Information transition prediction device, method, and program

A constrained neural network and language model combination addresses the limitations of existing models by integrating social science knowledge and large-scale data utilization, enhancing opinion prediction accuracy.

JP7589832B2Active Publication Date: 2024-11-26NIPPON TELEGRAPH & TELEPHONE CORP
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2023559259
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-11-10
Publication Date
2024-11-26
Estimated Expiration
2041-11-10

AI Technical Summary

Technical Problem

Existing prediction models for opinion dynamics in social media face challenges in incorporating prior knowledge from the social science field, requiring large amounts of labeled data and time, and are limited by inflexible interaction mechanisms.

Method used

A model combining a constrained neural network and a language model, where the neural network approximates opinion transitions using differential equations representing user interactions, and the language model predicts user opinions, with parameters learned under constraints to reflect social science knowledge, allowing for large-scale data utilization without extensive labeling.

Benefits of technology

This approach enables accurate prediction of future opinions by integrating prior knowledge from the social science field, reducing the time and cost associated with labeling, and improving prediction accuracy using a large volume of social media data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007589832000012
    Figure 0007589832000012
  • Figure 0007589832000013
    Figure 0007589832000013
  • Figure 0007589832000014
    Figure 0007589832000014
Patent Text Reader

Abstract

One aspect of the present invention involves preparing a model obtained by rewriting a predetermined theoretical model with a differential equation representing an interaction relating to information dissemination by a plurality of users, and using this model to generate a first constrained learning model for approximating transition of information dissemination by the users while generating a second learning model for predicting information to be disseminated by the users on the basis of a plurality of pieces of information on posts on social media in the past. Then, at the time of training, one aspect of the present invention involves estimating, for the first learning model, a parameter for minimizing a total likelihood defined by a likelihood of a prediction value of the disseminated information for the posted information to which the information predicted by the second learning model is reflected and a constraint term represented by the differential equation. Then, at the time of prediction, attribute information on a prediction target user and a date and a time are input to the trained first constrained learning model with the set parameter to acquire a prediction result of the disseminated information for the prediction target user at the date and the time from the trained first constrained learning model.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] One aspect of the present invention relates to an information transition prediction device, method, and program for modeling the transition of opinions of a group, for example, and predicting the transition of an individual's opinion. [Background technology]

[0002] In groups, people exchange opinions on a variety of topics, such as policies, new products, and sports games, and as a result of this exchange (interaction), each person's opinion may change.

[0003] Predicting the evolution of people's opinions is an important task in a wide range of fields, such as marketing and election strategy planning. Predictive models are generally used to predict the evolution of opinions. For example, theoretical models called opinion dynamics models are widely used as predictive models. This type of theoretical model is analyzed using, for example, an agent-based model or a differential equation solver. For example, the interaction mechanism is described mathematically, and the evolution of each person's opinion is simulated based on the predefined interaction mechanism.

[0004] Well-known theoretical models of opinion dynamics include the DeGroot model and the Bounded confidence model (BCM). In the framework of these theoretical models, each individual (hereafter referred to as user u) has an opinion x about a specific topic (politics, products, events, etc.) at time t, which is expressed as a continuous value from -1 to 1. u Assume we have (t) ∈ [-1, 1], where -1 represents the strongest negative opinion, 1 represents the strongest positive opinion, and 0 represents a neutral opinion. Below we explain the outline of each model using mathematical formulas.

[0005] The DeGroot model assumes that opinions change in discrete time according to the following rules:

number

[0006] where x v (t-1) is the opinion of user v at a time point a uv is a parameter that represents the strength of influence of user v on user u. The DeGroot model is based on the opinion x of each user u at a certain time t. u It is based on the assumption that (t) is determined by the weighted sum of opinions of oneself and other users at the previous time t-1.

