estimated action probability and the observed ground truth label is minimized. (B4) During this learning process, the engagement score, which is an intermediate output, is output as a value that reflects multiple actions. In other words, the DNN layers are trained to output the engagement score as a value that compresses information from multiple actions.

[0029] The above describes the training procedure. During inference, the process is as already described; the DNN model is given user features as input, and the engagement score is output.

[0030] Subsequently, the server device 100 applies a monotonically increasing nonlinear transformation to the score, where the action probability increases as the score increases, and uses this transformation as an estimate of the action probability. In this way, the server device 100 can calculate a score that reflects multiple action probabilities by compressing the intermediate output to one dimension. For the nonlinear transformation, one of the activation functions called the sigmoid function can be used.

[0031] Here, when building an engagement estimation model for each advertiser as a multi-action consideration model (creating a model like the one shown in Figure 4 for each advertiser), information about the advertiser is not required during training. However, creating a separate model for each advertiser has the following disadvantages: For example, it is not possible to utilize user behavior data for other advertisers that handle similar products. Also, estimation is not possible for advertisers with limited behavior data. Furthermore, because a different model is used for each advertiser, the analysis for each advertiser becomes somewhat complex. In addition, if there are countless (e.g., tens of thousands) advertisers, managing a separate model for each advertiser becomes difficult.

[0032] Therefore, in this embodiment, the server device 100 utilizes an integrated model that spans advertisers, rather than a model for each advertiser. Such a model can be realized, for example, by training the entire model (modifying connection coefficients) to output values ​​corresponding to the presence or absence of actions such as whether or not a user visited the site, whether or not they performed a search, whether or not they clicked on an ad, and whether or not they converted after clicking on an ad, when the user characteristics of a user are input.

[0033] [2-2. Integrated Model Across Advertisers] As shown in Figure 1, in this embodiment, by adding advertiser information (advertiser features) in addition to user information (user features) to the features input to the engagement estimation model, the engagement scores of multiple advertisers can be calculated with a single model. That is, the server device 100 can input user features and advertiser features into the engagement estimation model and obtain the advertiser's engagement score as the model's output.

[0034] Even when advertiser information is included, the method for calculating the score remains largely unchanged. The only differences are that advertiser information is added to the input features during training and also to the input during inference.

[0035] User features include, for example, user behavior, gender, age, and number of actions performed. In other words, user features include at least one of user behavior, gender, age, and number of actions performed. In practice, user features may also include other features obtained from user U's attribute information, history information, etc. Furthermore, user behavior may include multiple actions.

[0036] Advertiser features are information obtained from the advertisement placement application (insertion order: IO), such as the IO identifier (IO ID) and the IO industry category (IO category). Examples of IO industry categories include beverages, clothing, supplements, etc. In other words, advertiser features include at least one of the IO identifier (IO ID) and the IO industry category (IO category). In practice, advertiser features may also include the advertisement start date, advertisement period, advertisement medium, advertisement location, and brand strength (awareness, credibility) of the advertiser or product.

[0037] For example, as shown in Figure 1, the server device 100 extracts user features by referring to user information about user U (such as user U's attribute information and history information), and converts the extracted user features into a vector (vectorization) to generate a user feature vector (step S1). For example, the server device 100 generates a user feature vector from user behavior, gender, age, and the number of actual actions as user features.

[0038] Next, the server device 100 extracts advertiser features by referring to the advertisement placement application (IO) information, and converts the extracted advertiser features into vectors (vectorization) to generate an advertiser feature vector (step S2). For example, the server device 100 generates an advertiser feature vector from the IO identification information (IO ID) and the IO industry category (IO category) as advertiser features.

[0039] In other words, the server device 100 performs so-called embedding to generate an embedding vector based on user features or advertiser features. At this time, the server device 100 may convert the user features and advertiser features into multidimensional vectors. For example, the server device 100 uses the user features and advertiser features as datasets and converts them into multidimensional real-valued vectors using machine learning techniques such as neural networks. Note that the user feature vector may be generated and sent to the server device 100 by the user U's terminal device 10, rather than by the server device 100. Similarly, the advertiser feature vector may be generated and sent to the server device 100 by the advertiser's terminal device 10, rather than by the server device 100.

