JP2020004030A - Distribution environment providing system for providing distribution environment for artificial intelligence packages - Google Patents

Abstract

To provide an artificial intelligence package distribution environment providing system 1 that is highly convenient for both a person having an issue to be solved and a person having an artificial intelligence package.SOLUTION: Presentation means 13 in a distribution environment providing system 1 comprises: a function of accepting a specification of a certain issue to be solved and presenting an artificial intelligence package related to the certain specified issue to be solved based on the certain specified issue to be solved and information on the artificial intelligence package registered in artificial intelligence package registration means 12; and a function of accepting a specification of a certain artificial intelligence package and presenting the issue to be solved related to the specified certain artificial intelligence package or information on the specified certain artificial intelligence package based on the information on the specified certain artificial intelligence package or the information on the specified certain artificial intelligence package and the issue to be solved registered by issue registering means 11.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a distribution environment providing system that provides a distribution environment for artificial intelligence packages.

  In recent years, abnormality detection technology and optimal control technology using artificial intelligence have been researched, developed, and put to practical use. The flow of applying general artificial intelligence technology to a real system is as follows: AE (self-encoder), CNN (convolutional neural network), LSTM (principal processing of deep learning), F-CNN (forward-sequence neural network) Network), one or more learning inference modules that perform learning and inference functions such as SVM (support vector machine), preprocessing modules such as data reading modules and normalization modules, inference result classification modules, and input contribution calculation Arithmetic processing modules that perform arithmetic processing other than the learning function and the inference function, such as a post-processing module such as a module, are appropriately combined to constitute an artificial intelligence module to which data for learning and inference can be input. Then, a training data set and a verification data set are prepared, and, if necessary, these data sets are converted into data that can be input to the artificial intelligence module. Perform label preparation and prepare in advance. Thereafter, learning is performed on the configured artificial intelligence module using the learning data set, and inference is performed using the verification data set to verify the effectiveness. Thereafter, among the artificial intelligence modules determined to be valid, a learned model determined to be valid, which is included in the artificial intelligence module, is selected, and the learned model is applied to an actual system. ing.

  As described above, the artificial intelligence module is configured by various internal modules such as an arithmetic processing module such as a pre-processing module and a post-processing module, and a learning inference module, and is obtained by a combination of these internal modules. The effects of learning and inference on the module as a whole also vary. In addition, the effects of learning and inference of the artificial intelligence module vary greatly depending on the type of data set used for learning and verification and the type of labeling method applied to the data set as needed. However, the performance of the learned model that is ultimately required differs greatly.

On the other hand, there are various problems in the real system to which the learned model is applied, that is, the problems that the real system wants to solve using the artificial intelligence technology according to the environment of the real system.
Therefore, researchers and developers who apply artificial intelligence modules to real systems need to study the performance of artificial intelligence modules that combine various internal modules, artificial intelligence modules, training data sets, verification data sets, and labeling methods. After verifying the results and accumulating experience, based on the experience gained, the most appropriate artificial intelligence module for the application subject having various problems is selected and designed.

  Based on the accumulated experience, artificial intelligence modules, learning data sets, verification data sets, which have been found to be effective for more general-purpose applications without specializing in specific applications A trend has begun to appear in which the labeling methods and the like are appropriately grouped and sold as an artificial intelligence package to one or more application targets.

  Under such a background, for example, Patent Literature 1 discloses a machine learning model design support device for the purpose of easily designing a complex and advanced machine learning model without being an expert. ing. The technique disclosed in Patent Document 1 generates a machine learning model by connecting a plurality of function nodes each representing a function to be applied to input data and performing an operation as one node, and then performs a machine learning model transaction. It has a function of transmitting the generated machine learning model to the server.

Japanese Patent No. 60074430

  However, Patent Literature 1 discloses that a machine learning model generated by connecting a plurality of function nodes can be traded on a machine learning model transaction server. The detailed configuration is not clear, and does not disclose how the machine learning model transaction server distributes the machine learning model.

  That is, it is not clear how the machine learning model transaction server in Patent Literature 1 relates a machine learning model designed by a designer to an actual system. It was not very convenient for both the system side and those who provided the machine learning model.

  According to a first aspect of the present invention, there is provided a distribution environment providing system for providing a distribution environment of an artificial intelligence package for solving the problem using an artificial intelligence technology. The providing system includes a task registration unit, an artificial intelligence package registration unit, and a presentation unit, and the task registration unit applies, as a problem to be solved associated with an application target system to which the artificial intelligence package is applied. Type of application method of artificial intelligence technology to the target system, type or characteristic of specific variable of the target system, and type or characteristic of specific input variable and specific output variable of the target system One or more functions are accepted and registered, and the artificial intelligence package registration means has a function of generating a learned model. Along with the artificial intelligence package to be executed, as information on the artificial intelligence package, the type of artificial intelligence technology used when the artificial intelligence package learns, the type or characteristics of a specific variable used when the artificial intelligence package learns, the artificial intelligence The intelligent package has a function of accepting and registering one or more of a specific input variable type or characteristic and a specific output variable type or characteristic used when learning, and the presenting means is provided in the assignment registering means. A specification of a registered certain problem to be solved is received, and the specification is performed based on the specified certain problem to be solved and information on one or a plurality of the artificial intelligence packages registered in the artificial intelligence package registration unit. For presenting one or more artificial intelligence packages related to the certain problem to be solved And accepting designation of a certain artificial intelligence package or information related to a certain artificial intelligence package registered in the artificial intelligence package registration means, and information relating to the certain artificial intelligence package registered together with the designated certain artificial intelligence package or designated. The specified artificial intelligence package or the specified artificial intelligence package based on the information on the certain artificial intelligence package and the one or more problems to be solved registered by the problem registering means. There is provided a distribution environment providing system for providing a distribution environment of an artificial intelligence package, which has a function of presenting one or more problems to be solved related to information.

  According to the second aspect of the present invention, in a distribution environment providing system for providing a distribution environment of an artificial intelligence package for solving a problem using artificial intelligence technology, the distribution environment providing system includes: , An artificial intelligence package registering means, and a presenting means, wherein the task registering means includes, as a task to be solved associated with an application target system to which the artificial intelligence package is applied, Accepts and registers one or more of the type of application method, the type or characteristic of a specific variable of the applicable system, and the type or characteristic of a specific input variable and the specific output variable of the applicable system. An artificial intelligence package having a function, wherein the artificial intelligence package registration means has a function of generating a learned model; In addition, as information on the artificial intelligence package, the type of artificial intelligence technology used when the artificial intelligence package learns, the type or characteristics of specific variables used when the artificial intelligence package learns, the artificial intelligence package learns A function of accepting and registering one or more of the type or characteristic of a specific input variable and the type or characteristic of a specific output variable to be used at the time of execution, wherein the presenting means is registered in the assignment registering means. The designation of the problem to be solved is received, and the certain problem to be solved is specified based on the specified problem to be solved and information on one or a plurality of the artificial intelligence packages registered in the artificial intelligence package registration unit. A person having a function of presenting one or more artificial intelligence packages related to a problem to be solved. Distribution environment providing system for providing the distribution system intelligence package is provided.

  According to the third aspect of the present invention, in a distribution environment providing system for providing a distribution environment of an artificial intelligence package for solving a problem using artificial intelligence technology, the distribution environment providing system includes: , An artificial intelligence package registering means, and a presenting means, wherein the task registering means includes, as a task to be solved associated with an application target system to which the artificial intelligence package is applied, Accepts and registers one or more of the type of application method, the type or characteristic of a specific variable of the applicable system, and the type or characteristic of a specific input variable and the specific output variable of the applicable system. An artificial intelligence package having a function, wherein the artificial intelligence package registration means has a function of generating a learned model; In addition, as information on the artificial intelligence package, the type of artificial intelligence technology used when the artificial intelligence package learns, the type or characteristics of specific variables used when the artificial intelligence package learns, the artificial intelligence package learns A function of accepting and registering one or more of the type or characteristic of a specific input variable and the type or characteristic of a specific output variable to be used in the execution, and wherein the presenting unit is registered in the artificial intelligence package registering unit. Accepting the designation of information related to a certain artificial intelligence package or a certain artificial intelligence package, and information related to the certain artificial intelligence package registered with the specified certain artificial intelligence package or information related to the specified certain artificial intelligence package; One or more entries registered by the assignment register An artificial function having a function of presenting one or more problems to be solved related to the specified certain artificial intelligence package or information related to the specified certain artificial intelligence package based on the problem to be solved. A distribution environment providing system for providing a distribution environment of an intelligent package is provided.