[0007] In contrast, the bounded confidence model (BCM) is a model that takes into account conformity bias. Conformity bias is a bias that ignores opinions that are far from one's own. If two users have sufficiently similar opinions, these users will influence each other and become closer in opinion. If not, the users do not interact with each other, and each user's original opinion is maintained. In BCM, the opinion update rule is defined as follows:

number

[0008] Here, |Δ| is the absolute value of Δ, ε is a parameter that represents the threshold of difference in opinion, and 1{C} is an indicator function that takes the value 1 if condition C is true and 0 otherwise.

[0009] Theoretical models such as the DeGroot model and BCM are highly interpretable and can reflect knowledge from social science fields when designing interaction mechanisms. However, they have limitations in their prediction accuracy due to the difficulty in handling large amounts of data.

[0010] In recent years, social media such as Twitter (registered trademark), social networking services (SNS), electronic bulletin boards, and blogs have rapidly spread as a means of expressing personal opinions. These social media have become valuable sources of information for analyzing the formation of opinions in groups. When using these data to calibrate theoretical models, it is necessary to manually select models and set parameters, which is time-consuming and laborious.

[0011] Meanwhile, recently, attempts to utilize social media posting history to model and predict opinion evolution have attracted attention, and several approaches based on machine learning models have been proposed in particular. For example, Non-Patent Document 1 describes a method that uses a probabilistic generative model called a point process. This method learns the parameters of a point process model based on the history of opinion evolution estimated from social media posts. However, it lacks flexibility because it uses specific assumptions about the interaction mechanism.

[0012] In addition, Non-Patent Document 2 discloses a method using a deep learning model. This method extracts opinions from text posted on social media and estimates the opinions of each individual in an end-to-end manner. This method is capable of flexibly modeling interactions by utilizing a deep learning model. However, on the other hand, there is a problem in that it is not possible to take into account prior knowledge regarding the mechanism of interactions. [Prior art documents] [Non-patent literature]

[0013] [Non-Patent Document 1] Abir De et al., “Learning and Forecasting Opinion Dynamics in Social Networks.” In Advances in Neural Information Processing Systems 29: Annual Conference on Neural Information Processing Systems 2016. [Non-Patent Document 2] Lixing Zhu et al., “Neural opinion dynamics model for the prediction of user-level stance dynamics.” Information Processing & Management 57, 2 (2020), 102031. Summary of the Invention [Problem to be solved by the invention]

[0014] However, when trying to introduce prior knowledge from the social sciences into the framework of deep learning, the following issues arise. In general, in a physics-informed neural network, parameters are estimated using supervised learning. This requires a sufficient amount of observational data on the variables to be predicted. Opinions are inherently unobservable, and it is usually difficult to secure a sufficient amount of observational data.

[0015] On the other hand, it is possible to assign correct opinion labels to each and every post on social media. However, the task of assigning correct labels to each and every post is time-consuming and costly.

[0016] This invention has been made in light of the above-mentioned circumstances, and aims to provide technology that makes it possible to incorporate prior knowledge in the social sciences when training a learning model, and that enables the utilization of large-scale social media post data without requiring a lot of time or cost, thereby enabling improved accuracy in predicting changes in information. [Means for solving the problem]

[0017] In order to solve the above problem, one aspect of the information transition prediction device or prediction method according to the present invention prepares a model in which a predetermined theoretical model is rewritten by a differential equation that represents interactions regarding information transmission by multiple users, and generates a first learning model with constraints that approximates the transition of the user's transmitted information using this model, and generates a second learning model that predicts and outputs the user's transmitted information based on multiple past posted information on social media. Then, during learning, for the first learning model, parameters are estimated that minimize the overall likelihood defined by the likelihood of the predicted value of transmitted information for the posted information reflecting the output of the second learning model and the constraint term represented by the differential equation. Then, during prediction, attribute information and date and time of the user to be predicted are input to the trained first learning model in which the parameters are set, and a prediction result of the transmitted information of the user to be predicted at the date and time is obtained from the first learning model.

[0018] According to one embodiment of the present invention, parameters of a first learning model are learned under the constraints of a differential equation representing interactions between users and with information predicted from posted information by a second learning model reflected.