[0040] Next, the server device 100 collects information on multiple action probabilities, such as site visit probability, search probability, click probability, and conversion probability (step S3). In practice, the server device 100 may collect information on any one or any combination of site visit probability, search probability, click probability, and conversion probability, rather than all of them. It may also collect information on action probabilities other than site visit probability, search probability, click probability, and conversion probability (other action probabilities).

[0041] Next, the server device 100 trains the engagement estimation model with sets of user feature vectors, advertiser feature vectors, and multiple action probabilities (step S4). In the example in Figure 1, the engagement estimation model is represented as Dot. Dot represents the dot product operation between two vectors. That is, Dot in Figure 1 represents the mathematical operation between the user's embedding vector and the advertiser's embedding vector. The value after the dot product operation is not a vector but is aggregated into a one-dimensional numerical value.

[0042] Next, the server device 100 inputs the user feature vector and the advertiser feature vector into the trained engagement estimation model and calculates an engagement score that reflects multiple action probabilities by compressing the intermediate output to one dimension (step S5).

[0043] Next, the server device 100 estimates the action probability based on the engagement score for the advertiser (step S6). For example, using a trained engagement estimation model, the server device 100 obtains estimated action probabilities by applying a monotonically increasing nonlinear transformation to each of the engagement scores for site visits, searches, clicks, and conversions.

[0044] [2-3. The effects of using engagement estimation] [2-3-1. Estimation of the potential of each funnel] Refer to Figure 5 to explain the potential estimation for each funnel. Figure 5 is an explanatory diagram showing an overview of the potential estimation for each funnel. As shown in Figure 5, the server device 100 can estimate the number of potential users (potential number) for each level of engagement with advertisers (engagement score).

[0045] For example, as shown in Figure 5, the server device 100 estimates the action probability and the number of potential users for each engagement score for each of the following: site visits, searches, clicks, and conversions. This is done using a line graph showing the action probability according to the score, with the horizontal axis representing the engagement score and the vertical axis representing the action probability (action rate), and a bar graph showing the number of potential users according to the score, with the horizontal axis representing the engagement score and the vertical axis representing the number of potential users.

[0046] The source data for the graph is calculated by analyzing the engagement score of all users, focusing on users who saw ads from a specific account (advertiser). Specifically, it obtains the embedded vector for a particular account and takes the dot product (inner product) with each of the embedded vectors for the target users.

[0047] This enables the server device 100 to deliver advertisements tailored to each user's engagement score, according to the specified engagement score.

[0048] Furthermore, by understanding the relationship between multiple actions and scores, the server device 100 can consider advertising strategies that aim to boost specific actions.

[0049] [2-3-2. Visualization of advertising effectiveness] Refer to Figure 6 to explain the visualization of advertising effectiveness. Figure 6 is an explanatory diagram showing an overview of the visualization of advertising effectiveness. As shown in Figure 6, the server device 100 can evaluate advertising effectiveness by measuring the lift in the engagement score before and after ad delivery. The lift in the engagement score refers to the difference between the engagement score at a certain point in time and the engagement score after a certain period of time has passed. In other words, if the score was "0.6" at a past point in time and "0.7" at a new point in time, the lift in the engagement score is "+0.1". That is, the engagement score can be used as a lift value.

[0050] For example, as shown in Figure 6, when the user group is divided into an ad-exposed group and an ad-unexposed group, the engagement score of the ad-exposed group was "0.6" before ad delivery, but increased to "0.7" after ad delivery, showing an increase of "0.1" and indicating that the ad was effective. The engagement score of the ad-unexposed group remained unchanged at "0.6" before and after ad delivery.

[0051] Furthermore, it is possible to observe changes in user behavior during ad delivery for both the ad-exposed group and the ad-unexposed group. In other words, it is possible to observe changes in user behavior when ads are present and changes in user behavior when ads are not present.