  According to the present invention, the problem to be solved associated with the application target system and the artificial intelligence package for solving the problem using the artificial intelligence technology can be registered together in the distribution environment providing system. The problem to be solved registered by the person registering the problem to be distributed in the distribution environment providing system and the artificial intelligence package registered by the person registering the artificial intelligence package in the distribution environment providing system are displayed on the distribution environment providing system. Since they are presented in association with each other, a system that is highly convenient for both can be provided.

1 is a system configuration diagram of a distribution environment providing system 1 that provides a distribution environment of an artificial intelligence package. FIG. 7 is a diagram illustrating an operation of registering an issue to be solved in an issue registration unit 11 by an issue registrant. It is a figure showing the registration operation to the artificial intelligence package registration means 12 by the artificial intelligence package registrant. It is a figure showing the designation | designated operation to the artificial intelligence package presentation means 13 by an artificial intelligence package designator. FIG. 11 is a diagram illustrating a modification of the designation operation to the presentation unit 13.

  Hereinafter, embodiments for implementing the present invention will be described with reference to the drawings. In the following, the range necessary for the description for achieving the object of the present invention is schematically shown, and the range necessary for the description of the relevant part of the present invention will be mainly described. It shall be based on a known technique.

  First, FIG. 1 shows a system configuration diagram of a distribution environment providing system 1 that provides a distribution environment of an artificial intelligence package for solving the problem using the artificial intelligence technology in the present invention.

  The distribution environment providing system 1 shown in FIG. 1 includes an issue registering unit 11, an artificial intelligence package registering unit 12, and an artificial intelligence package presenting unit 13.

The distribution environment providing system 1 includes, for example, a computer having a user interface (not shown) such as a mouse, a keyboard, and a display device, and a server device capable of constructing a cloud system that provides a user interface to a client via a web page or the like. It is realized by. Each of the units included in the distribution environment providing system 1, such as the problem registering unit 11, the artificial intelligence package registering unit 12, and the artificial intelligence package presenting unit 13, is realized by, for example, an application or a partial program. The distribution environment providing system 1 executes the application and the partial program on the distribution environment providing system 1 to cause the respective units such as the task registration unit 11, the artificial intelligence package registration unit 12, the artificial intelligence package presentation unit 13, and the like to function. It has a configuration.
The function of each means will be described in detail below.

<About the task registration means 11>
First, as shown in FIG. 2, the assignment registering unit 11 accepts a registration operation from an assignment registrant and registers the assignment to be solved in the storage area such as an assignment library (not shown) with respect to the assignment associated with the application target system 2. It has a function of performing

  Here, the problem to be solved is a problem to be solved or a problem to be solved in the application target system 2 and is expected to be solved by applying the artificial intelligence technology to the application target system 2. Represents that.

  This can be organized in the form of a type of application method for applying the artificial intelligence technology to the application target system 2. For example, if the application target 2 is a plant, a specific example of the problem to be solved is to extract one or a plurality of sensors (input variables) having a high correlation with an abnormal event. When viewed as an application method, it can be expressed as "data mining" for extracting input variables highly correlated with abnormal events. Further, as another example, if the application target 2 is a vibration prevention control system of a robot installed in a specific environment, specific examples of problems to be solved include a robot vibration amount (an output value of an acceleration sensor) and Optimization of parameters having a known correlation such as current value, motor speed and motor temperature, which can be considered as a method of applying artificial intelligence technology, for variables that are known to correlate with each other in anti-vibration control. It can be expressed as "parameter optimization". As yet another example, if the application target 2 is an apparatus that inspects a product on a production line by using a camera image, a specific example of a problem to be solved is to determine a normal product and an abnormal product based on the characteristics of the camera image. This can be expressed as "clustering", which classifies the camera image of a normal product and the camera image of an abnormal product as a method of applying the artificial intelligence technology.

  Also, if other information is clear about the problem to be solved, that is, if the “variable type or characteristic” in the application system related to the problem to be solved is known, the problem should be solved. The task can be expressed as “variable type or characteristic”. This “variable type or characteristic” is, for example, when the application target 2 is an example of the vibration prevention control system of a robot installed in a specific environment, the vibration amount (variable ), The current value (variable type), the motor speed (variable type), and the motor temperature (variable type) are known to be used. By listing the “output value of the acceleration sensor)”, “current value”, “motor speed”, and “motor temperature”, it can be expressed as a type (or feature) of a variable (related to problem solving). In this case, the “vibration amount of the robot (output value of the acceleration sensor)” may be specified as a type of a variable like “output value of the acceleration sensor”, and an acceleration signal for extracting vibration may be specified. Since it has the property of causing a drastic change in a short time, etc., the "output value of the acceleration sensor" for extracting "vibration amount" is specified as a characteristic of the variable like "signal representing vibration" May be. Similarly, other variables may be specified as the type of the variable, such as the sensor value of “motor temperature”, and the temperature signal, which is a thermal system, has a large time constant of the response value and includes a dead time. Since it has such properties as “characteristics”, “motor temperature” may be specified as a characteristic of a variable like “signal representing temperature”.

  In addition, when further information is known about the problem to be solved, that is, the type of the specific input variable or the characteristic and characteristic of the target system, which indicates a known correlation related to the problem to be solved. If the "type or characteristic of the specific output variable" is known, the problem to be solved can be expressed as "the type or characteristic of the specific input variable and the type or characteristic of the specific output variable". The “type or characteristic of a specific input variable and the type or characteristic of a specific output variable” is, for example, the above-described example in which the application target 2 is a vibration prevention control system of a robot installed in a specific environment. When performing the prevention control, the relationship between the robot vibration amount (type of output variable), current value (type of input variable), motor speed (type of input variable), and motor temperature (type of input variable) If it is known that they are correlated, the problem to be solved is that the type of output variable is “the amount of vibration of the robot (output value of the acceleration sensor)”, the type of input variable is “current value”, It can be expressed as a correlation (adjustment or optimization) with “motor speed” and “motor temperature”. In this case, the “vibration amount of the robot (output value of the acceleration sensor)” may be specified as a type of an output variable like “output value of the acceleration sensor”, and the acceleration for extracting the vibration may be specified. Since the signal has a characteristic of causing a drastic change in a short time, etc., the output value of the acceleration sensor for extracting the "vibration amount" is a characteristic of the output variable like "signal representing vibration". May be specified. Similarly, the input variable may be specified as the type of the input variable, such as the sensor value of “motor temperature”, and the temperature signal which is a thermal system has a large response time constant and includes a dead time. Since it has characteristics such as “motor temperature”, “motor temperature” may be specified as a feature of the input variable like “signal representing temperature”.

  Then, as described above, the registrant assigns “type of application method for applying artificial intelligence technology to application target system 2”, “variable type or characteristic” of application target system 2 and application target system 2 One or more of the problems to be solved expressed as “types or characteristics of specific input variables and types or characteristics of specific output variables” are displayed on a user interface (not shown) of the distribution environment providing system 1. The registration is made to the assignment registering means 11 via the server.

  As for a specific method of registration, the distribution environment providing system 1 may present a user interface (not shown) to directly input the problem to be solved by the problem registrant, or may be registered in the problem registration unit 11 in advance. By displaying a list of problems to be solved or problems to be solved registered in the past, a problem registrant may be allowed to select a problem to be solved.