[0019] Therefore, it is possible to predict future posts by users by incorporating prior knowledge in the social sciences that is the basis of the theoretical model of opinion dynamics and by utilizing a large amount of past posted information. This makes it possible to predict future posts by users with high accuracy while reducing the significant amount of time and cost required for the task of labeling correct answers. Effect of the Invention

[0020] In other words, according to one aspect of the present invention, it is possible to provide a technology that makes it possible to incorporate prior knowledge in the social sciences when training a learning model, and that makes it possible to utilize large-scale social media posting data without requiring a lot of time or cost, thereby improving the accuracy of predicting changes in information. [Brief description of the drawings]

[0021] [Figure 1] FIG. 1 is a diagram showing an example of the overall configuration of a social media network system including an information transition prediction device according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a block diagram showing an example of a hardware configuration of an information transition prediction device according to an embodiment of the present invention. [Diagram 3] FIG. 3 is a block diagram showing an example of a software configuration of an information transition prediction device according to an embodiment of the present invention. [Figure 4] FIG. 4 is a flowchart showing the procedure and contents of the model learning process executed by the control unit of the information transition prediction device shown in FIG. [Diagram 5] FIG. 5 is a flowchart showing the procedure and contents of the opinion transition prediction process executed by the control unit of the information transition prediction device shown in FIG. [Figure 6] FIG. 6 is a diagram showing an outline of the configuration of a learning model used in the information transition prediction device shown in FIG. [Figure 7] FIG. 7 is a diagram showing an example of social media posting history information acquired by the information transition prediction device shown in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0022] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0023] [One embodiment] (overview) In one embodiment of the present invention, a prediction method using semi-supervised learning of a constrained neural network is described. The prediction method using semi-supervised learning predicts the opinions of individuals at any time in the future, given a social media posting history in which some opinions are labeled with correct answer labels.

[0024] It is also possible to use unsupervised learning methods. Prediction methods using unsupervised learning predict the opinions of individuals at any time in the future based on their social media posting history, to which no correct opinion labels have been assigned.

[0025] Fig. 6 shows an example of the configuration of a learning model used in an embodiment of the present invention. As shown in the figure, the learning model used in the embodiment is composed of a constrained neural network MD1 as a first learning model and a language model MD2 as a second learning model. The constrained neural network MD1 has a function of approximating the time change of opinions. The language model MD2 has a function of predicting the opinions of individual users based on social media posts.

[0026] In one embodiment, when estimating parameters of the constrained neural network MD1, the parameters are learned under the constraints of a differential equation that represents the mechanism of interaction in the exchange of opinions among a plurality of users.

[0027] (Configuration example) (1) System FIG. 1 is a diagram showing an example of the overall configuration of a social media network system according to an embodiment of the present invention.

[0028] A social media network system according to one embodiment enables multiple user terminals UT1 to UTn to access a server device SV via a network NW, thereby enabling multiple users to register and view posted information via the server device SV.

[0029] The user terminals UT1 to UTn are, for example, smartphones, tablet terminals, or personal computers, and are equipped with a browser and predetermined applications in order to use the above social media networks.

[0030] The server device SV is composed of a server computer located on the Web or the cloud, for example, and performs system management and transmission / reception of posted information transmitted from the user terminals UT1 to UTn. The server device SV also has a large-capacity data storage device, and manages the posted information transmitted from each of the user terminals UT1 to UTn as posting history information using the data storage device.

[0031] The network NW includes a wide area network with the Internet at its core, and an access network for accessing the wide area network. For example, a public communication network using wired or wireless communication, or a LAN (Local Area Network) using wired or wireless communication may be used as the access network.

[0032] (2) Information transition prediction device PD In one embodiment of the system, an information transition prediction device PD is provided to model the transition of opinions of a set of users and predict the change in the opinion of any user. The information transition prediction device PD is, for example, a personal computer or a server computer.

[0033] 2 and 3 are block diagrams showing an example of a hardware configuration and a software configuration, respectively, of the information transition prediction device PD.