[0052] In this way, the server device 100 can observe changes in user behavior considering multiple actions, without being limited to specific actions such as conversions (CVs). Specifically, it can observe changes in user features such as "user behavior" and "number of actions performed." Furthermore, by observing changes in the engagement score, it can indirectly observe when changes in features that increase user engagement have occurred.

[0053] [2-3-3. Brand preference evaluation compared to competitors] Refer to Figure 7 to explain brand preference evaluation compared to competitors. Figure 7 is an explanatory diagram illustrating the overview of brand preference evaluation compared to competitors. When there are two competing advertisers, A and B, we are interested in what factors make users prefer our company over the competitors. In the advertiser integration model, since the user feature vectors are shared, it is possible to analyze later using the model which features contributed to brand preference. This allows us to analyze the factors that led to a preference for one of the brands, as shown in Figure 7.

[0054] For example, as shown in Figure 7, vectorized data with vectorized features can be represented as [1st element, 2nd element, 3rd element], and the features contributing to each element of the vector can be analyzed by model analysis. For example, if the user feature vector is User:[0.1,-0.3,0.6], the advertiser feature vector for advertiser A is Advertiser A:[1.0,1.0,2.0], and the advertiser feature vector for advertiser B is Advertiser B:[2.0,1.0,1.0], then it can be inferred that the 3rd element "2.0" contributes significantly to engagement for advertiser A. Similarly, it can be inferred that the 1st element "2.0" contributes significantly to engagement for advertiser B. This is because, due to the properties of the dot product, the score tends to be higher when elements at the same position on the vector are large. For advertisers with a large 1st element, users with a large 1st element tend to have higher scores.

[0055] [2-4. Another perspective] From another perspective, the server device 100 according to this embodiment acquires user information of user U and scores indicating the likelihood of multiple actions performed by user U, each of which is a different action. The server device 100 also learns a model based on the user information and scores acquired by the acquisition unit, which has a structure that calculates a predetermined score based on the user information and calculates a score indicating the likelihood of performing each action from the predetermined score. This model is a model that estimates various scores from the engagement score.

[0056] The server device 100 may implement the above mechanism using AI (Artificial Intelligence) such as GPT (Generative Pre-trained Transformer). GPT is a text generation AI and a language model capable of generating text using natural language processing.

[0057] [3. Example of terminal device configuration] Next, the configuration of the terminal device 10 will be described using Figure 8. Figure 8 is a diagram showing an example of the configuration of the terminal device 10 according to this embodiment. As shown in Figure 8, the terminal device 10 comprises a communication unit 11, a display unit 12, an input unit 13, a positioning unit 14, a sensor unit 20, a control unit 30 (controller), and a storage unit 40.

[0058] (Communications Section 11) The communication unit 11 is connected to the network N by wire or wireless connection and transmits and receives information to and from the server device 100 via the network N. For example, the communication unit 11 can be implemented using a NIC (Network Interface Card) or an antenna.

[0059] (Display section 12) The display unit 12 is a display device that displays various information such as location information. For example, the display unit 12 may be a liquid crystal display (LCD) or an organic electro-luminescent display (OLED). The display unit 12 may also be a touch panel display, but is not limited to this.

[0060] (Input section 13) The input unit 13 is an input device that receives various operations from the user U. For example, the input unit 13 has buttons for inputting characters, numbers, etc. The input unit 13 may also be an input / output port (I / O port) or a USB (Universal Serial Bus) port. If the display unit 12 is a touch panel display, a part of the display unit 12 functions as the input unit 13. The input unit 13 may also be a microphone that receives voice input from the user U. The microphone may be wireless.

[0061] (Positioning unit 14) The positioning unit 14 receives signals (radio waves) transmitted from GPS (Global Positioning System) satellites and, based on the received signals, acquires position information (e.g., latitude and longitude) indicating the current position of the terminal device 10. In other words, the positioning unit 14 determines the position of the terminal device 10. Note that GPS is just one example of a GNSS (Global Navigation Satellite System).