  In addition, "variable type or characteristic" in the problem to be solved registered in the past or "type or characteristic of specific input variable and type or characteristic of specific output variable" can be input and output in any form. The extracted `` variable type or characteristic '' is displayed to the project registrant in a list, and a desired `` variable type or characteristic '' is selected from the displayed `` variable type or characteristic ''. The "variable type or characteristic" of the application target system 2, which is a problem to be solved, may be selected, or a desired "variable type or characteristic" may be selected from the "variable type or characteristic" displayed in the list. Is selected together with the information as to whether it is an input variable or an output variable, and the "specific Or instead of the input or selection of the force variable type or features and the type or characteristics of the particular output variable. "

  The input task to be solved or the selected task to be solved is the issue registrant such as an identification code that can uniquely identify the task registrant, the name of the task registrant, the contact information of the task registrant, and the date and time of the task registration. Along with the information, the information is additionally stored in an assignment library of the distribution environment providing system 1 which is a storage area (not shown).

  When registering the problem to be solved in the problem registering means 11, the "business type" of the application target 2, such as manufacturing, agriculture, medical treatment, etc., may be registered in association with the problem to be solved.

  If the application target 2 is an example of a vibration prevention control system of a robot installed in a specific environment, the assignment registrant performs the registration operation on the assignment registration unit 11 in order, and the assignment registration unit 11 performs the registration operation. The following are the issues to be solved: "manufacturing industry" as the "business type", "parameter optimization" as the "type of application method for applying the artificial intelligence technology to the applicable system 2," In the data format that can store each of “the amount of vibration of the robot (output value of the acceleration sensor)”, “current value”, “motor speed”, and “motor temperature” as “variable type or feature” of the system 2, It is conceivable that the information is listed and added to the assignment library together with the assignment registrant information.

  In another example, the assignment registrant performs registration operations on the assignment registration unit 11 in order, so that the assignment registration unit 11 sets "artificial intelligence" “Parameter optimization” as “type of application method for applying technology to application target system 2”, “robot vibration amount (output value of acceleration sensor)” as “variable type or characteristic” of application target system 2; “Current value”, “Motor speed”, “Motor temperature”, and “Robot vibration amount” as output variable types in “Specific input variable types or characteristics and specific output variable types or characteristics” of applicable system 2 (The output value of the acceleration sensor) "and the input variable in" the type or characteristic of the specific input variable and the type or characteristic of the specific output variable "of the application target system 2. "Current value" as the type and listed respectively can be stored in a database of data format "motor speed" and "motor temperature", it is conceivable to add storage challenges library with problems registrant information.

<About the artificial intelligence package registration means 12>
Next, the artificial intelligence package registration means 12 will be described. As shown in FIG. 3, the artificial intelligence package registration unit 12 receives a registration operation from the artificial intelligence package registrant with respect to the artificial intelligence package provided by the artificial intelligence package registrant and information on the artificial intelligence package, and It has a function of registering in a storage area (not shown) such as a library.

  Here, the artificial intelligence package is developed and integrated by an artificial intelligence package registrant, and provided to the distribution environment providing system 1, and is obtained by inputting a predetermined learning data set and performing learning. It is possible to generate a learned model applicable to the application target system 2 which is a real system.

  The artificial intelligence package has the necessary means for learning and generating a trained model.

  That is, the artificial intelligence package includes a reading module having a function of reading a data set used when the artificial intelligence module performs learning and inference, a normalization module for normalizing the data set, and a function of dividing a data set of the data set. A data extraction module having a function of extracting desired data from a data set, a data combining module having a function of combining data of a data set, an image data resizing module having a function of resizing image data, An image data filter module having a function of filtering image data, a pre-processing module such as a data labeling module having a function of labeling a data set as normal / abnormal, and a normal / abnormal for an inference result of a learning inference module. The inference result classification module having a function of classifying the determination result, the post-processing module such as the input contribution calculation module having the function of calculating the input contribution to the determination result, when a plurality of internal modules are connected in parallel , A data branching control module that is arranged immediately before a parallel internal module and controls how data is received by the parallel internal module, and each of the internal modules connected in parallel becomes one internal module. When connected, it is arranged immediately after the parallel-connected internal modules, and determines how data is received from each of the parallel-connected internal modules to the one internal module. The data merging control module to be controlled and the internal module connected in parallel When passing data to the one internal module from, awaiting of data output from each of the modules arranged in parallel and arranged immediately after each of the internal modules connected in parallel. A data queuing control module that controls data, a data flow control module such as a data loop control module that feeds back data output from an internal module to the previous stage, AE (self-encoder), CNN (convolutional neural network), LSTM (for deep learning) Principle processing), learning inference modules such as F-CNN (forward sequential neural network), SVM (support vector machine), decision tree learning, and GP (genetic programming) are performed to generate a trained model. So that you can , One learning inference module is specified so as to include at least one learning inference module, or another module is appropriately combined with at least one learning inference module.

  Each of the above-described modules can be realized as a program or an application, and an artificial intelligence package configured by appropriately combining them is also realized as a program or an application.

  In addition, the artificial intelligence package is provided with a data set that is known to be effective for learning and inference, a labeling module that is known to be effective for the data set, and the like, as necessary. You may. In adding each of the dataset and the module for labeling to the artificial intelligence package, the dataset may be stored in a data holding unit in the artificial intelligence package realized as an application, for example, and the module for labeling may be For example, it may be included in the artificial intelligence package as a program.

  An artificial intelligence package composed of an appropriate combination of these various modules generates a plurality of trained model candidates by changing the internal weight parameters and the like by inputting learning data by the action of the learning inference module. I do. After that, a data set for verification is input to a plurality of trained model candidates, a trained model with excellent verification results is selected from the output result, and the selected model is output to the outside as an executable file program By doing so, it is possible to generate a learned model to be applied to an application target that is a real system.

  On the other hand, the learning inference module and the learned model can be classified into, for example, “data mining”, “parameter optimization”, and “clustering” by a method for applying the artificial intelligence technology to the application target. An artificial intelligence package that includes a learning inference module and has a function of generating a learned model can be arranged based on “types of artificial intelligence technology used when the artificial intelligence package learns”.

  In addition, since the learning inference module and the learned model use variables having specific types and characteristics during learning, an artificial intelligence package that includes the learning inference module and has a function of generating a learned model is included in the artificial intelligence package. , Which have information on “the type or characteristic of a specific variable”. For example, for the purpose of optimizing the correlation between “the amount of vibration of the robot (the output value of the acceleration sensor)” and the types of the input variables “current value”, “motor speed”, and “motor temperature” Since the designed artificial intelligence package deals with information on “robot vibration (output value of acceleration sensor)”, “current value”, “motor speed”, and “motor temperature”, (The output value of the acceleration sensor), the “current value”, the “motor speed”, and the “motor temperature” are considered to be information on the artificial intelligence package. In this case, the “vibration amount of the robot (output value of the acceleration sensor)” may be specified as a type of a variable like “output value of the acceleration sensor”, and an acceleration signal for extracting vibration may be specified. Since it has the property of causing a drastic change in a short time, etc., the "output value of the acceleration sensor" for extracting "vibration amount" is specified as a characteristic of a variable like "signal representing vibration" May be. Similarly, the “motor temperature” may be specified as a type of a variable, such as a sensor value of the “motor temperature”, and the time signal of the thermal system has a large time constant of a response value and includes a dead time. Since it has such a characteristic, “motor temperature” may be specified as a feature of a variable like “signal representing temperature”.