[0034] The information transition prediction device PD includes a control unit 1 using a hardware processor such as a central processing unit (CPU). A storage unit having a program storage unit 2 and a data storage unit 3, a communication interface (hereinafter, the interface is referred to as I / F) unit 4, and an input / output I / F unit 5 are connected to the control unit 1 via a bus 8.

[0035] Under the control of the control unit 1, the communication I / F unit 4 performs data communication with the server device SV using a communication protocol defined by the network NW.

[0036] An input device 6 and an output device 7 are connected to the input / output I / F unit 5. The input device 6 includes, for example, a keyboard and a mouse, and is used to input various operation information related to the information transition prediction process. The output device 7 includes, for example, a display device, and displays the prediction information generated by the control unit 1. Note that a printer or an external storage device may also be used as the output device 7.

[0037] The program storage unit 2 is configured by combining a non-volatile memory such as a solid state drive (SSD) that can be written and read at any time as a storage medium, and a non-volatile memory such as a read only memory (ROM), and stores application programs required for executing various control processes according to an embodiment, in addition to middleware such as an operating system (OS). Hereinafter, the OS and each application program will be collectively referred to as a program.

[0038] The data storage unit 3 is, for example, a combination of a non-volatile memory such as an SSD that can be written to and read from at any time as a storage medium, and a volatile memory such as a RAM (Random Access Memory), and includes a posting history storage unit 31, a language model storage unit 32, a constrained learning model storage unit 33, a parameter storage unit 34, and a prediction information storage unit 35 as main storage units required to implement one embodiment.

[0039] The posting history storage unit 31 is used to store social media posting history information acquired from the server device SV.

[0040] The language model storage unit 32 stores the language model MD2 generated by the control unit 1. The language model MD2 is used to predict the transition of the user's opinions based on the posting history information.

[0041] The constrained learning model storage unit 33 stores information about the configuration of the constrained neural network MD1. The constrained neural network MD1 approximates the time change of opinions based on the posting history information.

[0042] The parameter storage unit 34 is used to store the optimization parameters of the constrained learning model obtained by the control unit 1.

[0043] The prediction information storage unit 35 is used to store information indicating the prediction results of changes in the opinions of individuals obtained by the control unit 1.

[0044] The control unit 1 includes, as processing functions necessary for implementing one embodiment, a posting history management processing unit 11, a posting history acquisition processing unit 12, a parameter estimation processing unit 13, a prediction processing unit 14, and a prediction information output processing unit 15. All of these processing units 11 to 15 are realized by causing a hardware processor of the control unit 1 to execute an application program stored in the program storage unit 2.

[0045] In addition, the above application programs may be stored in advance in the program storage unit 2, or may be downloaded from an external application server or the like when necessary and stored in the program storage unit 2.

[0046] When, for example, a system administrator inputs various operation information for the posting history information of the server device SV on the input device 6 of the information transition prediction device PD, the posting history management processing unit 11 performs information operation on the posting history information of the server device SV in response to this operation information. Types of information operation include, for example, registration, modification, and deletion of information. Note that the information operation function for the posting history information does not necessarily need to be provided by the information transition prediction device PD, and may be provided by a management terminal or the like attached to the server device SV.

[0047] The posting history acquisition processing unit 12 acquires the necessary posting history information from the server device SV via the communication I / F unit 4 when constructing a learning model consisting of a constrained neural network MD1 and a language model MD2, and performs a process of storing the acquired posting history information in the posting history memory unit 31.

[0048] The parameter estimation processing unit 13 performs a process of rewriting the theoretical model into the form of a differential equation, and generates a constrained neural network MD1 using the rewritten model. The parameter estimation processing unit 13 also reads the posting history information from the posting history storage unit 31, and constructs a language model MD2 based on the reading posting history information.

[0049] The parameter estimation processing unit 13 estimates parameters of the constrained neural network MD1 using the constraint conditions defined by the differential equation and the prediction information of the language model MD2, and stores the estimated parameters in the parameter storage unit 34. An example of the parameter estimation process will be described in detail in the operation example.