[0062] Furthermore, the positioning unit 14 can determine its position using various methods other than GPS. For example, the positioning unit 14 may use various communication functions of the terminal device 10 to determine its position as an auxiliary positioning means for position correction, etc., as described below.

[0063] (Wi-Fi positioning) For example, the positioning unit 14 determines the location of the terminal device 10 by utilizing the Wi-Fi® communication function of the terminal device 10 and the communication network provided by each telecommunications company. Specifically, the positioning unit 14 determines the location of the terminal device 10 by performing Wi-Fi communication, etc., and determining the distance to nearby base stations and access points.

[0064] (Beacon positioning) Furthermore, the positioning unit 14 may determine the location using the Bluetooth® function of the terminal device 10. For example, the positioning unit 14 determines the location of the terminal device 10 by connecting to a beacon transmitter connected via the Bluetooth® function.

[0065] (Geomagnetic positioning) Furthermore, the positioning unit 14 determines the position of the terminal device 10 based on the geomagnetic pattern of the structure, which has been measured in advance, and the geomagnetic sensor provided by the terminal device 10.

[0066] (RFID positioning) Furthermore, if, for example, the terminal device 10 is equipped with an RFID (Radio Frequency Identification) tag function equivalent to that of a contactless IC card used at a train station ticket gate or in a store, or if it is equipped with a function to read RFID tags, the location where it was used will be recorded along with the information on the payment or other transactions made by the terminal device 10. The positioning unit 14 may determine the location of the terminal device 10 by acquiring such information. Alternatively, the location may be determined by an optical sensor or infrared sensor equipped in the terminal device 10.

[0067] The positioning unit 14 may, if necessary, determine the position of the terminal device 10 using one or a combination of the positioning means described above.

[0068] (Sensor unit 20) The sensor unit 20 includes various sensors mounted on or connected to the terminal device 10. The connection can be wired or wireless. For example, the sensors may be detection devices other than the terminal device 10, such as wearable devices or wireless devices. In the example shown in Figure 8, the sensor unit 20 includes an acceleration sensor 21, a gyro sensor 22, a barometric pressure sensor 23, a temperature sensor 24, a sound sensor 25, a light sensor 26, a magnetic sensor 27, and an image sensor (camera) 28.

[0069] The sensors 21-28 described above are merely examples and not limiting. In other words, the sensor unit 20 may be configured to include some of the sensors 21-28, or it may include other sensors such as humidity sensors in addition to or instead of the sensors 21-28.

[0070] The acceleration sensor 21 is, for example, a 3-axis acceleration sensor and detects the physical movement of the terminal device 10, such as its direction of movement, velocity, and acceleration. The gyro sensor 22 detects the physical movement of the terminal device 10, such as its tilt in the three axes, based on its angular velocity. The barometric pressure sensor 23 detects the atmospheric pressure around the terminal device 10, for example.

[0071] Since the terminal device 10 is equipped with the acceleration sensor 21, gyroscope 22, barometric pressure sensor 23, etc., it becomes possible to determine the position of the terminal device 10 using technologies such as pedestrian dead-reckoning (PDR) that utilize these sensors 21 to 23. This makes it possible to obtain indoor location information that is difficult to obtain with positioning systems such as GPS.

[0072] For example, a pedometer using an accelerometer 21 can calculate the number of steps, walking speed, and distance walked. Additionally, a gyroscope 22 can be used to determine the user U's direction of movement, gaze direction, and body tilt. Furthermore, the barometric pressure detected by the barometric pressure sensor 23 can be used to determine the altitude and floor number of the user U's terminal device 10.

[0073] The temperature sensor 24 detects, for example, the ambient temperature around the terminal device 10. The sound sensor 25 detects, for example, the ambient sound around the terminal device 10. The light sensor 26 detects the ambient illumination around the terminal device 10. The magnetic sensor 27 detects, for example, the Earth's magnetic field around the terminal device 10. The image sensor 28 captures an image of the area around the terminal device 10.