  Further, the learning inference module and the trained model include a learning inference module because some models use “the type or characteristic of a specific input variable and the type or characteristic of a specific output variable” in learning. The artificial intelligence package having a function of generating a model includes a package having information on “the type or characteristic of a specific input variable and the type or characteristic of a specific output variable”. For example, optimization is performed on the correlation between the type of output variable “robot vibration amount (output value of acceleration sensor)” and the types of input variables “current value”, “motor speed”, and “motor temperature”. The artificial intelligence package designed for the purpose of performing is to output the type of the output variable "the amount of vibration of the robot (output value of the acceleration sensor)", the type of the input variable "current value", "motor speed" And the correlation between "motor temperature" and the type of output variables, "the amount of vibration of the robot (output value of the acceleration sensor)", and the types of input variables "current value" and "motor". The correlation between "speed" and "motor temperature" is considered to be information about the artificial intelligence package. In this case, the “vibration amount of the robot (output value of the acceleration sensor)” may be specified as a type of a variable like “output value of the acceleration sensor”, and an acceleration signal for extracting vibration may be specified. Since it has the property of causing a drastic change in a short time, etc., the "output value of the acceleration sensor" for extracting "vibration amount" is specified as a characteristic of the variable like "signal representing vibration" May be. Similarly, the “motor temperature” may be specified as a type of a variable, such as a sensor value of the “motor temperature”, and the time signal of the thermal system has a large time constant of a response value and includes a dead time. Since it has such a characteristic, “motor temperature” may be specified as a feature of a variable like “signal representing temperature”.

  Then, the artificial intelligence package registrant registers the artificial intelligence package in the artificial intelligence package registration unit 12 and, as information on the artificial intelligence package, “type of artificial intelligence technology used when the artificial intelligence package learns”, Regarding the "types or characteristics of variables used when the artificial intelligence package learns" or "the types or characteristics of specific input variables and the types or characteristics of specific output variables used when the artificial intelligence package learns", One or more of them are registered in the artificial intelligence package registration unit 12 via a user interface (not shown) of the distribution environment providing system 1.

  Regarding the specific method of registration, the artificial intelligence package may be configured such that the distribution environment providing system 1 uploads a program or an application in the form of an executable file of a program that realizes the artificial intelligence package via a user interface (not shown). Regarding the information on the artificial intelligence package, the distribution environment providing system 1 may present a user interface (not shown) and directly input the information to the registrant of the artificial intelligence package, or may be registered in the artificial intelligence package registration unit 12 in advance. By displaying a list of information on the artificial intelligence package and information on the artificial intelligence package registered in the past, the registrant of the artificial intelligence package may select information on the artificial intelligence package.

  In addition, in the information on the artificial intelligence package registered in the past, “type or feature of variable used when learning by artificial intelligence package” or “type or feature of specific input variable used when learning by artificial intelligence package and identification” `` Types or characteristics of output variables '' from `` types or characteristics of variables '', extract `` types or characteristics of variables '' in any form regardless of input or output, display the list to the artificial intelligence package registrant, and By selecting a desired `` variable type or feature '' from among `` variable types or characteristics used in learning '' and information on the artificial intelligence package, `` variable type or characteristics used by the artificial intelligence package when learning "Characteristic" or select a desired one from "variable types or characteristics" displayed in the list. The type or characteristic of the variable ”is used as an input variable or an output variable, and the selected system is used as the information on the artificial intelligence package. May be replaced with the input or selection of “type or characteristic of specific input variable and type or characteristic of specific output variable used when learning by artificial intelligence package”.

  The information about the input artificial intelligence package or the information about the selected artificial intelligence package is an identification code capable of uniquely identifying the artificial intelligence package registrant, the name of the artificial intelligence package registrant, the contact information of the artificial intelligence package registrant, Along with the artificial intelligence package registrant information such as the registration date of the artificial intelligence package, the information is additionally stored in an artificial intelligence package library, which is a storage area (not shown) of the distribution environment providing system 1.

  Further, when registering the artificial intelligence package and the information on the artificial intelligence package in the artificial intelligence package registration means 12, the “industry” such as manufacturing, agriculture, medical treatment, etc. targeted by the artificial intelligence package is registered and associated. May be.

If the application target 2 is an example of an anti-vibration control system of a robot installed in a specific environment, a program that realizes an artificial intelligence package that generates a trained model for the anti-vibration control system of the robot by an artificial intelligence package registrant By uploading the program or application in the executable file format through the user interface, and performing the registration operation on the artificial intelligence package registration unit 12 in order, the artificial intelligence package registration unit 12 accepts the registration operation. In contrast, information on artificial intelligence packages includes "manufacturing industry" and "robots" as "industry types" and "parameter optimization" and "artificial intelligence" as "types of artificial intelligence technology used when learning artificial intelligence packages." Types or characteristics of variables used by the intelligence package for learning The robot vibration amount (output value of the acceleration sensor), the current value, and the motor speed are listed in a data format that can be stored in the database. Save to library.
In another example, an artificial intelligence package registrant uploads, via a user interface, a program or an application in the form of an executable file of a program for realizing an artificial intelligence package for generating a learned model for a robot anti-vibration control system. In addition, by performing the registration operation on the artificial intelligence package registration unit 12 in order, the artificial intelligence package registration unit 12 sets the “parameter optimization” as “a type of the artificial intelligence technology used when the artificial intelligence package learns”. "The type or characteristics of variables used by the artificial intelligence package for learning" include "the amount of vibration of the robot (output value of the acceleration sensor)", "current value", "motor speed", and "the artificial intelligence package learns". The type or characteristics of a particular input variable As the types of output variables in “types or characteristics of force variables”, “amount of vibration of robot (output value of acceleration sensor)” and “type or characteristics of specific input variables used when learning by the artificial intelligence package and specific output” In the data type that can be stored in the database, "current value", "motor speed", and "motor temperature" are listed as input variable types in the "variable type or characteristics", and the assignment library together with artificial intelligence package registrant information. And save it.

<About the presentation means 13>
Next, the presentation means 13 will be described. As shown in FIG. 4, the presenting unit 13 registers, for the certain problem to be solved specified based on the specifying operation from the designator, the specified certain problem to be solved and the artificial intelligence package registration unit 12. One or more artificial intelligence packages related to the certain problem to be solved are extracted based on information about one or a plurality of artificial intelligence packages, and the extracted one or more artificial intelligence packages related to the specified certain problem to be solved are extracted. A function of presenting one or more artificial intelligence packages, and / or a specific artificial intelligence package or information on a specific artificial intelligence package designated based on a designation operation from a designated person, together with the designated artificial intelligence package and the artificial intelligence package Information on the artificial intelligence package registered in the registration means 13 or a designated artificial intelligence package Based on the information on the information and the one or more tasks to be solved registered by the task registering means 11, the solution related to the specified artificial intelligence package or the information on the specified artificial intelligence package is solved. One or more tasks to be extracted, and a function of presenting one or more tasks to be solved related to the extracted artificial intelligence package or the information on the specified certain artificial intelligence package. .

  Based on the specified problem to be solved among the functions of the presentation unit 13 and the information on one or a plurality of artificial intelligence packages registered by the artificial intelligence package registration unit 12, the specified certain solution is solved. As for a part for extracting one or more artificial intelligence packages related to a task to be solved, “type of application method of artificial intelligence technology to target system” in a specified task to be solved and artificial intelligence package registration means 12 "Type of artificial intelligence technology used when artificial intelligence package learns" in information on one or a plurality of artificial intelligence packages registered by, "type of variable or One or more artificial intelligence packages registered by the "Types or characteristics of variables used by the artificial intelligence package for learning" in the information about the page, and "types or characteristics of specific input variables and specific outputs of the system to be applied" in a specified task to be solved. "Type or characteristic of variable" and "type or characteristic of specific input variable used when learning by artificial intelligence package and specific output variable" in information on one or more artificial intelligence packages registered by artificial intelligence package registration means 12 For each of the "types or characteristics", the matching is performed by a known method, the degree of matching, the degree of relevance, the degree of similarity, etc. are determined to evaluate the degree of matching, and from the evaluation result, a predetermined standard is determined. And an artificial intelligence package related to the specified problem to be solved. Information about one or a plurality of selected, is carried out by extracting the respective AI package registered in the AI package registration unit 12 with the respective information about artificial intelligence packages selected.