[0050] The prediction processing unit 14 sets the optimization parameters stored in the parameter storage unit 34 for the constrained neural network MD1 stored in the constrained learning model storage unit 33, thereby constructing a trained constrained neural network MD1.

[0051] The prediction processing unit 14 receives the user ID of the user u to be predicted and the date and time t to be predicted, which are inputted to the input device 6, via the input / output I / F unit 5, and inputs the received user u to be predicted and the date and time t to be predicted to the trained constrained neural network MD1. Then, the value outputted from the trained constrained neural network MD1 is stored in the prediction information storage unit 35 as a prediction result.

[0052] The prediction information output processing unit 15 reads prediction information from the prediction information storage unit 35 in response to a request to output a prediction result inputted via the input device 6, and performs processing to output the read prediction information to the output device 7 via the input / output I / F unit 5.

[0053] (Example of operation) Next, an example of the operation of the information transition prediction device PD configured as above will be described.

[0054] (I) Model learning process 4 is a flowchart showing an example of the processing procedure and processing contents of the model learning processing executed by the control unit 1 of the information transition prediction device PD. Note that, in this example, a case where a semi-supervised learning method is used as a constrained neural network will be described as an example, but it can be easily extended to a case where an unsupervised learning method is used.

[0055] The control unit 1 of the information transition prediction device PD determines whether or not the learning mode has been instructed in step S10. In this state, when, for example, a system administrator instructs execution of the learning mode, the control unit 1 of the information transition prediction device PD executes a model learning process as follows.

[0056] (1) Acquisition of posting history information The control unit 1 of the information transition prediction device PD first accesses the server device SV under the control of the posting history acquisition processing unit 12, and receives social media posting history information from the server device SV via the communication I / F unit 4. Then, the received posting history information is stored in the posting history storage unit 31. At this time, for example, a topic to be acquired (e.g. politics, products, events) may be specified, and posting information related to the specified topic may be selectively acquired. An example of the acquired posting history information is shown in FIG. 7.

[0057] The social media posting history information to be analyzed consists of a series of posted sentences on a specific topic, i , t i , d i )} I i=1 where u i is the user, d i is a social media post on a particular topic, i indicates the posting time, and I indicates the number of data. Let U be the number of users.

[0058] Here, for some sentences, the correct answer label y i The correct label is, for example, the opinion category y i ∈ {-1 , 0 , 1}, or a score y indicating the degree of positivity i ∈ {1 , 2 , 3 , 4 , 5}, etc. For example, if we collect political tweets during an election period and select tweets that are positive about a political party, i = 1 for negative posts i = -1 for neutral posts i = 0 to generate labeled data. The history of posted sentences with correct opinion labels is defined as Ds = {(u i , t i , y i )} Is i=1 , the history of posted sentences without correct labels is Du = {(u i , t i , d i)} Iu i=1 Here, Is denotes the number of labeled data, and Iu denotes the number of unlabeled data.

[0059] (2) Parameter estimation of the learning model (2-1) Rewriting the theoretical model In step S11, the control unit 1 of the information transition prediction device PD, under the control of the parameter estimation processing unit 13, first rewrites the theoretical model in the form of a differential equation in order to introduce knowledge in the social science field that is the basis of the theoretical model of opinion dynamics into the machine learning model. For example, theoretical models such as the DeGroot model and the Bounded confidence model (BCM) are generally rewritten using the following differential equation.

number

[0060] where t is the time, x u (t) is the opinion of user u at time t, and Λ is the parameter of the differential equation. Φ(·) is a function that represents the interaction mechanism and is designed based on the assumptions of the theoretical model. For example, when using the assumptions of the bounded confidence model (equation (2)), Φ(·) is defined as follows:

number

[0061] When using the DeGroot model (equation (1)), Φ(·) is defined as follows:

number

[0062] Here, a uv is a parameter representing the strength of interaction between user u and user v. Note that, although two representative theoretical models are used in this example, the present invention is also applicable to other theoretical models.