[0074] The aforementioned pressure sensor 23, temperature sensor 24, sound sensor 25, light sensor 26, and image sensor 28 can detect the surrounding environment and conditions of the terminal device 10 by detecting atmospheric pressure, temperature, sound, and illuminance, respectively, and by capturing images of the surroundings. Furthermore, it becomes possible to improve the accuracy of the location information of the terminal device 10 based on the surrounding environment and conditions.

[0075] (Control Unit 30) The control unit 30 includes, for example, a microcomputer having a CPU (Central Processing Unit) or MPU (Micro Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), input / output ports, and various circuits. Alternatively, the control unit 30 may be composed of hardware such as an integrated circuit (ASIC) or FPGA (Field Programmable Gate Array). The control unit 30 includes a transmission unit 31, a reception unit 32, and a processing unit 33.

[0076] (Transmitter 31) The transmission unit 31 can transmit various information, such as information input by the user U using the input unit 13, various information detected by sensors 21-28 mounted on or connected to the terminal device 10, and location information of the terminal device 10 determined by the positioning unit 14, to the server device 100 via the communication unit 11.

[0077] (Receiving unit 32) The receiving unit 32 can receive various information provided by the server device 100, as well as requests for various information from the server device 100, via the communication unit 11.

[0078] (Processing 33) The processing unit 33 controls the entire terminal device 10, including the display unit 12. For example, the processing unit 33 can output and display various information transmitted by the transmission unit 31 and various information received from the server device 100 by the reception unit 32 to the display unit 12.

[0079] (Storage unit 40) The storage unit 40 is implemented by, for example, semiconductor memory elements such as RAM (Random Access Memory) and flash memory, or by storage devices such as HDD (Hard Disk Drive), SSD (Solid State Drive), and optical discs. Various programs and various data are stored in this storage unit 40.

[0080] [4. Example of Server Device Configuration] Next, the configuration of the server device 100 according to the embodiment will be described using Figure 9. Figure 9 is a diagram showing an example of the configuration of the server device 100 according to the embodiment. As shown in Figure 9, the server device 100 includes a communication unit 110, a storage unit 120, and a control unit 130.

[0081] (Communications Department 110) The communication unit 110 is implemented, for example, by a NIC (Network Interface Card). The communication unit 110 is connected to the network N by wire or wireless connection.

[0082] (Storage unit 120) The storage unit 120 is implemented by, for example, semiconductor memory elements such as RAM (Random Access Memory) and flash memory, or by storage devices such as HDDs, SSDs, and optical discs. The storage unit 120 may store identification information (such as a user ID) indicating user U, as well as attribute information and history information (log data) of user U.

[0083] (Control unit 130) The control unit 130 is a controller, and is realized by executing various programs (corresponding to an example of an information processing program) stored in the internal storage device of the server device 100 using a storage area such as RAM as a working area, for example, by a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), or FPGA (Field Programmable Gate Array). In the example shown in Figure 9, the control unit 130 has an acquisition unit 131, a generation unit 132, a learning unit 133, a calculation unit 134, an estimation unit 135, an evaluation unit 136, and a provision unit 137.

[0084] (Acquisition part 131) The acquisition unit 131 acquires the search query entered by the user U. For example, when the user U enters a search query into a search engine or the like and performs a keyword search, the acquisition unit 131 acquires the search query via the communication unit 110. In other words, the acquisition unit 131 acquires the keyword entered by the user U into the search box of a search engine, website, or application via the communication unit 110.

[0085] Furthermore, the acquisition unit 131 acquires user information about user U via the communication unit 110. For example, the acquisition unit 131 acquires identification information (such as user ID), location information, and attribute information of user U from user U's terminal device 10. The acquisition unit 131 may also acquire identification information and attribute information of user U when user U is registered. The acquisition unit 131 then stores the user information in the storage unit 120.

[0086] Furthermore, the acquisition unit 131 acquires various historical information (log data) indicating the user U's actions via the communication unit 110. For example, the acquisition unit 131 acquires various historical information indicating the user U's actions from the user U's terminal device 10, or from various servers based on the user ID, etc. The acquisition unit 131 then stores the various historical information in the storage unit 120.