At this time, when the collation is performed and the degree of matching is evaluated, the “business type” registered in association with the problem to be solved registered in the problem registration unit 11 and the “industry type” registered in the artificial intelligence package registration unit 12 are registered. The “business type” registered in association with the registered artificial intelligence package may be taken into consideration.
Further, if necessary, a natural language processing technique may be used.

  For example, in the problem registration library, the specified problem to be solved is “manufacturing” as the “industry type” and “parameter optimization” as the “type of application method for applying the artificial intelligence technology to the application target system 2”. , As the “variable type or characteristic” of the application target system 2, the “vibration amount of the robot (output value of the acceleration sensor)”, “current value”, “motor speed”, and “motor temperature”. Also, among the items stored in the artificial intelligence package library, those selected as items to be collated (A) are “manufacturing”, “robot”, and “artificial intelligence used when the artificial intelligence package learns” as the “business type”. The “type of intelligent technology” is “parameter optimization”, and the “types or characteristics of variables used by the artificial intelligence package for learning” are In the case of "moving amount (output value of acceleration sensor)", "current value", and "motor speed", six data of the designated problem to be solved, namely, "manufacturing industry", "parameter optimization", "robot" Of the vibration amount (output value of the acceleration sensor), “current value”, “motor speed”, and “motor temperature” are selected as those to be compared among those stored in the artificial intelligence package library ( A) The six data, namely, “manufacturing industry”, “robot”, “parameter optimization”, “robot vibration amount (output value of acceleration sensor)”, “current value”, and “motor speed” are 5 Since there are two common data, among the ones stored in the artificial intelligence package library, the degree of matching of the one selected as the one to be compared (A) is set to 5/6 * 10 = It can be determined as 83.3%.

  In the same example, among those stored in the artificial intelligence package library, the one selected as the one to be collated (a) is referred to as “a type of artificial intelligence technology used when the artificial intelligence package learns” as “parameter optimization”. If the "type or characteristics of variables used by the artificial intelligence package for learning" are "robot vibration amount (output value of acceleration sensor)", "voltage value", and "motor speed", the specified solution Six data items to be solved, namely “manufacturing industry”, “parameter optimization”, “robot vibration amount (output value of acceleration sensor)”, “current value”, “motor speed”, and “motor temperature” As a criterion, among the data stored in the artificial intelligence package library, four data items (a) selected to be collated, that is, “parameter optimization” , "The amount of vibration of the robot (the output value of the acceleration sensor)", "voltage value", and "motor speed" have three common data. The high degree of matching of the item (a) selected as a kimono can be determined as 3/6 * 100 = 50.0%.

  Here, when obtaining a high degree of matching, the “type of variable” and the “characteristic of the variable” are stored in advance in the presentation unit 13 in a correspondence relationship, and whether or not the two match. May be determined. For example, by registering the “motor temperature” and the “signal representing the temperature” in advance in the presentation unit 13 so as to represent the same one, it may be determined that the two match at the time of comparison.

  Then, among the items stored in the artificial intelligence package library, those selected as items to be compared (A) and (A) are evaluated for the degree of matching, and based on the evaluation result, a predetermined criterion is determined. Satisfaction, that is, for example, the one with the highest matching degree (in this case, only the one (a) selected as the one to be compared among the ones stored in the artificial intelligence package library) or the matching is selected. Those whose degree exceeds a predetermined reference value (40%) (in this case, of the ones stored in the artificial intelligence package library, both (a) and (b) selected as the ones to be collated are selected) Etc., and select one or more pieces of information related to the artificial intelligence package related to the specified task to be solved. It is performed to extract the respective AI package registered in the AI package registration unit 12 with the respective information about artificial intelligence packages.

  At this time, if necessary, the matching may be performed using a natural language processing technique. For example, both the “parameter optimization” and the “parameter optimization” use a known natural language processing technology to perform processing for recognizing that both “parameter” and “optimization” are the main words. By performing, for example, it may be determined that they are the same.

  Further, the method for obtaining the high degree of matching is not limited to the above-described method, and an appropriate method may be used.

  Further, among the functions of the presentation unit 13, information on the artificial intelligence package registered in the artificial intelligence package registration unit 13 together with the designated artificial intelligence package or information on a specified artificial intelligence package, and the assignment registration unit 11 Based on the registered one or a plurality of problems to be solved, a part for extracting one or a plurality of problems to be solved related to the certain artificial intelligence package is registered with the designated artificial intelligence package together with the artificial intelligence package. The “type of artificial intelligence technology used when the artificial intelligence package learns” in the information on the artificial intelligence package registered in the means 13 or the information on the specified artificial intelligence package and one or more registered in the task registration means 11 Applying Artificial Intelligence Technology "Type of application method for applying to elephant system 2", "artificial intelligence" in information on an artificial intelligence package registered in artificial intelligence package registration means 13 together with a designated artificial intelligence package, or information on a designated artificial intelligence package. “Types or characteristics of variables used when package learns” and “variable types or characteristics” of target system 2 in one or more tasks to be solved registered by task registration unit 11, "The type or characteristic of a specific input variable and the type or characteristic of a specific output variable", and information on the artificial intelligence package registered in the artificial intelligence package registration means 13 together with the designated artificial intelligence package or a designated artificial intelligence See “Artificial Intelligence “Type or characteristic of specific input variable and type or characteristic of specific output variable used when cage learns” and “artificial intelligence package learns in one or more tasks to be solved registered by task registering means 11 The type or characteristic of the specific input variable and the type or characteristic of the specific output variable used in this case are collated by a known method, and the degree of matching is determined by obtaining the degree of coincidence, the degree of relevance, the degree of similarity, and the like. The height is evaluated, and a result satisfying a predetermined criterion is determined from the evaluation result, and information on the artificial intelligence package registered in the artificial intelligence package registration unit 13 together with the designated artificial intelligence package or a designated certain artificial intelligence package is registered. Performed by extracting one or more issues to be solved related to the information about the artificial intelligence package .

At this time, when the collation is performed and the degree of matching is evaluated, the “business type” registered in association with the problem to be solved registered in the problem registration unit 11 and the “industry type” registered in the artificial intelligence package registration unit 12 are registered. The “business type” registered in association with the registered artificial intelligence package may be taken into consideration.
Further, if necessary, a natural language processing technique may be used.

  For example, information on the artificial intelligence package registered in the artificial intelligence package registration means 13 together with the designated artificial intelligence package or information on a specified artificial intelligence package is stored in the artificial intelligence registration library as “manufacturing” as “industry type”. , "Robot", "parameter optimization" as "type of artificial intelligence technology used when learning by artificial intelligence package", and "vibration of robot" as "type or characteristic of variable used when learning by artificial intelligence package" Amount (output value of acceleration sensor) "," current value ", and" motor speed ". Of those stored in the task library, the one selected as the one to be compared (c) "Manufacturing" as "industry", "Application to apply artificial intelligence technology to applicable system 2" “Parameter optimization” as the type of method, and “Robot vibration amount (output value of acceleration sensor)”, “Current value”, “Motor speed”, “Motor” as the “variable type or feature” of the applicable system 2 When stored as "temperature", six data of information on the artificial intelligence package registered in the artificial intelligence package registration unit 13 together with the designated artificial intelligence package or information on a specified artificial intelligence package, that is, " Based on “Manufacturing”, “Robot”, “Parameter optimization”, “Robot vibration amount (output value of acceleration sensor)”, “Current value”, and “Motor speed”, collate among those stored in the task library. The six data items (c) selected to be processed, namely, “manufacturing”, “parameter optimization”, and “robot” The vibration amount (output value of the acceleration sensor), the “current value”, the “motor speed”, and the “motor temperature” have five common data. The degree of matching of the selected item (c) may be determined as 5/6 * 100 = 83.3%.