[0063] (2-2) Approximation of neural networks Next, in step S12, the parameter estimation processing unit 13 estimates the variable x u (t) is a neural network f(·;θ f ) is approximated as follows:

number

[0064] (2-3) Building a language model The parameter estimation processing unit 13 then constructs a language model in step S13. The language model is used to extract opinions from social media posts, and takes a post text d on a specific topic as input and extracts opinions o ~ The model outputs the following:

number

[0065] where g(·) is the language model, θ g are the parameters of the language model. The language model g(·) can be freely designed to suit the problem. For example, one possible approach to designing the language model g(·) is to fine-tune a pre-trained deep learning model.

[0066] (2-4) Parameter estimation In the learning mode, the parameters Λ of the differential equation and the parameters θ of the constrained neural network MD1 are set to minimize the likelihood L defined by the following equation (8). f, the parameters θ of the language model MD2 g are estimated respectively.

number

[0067] Here, L s is the likelihood of a social media posting history with a correct label, L u is the likelihood for the posting history without the correct label, L ode is a term that represents the constraint by the differential equation. α and λ are hyperparameters that represent the importance of each term.

[0068] The first term L from equation (8) s (θ f ;D s ), the above equation (8) is the same as the likelihood in unsupervised learning. The first term on the right-hand side of equation (8) is the predicted value of the opinion by the neural network x u ^ (t) and the correct answer label y i It is defined as the error between

number

[0069] where L(·) is the cross-entropy. σ(·) is a continuous value x representing the strength and polarity (positive or negative) of the opinion. u ^ (·) is a function that embeds the label space. When there are two types of correct opinion labels, a sigmoid function is used, and when there are multiple classes, a softmax function is used. Also, Ds = {(u i , t i , y i )} Is i=1 is the posting history with the correct label, and Is is the number of labeled data.

[0070] Unlabeled Post History Du = {(u i , t i , d i )} Iu i=1For x, we use the prediction result of the language model MD2 instead of the correct label. Specifically, the output x of the constrained neural network MD1 u ^ (t) and the output of the language model MD2, o ~ Minimize the squared error between

number

[0071] The third term on the right side of the above equation (8) is defined as follows:

number

[0072] where {T1, …,T J} is a predefined representative point on the time axis. ode represents the error of the differential equation at the J representative points. In this example, the mean square error is used as an index of error, but other error functions can also be used. The above equation (11) is expressed as the time Τ of the representative point by taking the difference between the left and right sides of the differential equation (equation (3)). J By adding the constraint term in the above equation (11) to the likelihood L, the output x u ^ The parameters θ of the constrained neural network MD1 such that (t) follows the differential equation defined by (3) f can be learned.

[0073] In addition, the differential dx of the first term on the right side of equation (11) u ^ (t) / dt can be easily calculated using automatic differentiation implemented in deep learning packages such as PyTorch.

[0074] In addition, the parameter θ f Any method may be used for optimizing (11). However, in order to efficiently calculate the derivatives included in the equation (11), it is preferable to use the backpropagation method.

[0075] (2-5) Saving parameters The parameter estimation processing unit 13 estimates the optimization parameters θ of the constrained neural network MD1 estimated as described above. f ^ and the optimization parameters θ of the language model MD2 g The set of ^ is stored in the parameter storage unit 34 in step S15.

[0076] (II) Prediction Processing The control unit 1 of the information transition prediction device PD uses the learning model constructed as described above to predict future transmitted information of a specific individual user as follows. FIG. 5 is a flowchart showing an example of the procedure and contents of the prediction process executed by the control unit 1 of the information transition prediction device PD.

[0077] In order to predict future information to be transmitted by a particular individual user, for example, when a system administrator specifies a prediction mode on an input device 6, the control unit 1 of the information transition prediction device PD receives the specified input of the prediction mode via the input / output I / F unit 5 and sets the prediction mode.