[0087] Furthermore, the acquisition unit 131 acquires multiple action probabilities related to the merchandise via the communication unit 110. In other words, the acquisition unit 131, acting as a collection unit, collects multiple action probabilities related to the merchandise.

[0088] (Generation unit 132) The generation unit 132 generates a user feature vector by vectorizing user features, including user behavior based on multiple behavior logs. Furthermore, the generation unit 132 generates an advertiser feature vector by vectorizing advertiser features based on advertisement placement applications.

[0089] (Learning Section 133) The learning unit 133 trains a single model with sets of user feature vectors and multiple action probabilities. For example, the learning unit 133 trains a single advertiser integrated model with sets of user feature vectors, advertiser feature vectors, and multiple action probabilities.

[0090] (Calculation section 134) The calculation unit 134 inputs the user feature vector into a single model and compresses it to one dimension in the intermediate output to calculate an engagement score that reflects multiple action probabilities. For example, the calculation unit 134 inputs the user feature vector and the advertiser feature vector into an advertiser integration model and compresses it to one dimension in the intermediate output to calculate an engagement score that reflects multiple action probabilities. Alternatively, the calculation unit 134 uses a DNN model and outputs a single engagement score for each user in the intermediate output of the DNN model.

[0091] (Estimation part 135) The estimation unit 135 performs engagement estimation using the engagement score. For example, the estimation unit 135 estimates the number of users for each engagement score for advertisers. The estimation unit 135 also uses the engagement score as the basis for the estimated action probability by applying a monotonically increasing nonlinear transformation.

[0092] (Evaluation Section 136) The evaluation unit 136 assesses the effectiveness of advertising by measuring the lift in engagement scores before and after ad delivery. For example, the evaluation unit 136 observes changes in user behavior during ad delivery that reflect multiple actions, regardless of specific actions. In addition, the evaluation unit 136 performs brand preference evaluations for multiple competing advertisers in comparison to competitors.

[0093] (Providing Department 137) The provision unit 137 provides advertisers with engagement estimation and evaluation results via the communication unit 110. The provision unit 137 also provides advertisers with suggestions for advertising strategies that will boost specific actions, based on the relationship between multiple actions and scores.

[0094] [5. Processing Procedure] Next, the processing procedure by the server device 100 according to the embodiment will be described using Figure 10. Figure 10 is a flowchart of the processing procedure according to the embodiment. Note that the processing procedure shown below is repeatedly executed by the control unit 130 of the server device 100.

[0095] For example, as shown in Figure 10, the acquisition unit 131 of the server device 100, acting as a collection unit, collects multiple action probabilities related to the product (step S101).

[0096] Next, the generation unit 132 of the server device 100 vectorizes user features, including user behavior based on multiple activity logs, to generate a user feature vector (step S102).

[0097] Furthermore, the generation unit 132 of the server device 100 vectorizes advertiser features based on the advertisement placement application to generate an advertiser feature vector (step S103).

[0098] Next, the learning unit 133 of the server device 100 trains a single advertiser integrated model with sets of user feature vectors, advertiser feature vectors, and multiple action probabilities (step S104).

[0099] Next, the calculation unit 134 of the server device 100 inputs the user feature vector and the advertiser feature vector into the advertiser integration model and compresses them to one dimension in the intermediate output to calculate an engagement score that reflects multiple action probabilities (step S105).

[0100] Next, the estimation unit 135 of the server device 100 uses the engagement score as the basis for an estimated value of the action probability by applying a monotonically increasing nonlinear transformation (step S106).

[0101] Furthermore, the estimation unit 135 of the server device 100 estimates the number of users for each engagement score for advertisers (step S107).

[0102] Furthermore, the evaluation unit 136 of the server device 100 evaluates the effectiveness of the advertisement by measuring the lift in the engagement score before and after the advertisement delivery (step S108).

[0103] Furthermore, the evaluation unit 136 of the server device 100 performs a brand preference evaluation for multiple advertisers in a competitive relationship, comparing them with their competitors (step S109).

[0104] Furthermore, the provision unit 137 of the server device 100 provides the advertiser with the engagement estimation results and evaluation results via the communication unit 110 (step S110).