  In the same example, the one selected as a collation target (d) is “parameter optimization” as “type of application method for applying artificial intelligence technology to application target system 2” and “variable of application target system 2” "Type or characteristic of the robot", "vibration amount of robot (output value of acceleration sensor)", "voltage value", "motor speed", "type or characteristic of specific input variable and specific output variable The type of the output variable in the “type or characteristic” is “the amount of vibration of the robot (output value of the acceleration sensor)”, and the input in the “type or characteristic of the specific input variable and the type or characteristic of the specific output variable” of the application target system 2 is If the variable type is stored as “voltage value” or “motor speed”, the artificial intelligence package is Six data of information on the artificial intelligence package registered in the registration means 13 or information on a specified artificial intelligence package, that is, “manufacturing industry”, “robot”, “parameter optimization”, “vibration amount of robot ( The output value of the acceleration sensor), the “current value”, and the “motor speed” are used as a reference. Of those stored in the assignment library, those selected as the ones to be compared among those stored in the assignment library (d) The data, that is, “parameter optimization”, “robot vibration amount (output value of acceleration sensor)”, “voltage value”, “motor speed”, and the type of output variable “robot vibration amount (output value of acceleration sensor) )), The input variable type “voltage value” and the input variable type “motor speed” have three common data. Among those stored in the library, the height of the degree of matching of those selected to be collated (d), may be determined as 3/6 * 100 = 50.0%.

  Here, when obtaining a high degree of matching, the “type of variable” and the “characteristic of the variable” are stored in advance in the presentation unit 13 in a correspondence relationship, and whether or not the two match. May be determined. For example, by registering in advance that the “motor temperature” and the “signal representing the temperature” represent the same thing, it may be determined that they match.

  Then, among the items stored in the assignment library, those selected as items to be compared (C) and (D) are evaluated for the degree of matching, and from the results of the evaluation, those satisfying a predetermined standard That is, for example, the one with the highest matching degree (in this case, only the one (c) selected as the one to be compared among the ones stored in the assignment library) or the degree of matching is a predetermined one Those exceeding the reference value (40%) (in this case, among the ones stored in the assignment library, both (c) and (d) selected to be compared) are determined and designated. Information on the artificial intelligence package registered in the artificial intelligence package registration means 13 together with the registered artificial intelligence package, or a designated artificial intelligence package. Information related to solving problems one or more selected should related, it is performed to extract the respective problems to be solved that are selected.

  At this time, if necessary, the matching may be performed using a natural language processing technique. For example, both of “parameter optimization” and “parameter optimization” perform a process of recognizing that “parameter” and “optimization” are the main words by a known natural language processing technique, or the like. May be collated as being the same.

  Further, the method for obtaining the high degree of matching is not limited to the above-described method, and an appropriate method may be used.

Specifically, for example, the following person can be considered as the designator who operates the presentation unit 13.
(1) A person who belongs to the application target system 2 which is a real system (eg, a person who works for the application target system 2) and who wants to solve a problem to be solved of the application target system 2 by artificial intelligence technology. (2) A person who can develop and integrate an artificial intelligence package, and wants to solve various problems to be solved in various applicable systems 2 using his / her own artificial intelligence package equipped with artificial intelligence technology. Thinker The following describes how (1) and (2) described above operate the presentation unit 13 and use the distribution environment providing system 1 to achieve what effect. I do.

First, the person (1) described above will be described.
A person who belongs to the application target system 2 which is the actual system (1) (a person who works for the application target system 2) and wants to solve the problem to be solved which the application target system 2 has by using artificial intelligence technology. The user can perform the registration operation on the problem registration unit 11 for the problem to be solved in advance, as described above.

  Thereafter, a certain task, for example, a task to be solved, which has been registered in the task registering means 11 by itself, is specified to the presenting means 13 by performing the specifying operation as described above. Then, the presentation unit 13 determines the “type of application method of the artificial intelligence technology to the application target system”, the “variable type or characteristic” of the application target system 2, and the “identification of the application target system” in the designated task to be solved. Or the type or characteristic of the input variable and the type or characteristic of the specific output variable, and the “artificial intelligence package” in the information on the one or more artificial intelligence packages registered by the artificial intelligence package registration unit 12 so far. Type of artificial intelligence technology used when learning by AI "," Types or characteristics of variables used by artificial intelligence package to learn "," Types or characteristics of specific input variables used by artificial intelligence package to learn and specification " Or more than one of "types or characteristics of output variables of For, as described above, by known techniques, collating, matching degree, relevance and, seeking the similarity or the like to assess the height of the degree of matching. Then, from the results of the evaluation, one that satisfies a predetermined criterion is determined, and based on the determination result, one or more pieces of information on the artificial intelligence package related to the specified task to be solved are selected and selected. Each of the artificial intelligence packages registered in the artificial intelligence package registration means 12 is extracted together with the information on the registered artificial intelligence package.

  Then, the presenting unit 13 sends, to the person (1) described above, the artificial intelligence package selected as described above or the information on the selected artificial intelligence package as described above, the user interface (not shown) of the distribution environment providing system 1. To be presented via

  Then, the presenting unit 13 confirms whether or not the above-mentioned person (1) selects and uses the artificial intelligence package presented via the user interface, and selects a specific artificial intelligence package. When the answer to use is obtained, the above-mentioned person (1) is caused to download the program or the learned model of the executable file format of the selected artificial intelligence package.

  Further, the presentation means 13 provides the above-mentioned person (1) with the artificial intelligence package registrant information stored in the artificial intelligence package library together with the information on the artificial intelligence package selected as described above. Thus, information such as the name and contact information of the registrant of the artificial intelligence package that has registered the information on the selected artificial intelligence package can be presented.

  Accordingly, the above-mentioned person (1) uses the distribution environment providing system 1 to provide an artificial intelligence package that is expected to be solved by the artificial intelligence technology for the problem to be solved by the application target system 2 of the user himself / herself. And easy to get.

  In particular, regarding the problem to be solved, the “type of application method of artificial intelligence technology to the applicable system”, the “variable type or characteristic” of the applicable system 2, the “type of specific input variable of the applicable system or As a characteristic and a type or characteristic of a specific output variable, the artificial intelligence package is registered on the distribution environment providing system 1 and, regarding the artificial intelligence package, “type of artificial intelligence technology used when the artificial intelligence package learns”, “artificial intelligence package” On the distribution environment providing system 1 as “types or characteristics of variables used when learning” or “types or characteristics of specific input variables and specific output variables used when the artificial intelligence package learns”. By registering, the problem to be solved and the artificial intelligence package are described as "type of application method of artificial intelligence technology" Since the matching can be performed by comparing “the type or characteristic of a specific variable” or “the type or characteristic of a specific input variable and the type or characteristic of a specific output variable”, the distribution environment providing system 1 should solve the problem. In order to solve the problem, it is possible to efficiently obtain a more effective artificial intelligence package.

Further, even if one or a plurality of artificial intelligence packages presented by the presenting unit 13 cannot completely match the problem to be solved, the presenting unit 13 sends the above-mentioned person (1) to the above-mentioned person (1). Name of the artificial intelligence package registrant who registered the information on the selected artificial intelligence package based on the artificial intelligence package registrant information registered in the artificial intelligence package library with each of the information on the artificial intelligence package selected as described above. Since it is possible to present information such as information and contact information, for example, the presented artificial intelligence package registrant can be given an opportunity to individually consult about a problem to be solved.
Then, it is possible to register a new problem to be solved, which has become clear as a result of individual consultation, separately in the problem registering means 11 and repeat the above-described processing, thereby performing matching with a new artificial intelligence package. Become.

Next, the person (2) described above will be described.
A person who can develop and integrate the artificial intelligence package of (2), and wants to solve the problems to be solved in various applicable systems 2 and the like using his own artificial intelligence package including the artificial intelligence technology. The person who thinks that he or she can perform the registration operation on the artificial intelligence package registration means 12 as described above with respect to the own artificial intelligence package.