[0078] (1) Prediction When the prediction mode is set, the control unit 1 of the information transition prediction device PD, under the control of the prediction processing unit 14, first receives, in step S21, attribute information of the user u to be predicted, such as a user ID or profile, and a date and time t to be predicted, which are inputted to the input device 6, via the input / output I / F unit 5. Next, in step S22, the prediction processing unit 14 receives the optimization parameters θ f ^ and the above optimization parameters θ f ^ is set to the constrained neural network MD1. That is, a trained constrained neural network MD1 is constructed.

[0079] Next, in step S23, the prediction processing unit 14 inputs the identification ID of the individual user u to be predicted and the time t to be predicted, which have been input from the input device 6, to the trained constrained neural network MD1. Then, the prediction processing unit 14 receives a prediction result of the transmitted information of the individual user u to be predicted at the prediction time t, which is output from the trained constrained neural network MD1, and stores information representing the received prediction result in the prediction information storage unit 35.

[0080] (2) Output of prediction results In step S24, under the control of the predicted information output processing unit 15, the control unit 1 of the information transition prediction device PD reads out information representing the prediction result from the predicted information storage unit 35 autonomously at a predetermined timing or in response to an input of an output request from the input device 6. Then, the predicted information output processing unit 15 generates, for example, display data of the read out information representing the prediction result, outputs the generated display data from the input / output I / F unit 5 to the output device 7, and displays it on the output device 7. The display data includes, for example, the user ID of the user u to be predicted, the prediction target date and time t, and information representing the prediction result.

[0081] Thus, the prediction result of the outgoing information of the prediction target user u at the prediction target date and time t is displayed on the output device 7. Note that the information representing the prediction result may not only be displayed on the output device 7, but may also be printed out on the output device 7 or stored in an external storage medium such as a USB memory, or may be transmitted from the communication I / F unit 4 to another user terminal or the like via the network NW.

[0082] (Action and effect) As described above, in one embodiment, the system is based on a learning model described using differential equations representing interactions between a plurality of users who transmit information, and includes a constrained neural network that approximates the transition of the transmitted information of the user, and a language model that predicts and outputs the transmitted information of the user based on posted information on social media. During learning, parameters that minimize the likelihood of the predicted value of the transmitted information for the posted information reflecting the output of the language model and the constraint term by the differential equation are estimated, and during prediction, attribute information of the user to be predicted and date and time are input as explanatory variables to the trained constrained neural network in which the estimated parameters are set, and information output from the constrained neural network as a target variable is output as a predicted result of the transmitted information of the user to be predicted at the date and time.

[0083] Therefore, according to one embodiment, a learning model is constructed by combining a constrained neural network and a language model that predicts outgoing information based on a large amount of posted information, and the parameters of the learning model are trained under the constraints of a differential equation that represents interactions between users and in a state that reflects the outgoing information predicted from a large amount of posted information on social media.

[0084] This makes it possible to predict future posts by users by incorporating prior knowledge from the social sciences that underpins the theoretical model of opinion dynamics and by utilizing a large amount of past posted information. In other words, it is possible to predict future posts by users with high accuracy without requiring a large amount of time and cost for annotation.

[0085] [Other embodiments] In the embodiment described above, an example was given in which all the functions of the information transition prediction device PD are provided in a terminal such as a personal computer used by a user. However, all or some of the functions of the information transition prediction device PD may be provided in a server device provided on the Web or the cloud. For example, a parameter estimation processing function during learning may be provided in the server device, and the terminal may download and use a trained constrained neural network from the server device during prediction. In this way, the processing load on the terminal can be reduced.

[0086] In addition, the types and configurations of the first learning model and the second learning model, the processing procedures and processing contents of the parameter estimation processing during learning, the processing procedures and processing contents of the prediction processing during prediction, the type of information to be predicted, etc. can be modified in various ways without departing from the spirit of this invention.