[0105] [6. Variant Example] The terminal device 10 and server device 100 described above may be implemented in various other forms besides those of the embodiment described above. Therefore, the following describes modifications of the embodiment.

[0106] In the above embodiment, some or all of the processing performed by the server device 100 may actually be performed by the terminal device 10 (or an application running on the terminal device 10). For example, the processing may be completed in a standalone manner (by the terminal device 10 alone). In this case, the terminal device 10 is assumed to have the same functions as the server device 100 in the above embodiment. Furthermore, in the above embodiment, since the terminal device 10 is in cooperation with the server device 100, from the perspective of the user U, it appears as if the processing of the server device 100 is also being performed by the terminal device 10. In other words, from another perspective, it can be said that the terminal device 10 is equipped with the server device 100.

[0107] [7. Effects] As described above, the information processing device (terminal device 10 and server device 100) according to the present invention is characterized by comprising: a generation unit 132 that generates a user feature vector by vectorizing user features including user behavior based on multiple behavior logs; a collection unit (acquisition unit 131) that collects multiple action probabilities related to a product; a learning unit 133 that trains a single model with sets of user feature vectors and multiple action probabilities; a calculation unit 134 that inputs the user feature vector to the single model and compresses it to one dimension at the intermediate output to calculate an engagement score that reflects multiple action probabilities; and an estimation unit 135 that performs engagement estimation using the engagement score.

[0108] This makes it possible to estimate engagement that reflects multiple actions.

[0109] The generation unit 132 further vectorizes advertiser features based on ad placement applications to generate advertiser feature vectors. The learning unit 133 trains a single advertiser integrated model with sets of user feature vectors, advertiser feature vectors, and multiple action probabilities. The calculation unit 134 inputs the user feature vectors and advertiser feature vectors into the advertiser integrated model and compresses them to one dimension at the intermediate output to calculate an engagement score that reflects multiple action probabilities. The estimation unit 135 uses the engagement score to perform engagement estimation.

[0110] This allows the engagement scores of multiple advertisers to be calculated using a single model.

[0111] The estimation unit 135 estimates the number of users for each engagement score for advertisers.

[0112] This makes it possible to deliver ads tailored to each user's engagement level, based on their specified engagement score.

[0113] The information processing device according to the present invention further comprises an evaluation unit 136 that evaluates the effectiveness of an advertisement by measuring the lift in the engagement score before and after the delivery of the advertisement.

[0114] This allows for the evaluation and visualization of advertising effectiveness using engagement scores.

[0115] The evaluation unit 136 observes changes in user behavior during ad delivery that reflect multiple actions, regardless of any specific action.

[0116] This makes it possible to observe and visualize changes in user behavior during ad delivery that reflect multiple actions.

[0117] The evaluation unit 136 performs brand preference evaluations on multiple advertisers that are in a competitive relationship, comparing them with their competitors.

[0118] This allows us to analyze the factors behind a preference for a particular brand.

[0119] The calculation unit 134 uses a DNN model and outputs a single engagement score for each user as an intermediate output of the DNN model.

[0120] This allows for the calculation of a single score that reflects the probability of multiple actions for each user.

[0121] The estimation unit 135 uses the engagement score as the basis for the estimated action probability, applying a monotonically increasing nonlinear transformation to obtain the result.

[0122] By solving this problem as an estimation of engagement levels, it becomes possible to estimate engagement that reflects multiple actions.

[0123] Through any or a combination of the above-described processes, the information processing device according to the present invention can enable the estimation of engagement that reflects multiple actions.

[0124] [8. Hardware Configuration] Furthermore, the terminal device 10 and server device 100 according to the above-described embodiment are realized by a computer 1000 having a configuration such as that shown in Figure 11. The following explanation will use the server device 100 as an example. Figure 11 is a diagram showing an example of the hardware configuration. 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 interface 1060, an input interface 1070, and a network interface 1080 are connected by a bus 1090.

[0125] 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 executes various processes. The arithmetic unit 1030 can be implemented using, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field Programmable Gate Array).