  Thereafter, the information about the artificial intelligence package registered in the artificial intelligence package registration means 12 together with a certain artificial intelligence package or a certain artificial intelligence package, for example, the artificial intelligence package is registered in the artificial intelligence package registration means 12 with the presentation means 13. The object is specified by performing the specifying operation as described above. Then, the presentation unit 13 reads “the artificial intelligence technology used when the artificial intelligence package learns” in the information about the designated artificial intelligence package or the artificial intelligence package registered in the artificial intelligence package registration unit 12 together with the designated artificial intelligence package. Type, the types or characteristics of variables used when the artificial intelligence package learns, and the types or characteristics of specific input variables and specific output variables used when the artificial intelligence package learns. One or more of these, and one or more tasks to be solved registered by the task registration means 11 so far, “type of application method of artificial intelligence technology to applicable system”, “variable type of applicable system 2” Or characteristics "," types or characteristics of specific input variables and specific One or more of the types or characteristics of force variables ", for which the information to be collated is complete, is collated by a known method as described above, and the degree of coincidence, relatedness, similarity, etc. is determined. Then, the degree of matching is evaluated. Then, from the results of the evaluation, those that satisfy a predetermined criterion are determined. Based on the determination result, the designated artificial intelligence package or the artificial intelligence registered in the artificial intelligence package registration means 12 together with the designated artificial intelligence package is determined. One or more issues to be solved related to the information on the intelligent package are selected, and each of the selected issues to be solved is extracted.

  Then, the presentation unit 13 presents each of the problems to be solved extracted as described above to the person (2) above through a user interface (not shown) of the distribution environment providing system 1.

  Then, the presenting unit 13 selects the problem to be solved presented via the user interface for the above-mentioned person (2), and confirms whether or not to attempt to apply the artificial intelligence package designated by himself / herself. If a specific problem to be solved is selected and an answer that the application is to be obtained is obtained, the above-mentioned person (2) is registered in the problem library together with the selected specific problem to be solved. Information such as the name and contact information of the registrant who registered the selected specific problem to be solved can be presented based on the registered registrant information.

  Thus, the above-mentioned person (2) can easily obtain, using the distribution environment providing system 1, a problem to be solved that is expected to function effectively for his or her own artificial intelligence package.

  In particular, regarding the problem to be solved, the “type of application method of artificial intelligence technology to the applicable system”, the “variable type or characteristic” of the applicable system 2, the “type of specific input variable of the applicable system or As a characteristic and a type or characteristic of a specific output variable, the artificial intelligence package is registered on the distribution environment providing system 1 and, regarding the artificial intelligence package, “type of artificial intelligence technology used when the artificial intelligence package learns”, “artificial intelligence package” On the distribution environment providing system 1 as “types or characteristics of variables used when learning” or “types or characteristics of specific input variables and specific output variables used when the artificial intelligence package learns”. By registering, the problem to be solved and the artificial intelligence package are described as "type of application method of artificial intelligence technology" Since the matching can be performed by “type or characteristic of specific variable” or “type or characteristic of specific input variable and type or characteristic of specific output variable”, the distribution environment providing system 1 Thus, it is possible to efficiently obtain a problem to be solved that can be expected to be more effective.

Further, even if one or a plurality of problems to be solved presented by the presenting means 13 cannot be completely matched with the artificial intelligence package, the presenting means 13 can provide the above-mentioned (2) to the above-mentioned person. Based on the problem registrant information registered in the problem library along with each of the problems to be solved as described above, the name and contact information of the problem registrant who registered the selected problem to be solved are presented. Therefore, for example, the presented registrant can be given an opportunity to sell his or her own artificial intelligence package as a separate business.
Then, a new artificial intelligence package acquired through the individual business is separately registered in the artificial intelligence package registration means 12, and by repeating the above processing, it is possible to perform matching with a new problem to be solved. Become.

  Here, in the above description, the problem registration unit 11 determines, as the problem to be solved associated with the application target system 2 to which the artificial intelligence package is to be applied, “the method of applying the artificial intelligence technology to the application target system 2”. One or more of “type”, “type or characteristic of specific variable” of applicable system 2, and “type or characteristic of specific input variable and specific type or characteristic of output variable” of applicable system 2 Although the distribution environment providing system 1 has a function of accepting and registering the information, the problem registration means 11 determines that the problem registration means 11 applies the problem as a problem to be solved associated with the application target system 2 to which the artificial intelligence package is applied. Only "type of application method of artificial intelligence technology" to target system 2, only "type or characteristic of specific variable" of target system 2, or The application target system 2 may have a function of accepting and registering only “the type or characteristic of a specific input variable and the type or characteristic of a specific output variable”, and associates the function with the application target system 2 to which the artificial intelligence package is applied. As a problem to be solved, the system may have a function of receiving and registering “type of application method of artificial intelligence technology” to application target system 2 and “type or characteristic of specific variable” of application target system 2. As a problem to be solved associated with the application target system 2 to which the artificial intelligence package is applied, “type of application method of the artificial intelligence technology” to the application target system 2 and “specific input variables of the application target system 2” Has the function of accepting and registering the type or characteristic of the Often, the "business type" of the applicable system 2, the "type of application method of the artificial intelligence technology" to the applicable system 2, the "type or characteristic of a specific variable" of the applicable system 2, or the "applicable system 2" And a function of accepting and registering only one of “type or characteristic of specific input variable and type or characteristic of specific output variable”.

  Similarly, in the above description, the artificial intelligence package registration unit 12 outputs the information on the artificial intelligence package together with the artificial intelligence package having the function of generating the learned model as “artificial intelligence technology used when the artificial intelligence package learns. Type "," Types or characteristics of specific variables used when learning by the artificial intelligence package "," Types or characteristics of specific input variables used when learning by the artificial intelligence package "and types or characteristics of specific output variables Of the distribution environment providing system 1, the artificial intelligence package registration unit 12 determines that the artificial intelligence package learns as information on the artificial intelligence package. Types of artificial intelligence technology used when learning "," The artificial intelligence package used when learning It may have a function of accepting and registering only "types or characteristics of specific variables" or "types or characteristics of specific input variables and specific output variables used when the artificial intelligence package learns". , A function that accepts and registers “types of artificial intelligence technology used when artificial intelligence packages learn” and “types or characteristics of specific variables used when artificial intelligence packages learn” as information about artificial intelligence packages. The information on the artificial intelligence package may include `` types of artificial intelligence technology used when the artificial intelligence package learns '' and `` types or characteristics of specific input variables used when the artificial intelligence package learns and specific With the function of accepting and registering "types or characteristics of output variables" Often, as information on the artificial intelligence package, the "business type" of the applicable system 2, the "type of the artificial intelligence technology used when the artificial intelligence package learns", and the "specific variables used when the artificial intelligence package learns" Even if it has a function of accepting and registering only one of “type or characteristic” and “type or characteristic of specific input variable and specific output variable used when learning by the artificial intelligence package”. Good.

<About the modification>
Further, a modified example of the distribution environment providing system 1 will be described below.
In this modification, the configuration described using the same name and the same reference numeral has the same function as the configuration described above, and the description thereof will not be repeated.

  In this modification, as shown in FIG. 5, the distribution environment providing system 1 further includes a designation device 3.

  The designation device 3 is used by the designation device 3 to designate a certain problem to be solved registered in the problem registration device 11 or to be registered in the artificial intelligence package registration device 12 by the designation device 3. A designation operation for designating information relating to a certain artificial intelligence package or an artificial intelligence package registered in the artificial intelligence package registration means 12 together with the artificial intelligence package may be performed at an appropriate timing for an appropriate target.

  Here, the appropriate timing is a timing appropriately selected by the administrator of the distribution environment providing system 1 or, for example, once every hour in the daytime (for example, from 9:00 am to 5:00 pm), once every minute. , Etc., represents an appropriate period. Further, the appropriate target is registered in the assignment registering unit 11 or the artificial intelligence package registering unit 12 based on a predetermined condition, for example, by the designation of the administrator of the distribution environment providing system 1 or at random. One or a plurality of items extracted from the information, the information to be solved, and information on the artificial intelligence package and the artificial intelligence package registered in the artificial intelligence package registration unit 12 together with the artificial intelligence package.