[0087] Although the embodiment of the present invention has been described in detail above, the above description is merely an example of the present invention in every respect. It goes without saying that various improvements and modifications can be made without departing from the scope of the present invention. In other words, in implementing the present invention, specific configurations according to the embodiment may be appropriately adopted.

[0088] In short, this invention is not limited to the above-mentioned embodiment as it is, and in the implementation stage, the components can be modified and embodied without departing from the gist of the invention. In addition, various inventions can be formed by appropriately combining multiple components disclosed in the above-mentioned embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, components from different embodiments may be appropriately combined. [Explanation of symbols]

[0089] PD: Information transition prediction device PT1~PTn: User terminal SV: Server device NW…Network MD1…Constrained Neural Networks MD2…Language model 1...Control unit 2...Program memory section 3. Data storage section 4. Communication I / F section 5...Input / output interface 6. Input Device 7. Output device 8…Bus 11...Posting history management processing section 12...Post history acquisition processing section 13...Parameter estimation processing unit 14...Prediction processing unit 15...Prediction information output processing unit 31...Post history memory section 32…Language model memory unit 33…Constrained learning model memory unit 34...Parameter storage section 35…Prediction information storage unit

Claims

1. An information transition prediction device for predicting a transition of information transmitted by a user in the future, comprising: a first learning model with constraints that approximates the transition of the transmitted information of the user using a theoretical model described by differential equations that represent interactions regarding information transmitted by the multiple users; a second learning model that predicts and outputs the user's transmitted information based on a plurality of pieces of past posted information on social media; a parameter estimation processing unit that estimates, during learning, parameters for the first learning model that minimize an overall likelihood defined by a likelihood of a predicted value of the transmitted information for the posted information reflecting an output of the second learning model and a constraint term represented by the differential equation; a prediction processing unit that inputs attribute information of a specific user to be predicted and a date and time into the first learning model in which the parameters have been set and obtains a prediction result of the transmitted information of the specific user at the date and time from the first learning model at the time of prediction; an output processing unit that outputs the prediction result; An information transition prediction device comprising:

2. The method further includes a post information acquisition processing unit that selectively acquires the post information related to a specific topic from a server device that accumulates and manages the post information, The parameter estimation processing unit constructs the second learning model based on the acquired posted information. The information transition prediction device according to claim 1 .

3. the first learning model is configured by a constrained neural network including the constraint term represented by the differential equation; The second learning model is configured by a language model that receives the posted information related to the specific topic as an input, extracts the transmitted information related to the specific topic from the posted information, and outputs the transmitted information. The information transition prediction device according to claim 2.

4. The parameter estimation processing unit estimates the parameters that minimize a first likelihood of a predicted value for the posted information with a correct answer label, a second likelihood of a predicted value for the posted information using an output of the second learning model as a correct answer label, and a third likelihood defined by a constraint term represented by the differential equation. The information transition prediction device according to claim 1 .

5. the second likelihood and the constraint term are weighted by a hyperparameter that represents an importance to the third likelihood; The information transition prediction device according to claim 4.

6. An information transition prediction method executed by an information processing device that predicts a transition of information transmitted by a user in the future, comprising: A process of rewriting a predetermined theoretical model using differential equations that represent interactions related to information transmission by a plurality of users; generating a first learning model with constraints that approximates the transition of the transmitted information of the user; generating a second learning model that predicts and outputs the user's transmitted information based on a plurality of pieces of past posted information on social media; a step of estimating, during learning, parameters for the first learning model that minimize an overall likelihood defined by a likelihood of a predicted value of the transmitted information for the posted information reflecting an output of the second learning model and a constraint term represented by the differential equation; At the time of prediction, inputting attribute information and a date and time of a specific user to be predicted into the first learning model having been trained and having the parameters set, and acquiring a prediction result of the transmitted information of the specific user to be predicted at the date and time from the first learning model; outputting the prediction result; An information transition prediction method comprising:

7. 7. A program for causing said information processing device to execute each process in each step of said information transition prediction method according to claim 6.

Citation Information

Patent Citations

  • Data processor, data processing method and program

    JP2007299365A