[0126] The primary storage device 1040 is a memory device, such as RAM (Random Access Memory), that temporarily stores data used by the arithmetic unit 1030 for various calculations. The secondary storage device 1050 is a storage device where 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.

[0127] 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.

[0128] 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.

[0129] 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.

[0130] 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.

[0131] 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.

[0132] 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.

[0133] 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.

[0134] [9. 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.

[0135] 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.

[0136] 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.

[0137] For example, the server device 100 described above may be implemented using multiple server computers, and the configuration can be flexibly changed, such as by calling external platforms via APIs (Application Programming Interfaces) or network computing depending on the function.

[0138] Furthermore, the embodiments and modifications described above can be combined as appropriate, provided that the processing content is not inconsistent.

[0139] 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]

[0140] 1. Information Processing System 10 Terminal devices 100 Server Devices 110 Communications Department 120 Storage section 130 Control Unit 131 Acquisition Department 132 Generation part 133 Learning Department 134 Calculation Section 135 Estimation Department 136 Evaluation Department 137 Provision Department

Claims

1. A generation unit that generates a user feature vector by vectorizing user features including user behavior based on multiple behavior logs, A data collection unit that collects multiple action probabilities related to the product, A learning unit that trains a single model on the combination of the user feature vector and the multiple action probabilities, A calculation unit inputs the user feature vector into the single model and compresses it to one dimension at the intermediate output to calculate an engagement score that reflects the multiple action probabilities, An estimation unit that performs engagement estimation using the aforementioned engagement score, An information processing device characterized by comprising:

2. The generation unit further vectorizes advertiser features based on the advertisement placement application to generate an advertiser feature vector, The learning unit trains a single advertiser integrated model with the user feature vector, the advertiser feature vector, and the set of multiple action probabilities. The calculation unit inputs the user feature vector and the advertiser feature vector into the advertiser integration model and compresses them to one dimension in the intermediate output to calculate an engagement score that reflects the multiple action probabilities. The estimation unit performs engagement estimation using the engagement score. The information processing apparatus according to feature 1.

3. The estimation unit estimates the number of users for each engagement score for advertisers. The information processing apparatus according to feature 2.

4. The evaluation unit evaluates the effectiveness of advertising by measuring the lift in the engagement score before and after ad delivery. The information processing apparatus according to claim 2, further comprising:

5. The aforementioned evaluation unit observes changes in user behavior during ad delivery that reflect multiple actions, regardless of any specific action. The information processing apparatus according to feature 4.

6. The aforementioned evaluation unit performs brand preference evaluations for multiple advertisers in a competitive relationship, comparing them with their competitors. The information processing apparatus according to feature 4.

7. The calculation unit utilizes a DNN model and outputs a single engagement score for each user using the intermediate output of the DNN model. The information processing apparatus according to feature 1.

8. The estimation unit uses the engagement score as the basis for its estimate of the action probability, applying a monotonically increasing nonlinear transformation to obtain the result. The information processing apparatus according to feature 7.

9. An information processing method performed by an information processing device, A generation process that generates a user feature vector by vectorizing user features that include user behavior based on multiple behavior logs, A data collection process to gather multiple action probabilities related to the product, A learning process in which a single model is trained with the set of user feature vectors and the multiple action probabilities, A calculation step involves inputting the user feature vector into the single model and compressing it to one dimension at the intermediate output to calculate an engagement score that reflects the multiple action probabilities, An estimation process that uses the aforementioned engagement score to estimate engagement, An information processing method characterized by including

10. A generation procedure for generating a user feature vector by vectorizing user features that include user behavior based on multiple behavior logs, A collection procedure for gathering multiple action probabilities related to a product, A learning procedure for training a single model with the aforementioned user feature vector and the aforementioned set of multiple action probabilities, A calculation procedure for calculating an engagement score that reflects the multiple action probabilities by inputting the user feature vector into the single model and compressing it to one dimension at the intermediate output, An estimation procedure for performing engagement estimation using the aforementioned engagement score, An information processing program characterized by causing a computer to execute it.

Images