  Thereby, the presenting unit 13 is operated by the designation operation from the designation device 3 as described above, and for example, identification of one or a plurality of problems to be solved registered in the problem registration unit 11 by the designation device 3 is performed. Is automatically specified, and one or more artificial intelligence packages having a high degree of matching with the specified problem to be solved are automatically extracted. At this time, the presenting unit 13 has a function of automatically sending an e-mail or the like with reference to the information of the artificial intelligence package registrant registered in the artificial intelligence package registering unit 12 together with the automatically extracted artificial intelligence packages. By preparing, the corresponding AI package registrant will receive a message such as "I found an issue that matches the AI package you registered with a high degree" and information on the specified issue to be solved In other words, “type of application method of artificial intelligence technology to the applicable system” for the specified problem to be solved, “variable type or characteristic” of applicable system 2, “specific input variables of the applicable system” And the types and characteristics of specific output variables "and the information of the project registrant.

  In addition, by causing the presenting means 13 to operate as described above by the designation operation from the designation device 3, the designation device 3 is used to generate one or more artificial intelligence packages or artificial intelligence packages registered in the artificial intelligence package registration means 12 together with the artificial intelligence package. Information relating to one or a plurality of artificial intelligence packages registered in the intelligent package registration means 12 is designated, and a problem to be solved having a high degree of matching with the designated one is automatically extracted. At this time, the presenting means 13 is provided with a function of automatically sending an e-mail or the like by referring to the information of the issue registrant registered in the issue registering means 11 together with each of the automatically extracted issues to be solved. In response to the subject registrant, a message such as "An artificial intelligence package with a high degree of matching with the task you registered was found" is sent to one or more designated artificial intelligence packages or Information on one or more artificial intelligence packages registered in the artificial intelligence package registration means 12 together with the designated artificial intelligence package, that is, one or more artificial intelligence registered in the artificial intelligence package registration means 12 together with the designated artificial intelligence package. "Types of Artificial Intelligence Technology Used When Artificial Intelligence Packages Learn" for information on intelligence packages "Types or characteristics of variables used when learning by artificial intelligence package", "Types or characteristics of specific input variables and types or characteristics of specific output variables used when learning by artificial intelligence package", artificial intelligence package Can be contacted with registrant information.

  By using the distribution environment providing system 1 shown in this modified example, the problem registrant or the artificial intelligence package registrant can register only the problem to be solved or his or her own artificial intelligence package in the distribution environment providing system. It is possible to automatically obtain information about each of the AI packages and issues to be solved that have a high degree of matching with each of the issues to be solved and each of the AI packages. It is possible to further improve the convenience of the registrant of the intelligent package.

  The present invention is not limited to the above-described embodiment, and can be implemented with various modifications without departing from the spirit of the present invention. And, they are all included in the technical idea of the present invention.

DESCRIPTION OF SYMBOLS 1 Distribution environment provision system 11 Issue registration means 12 Artificial intelligence package registration means 13 Presentation means 2 Applicable system 3 Designation device

Claims (4)

In a distribution environment providing system that provides a distribution environment of an artificial intelligence package for solving the problem using the artificial intelligence technology,
The distribution environment providing system includes a task registration unit, an artificial intelligence package registration unit, and a presentation unit,
The problem registration means may include, as a problem to be solved associated with an application target system to which the artificial intelligence package is applied, a type of an application method of the artificial intelligence technology to the application target system, a specific variable of the application target system. Has a function of receiving and registering one or more of the type or characteristic of a specific input variable and the type or characteristic of a specific output variable of the application target system,
The artificial intelligence package registration means, together with an artificial intelligence package having a function of generating a learned model, as information on the artificial intelligence package, a type of artificial intelligence technology used when the artificial intelligence package learns, the artificial intelligence package Accepts and registers one or more of a specific variable type or characteristic used for learning, a specific input variable type or characteristic used for learning by the artificial intelligence package, and a specific output variable type or characteristic used for learning. Has the function of
The presenting means receives designation of a certain task to be solved registered in the task registering means, and specifies the designated certain task to be solved and one or more of the artificial subjects registered in the artificial intelligence package registering means. A function of presenting one or a plurality of artificial intelligence packages related to the specified problem to be solved based on information about the intelligent package, and a certain artificial intelligence package registered in the artificial intelligence package registration means or The specification of the information on the artificial intelligence package is received, and the information on the certain artificial intelligence package registered with the specified certain artificial intelligence package or the information on the specified certain artificial intelligence package is registered by the assignment registering means. Based on one or more of the problems to be solved, It said certain distribution environment providing system for providing the distribution system of artificial intelligence packages characterized by having one or more presenting functions problem to be solved related to the information about artificial intelligence packages AI packaged or specified that. In a distribution environment providing system that provides a distribution environment of an artificial intelligence package for solving the problem using the artificial intelligence technology,
The distribution environment providing system includes a task registration unit, an artificial intelligence package registration unit, and a presentation unit,
The problem registration means may include, as a problem to be solved associated with an application target system to which the artificial intelligence package is applied, a type of an application method of the artificial intelligence technology to the application target system, a specific variable of the application target system. Has a function of receiving and registering one or more of the type or characteristic of a specific input variable and the type or characteristic of a specific output variable of the application target system,
The artificial intelligence package registration means, together with an artificial intelligence package having a function of generating a learned model, as information on the artificial intelligence package, a type of artificial intelligence technology used when the artificial intelligence package learns, the artificial intelligence package Accepts and registers one or more of a specific variable type or characteristic used for learning, a specific input variable type or characteristic used for learning by the artificial intelligence package, and a specific output variable type or characteristic used for learning. Has the function of
The presenting means receives designation of a certain task to be solved registered in the task registering means, and specifies the designated certain task to be solved and one or more of the artificial subjects registered in the artificial intelligence package registering means. A distribution environment providing a distribution environment of an artificial intelligence package having a function of presenting one or a plurality of artificial intelligence packages related to the specified problem to be solved based on information on the intelligent package. system. In a distribution environment providing system that provides a distribution environment of an artificial intelligence package for solving the problem using the artificial intelligence technology,
The distribution environment providing system includes a task registration unit, an artificial intelligence package registration unit, and a presentation unit,
The problem registration means may include, as a problem to be solved associated with an application target system to which the artificial intelligence package is applied, a type of an application method of the artificial intelligence technology to the application target system, a specific variable of the application target system. Has a function of receiving and registering one or more of the type or characteristic of a specific input variable and the type or characteristic of a specific output variable of the application target system,
The artificial intelligence package registration means, together with an artificial intelligence package having a function of generating a learned model, as information on the artificial intelligence package, a type of artificial intelligence technology used when the artificial intelligence package learns, the artificial intelligence package Accepts and registers one or more of a specific variable type or characteristic used for learning, a specific input variable type or characteristic used for learning by the artificial intelligence package, and a specific output variable type or characteristic used for learning. Has the function of
The presenting unit receives designation of a certain artificial intelligence package or information related to a certain artificial intelligence package registered in the artificial intelligence package registration unit, and information about the certain artificial intelligence package registered with the designated certain artificial intelligence package. Alternatively, the specified artificial intelligence package or the specified specific artificial package is determined based on the information on the specified specific artificial intelligence package and one or more of the problems to be solved registered by the problem registering means. A distribution environment providing system for providing a distribution environment of an artificial intelligence package, which has a function of presenting one or more problems to be solved related to information on an intelligent package. In the distribution environment providing system according to any one of claims 1 to 3,
The distribution environment providing system further provides the presenting means 13 with the task to be solved registered in the task registering means or the artificial intelligence package registering means selected under predetermined conditions at a predetermined timing. A distribution environment providing system for providing a distribution environment of an artificial intelligence package, comprising a designated device for performing a designated operation on a registered artificial intelligence package or information on the artificial intelligence package.