KR102185359B1 - Service restoration device linked with the change in service usage by menu chage - Google Patents

Abstract

The apparatus for implementing an artificial intelligence-based service and operating a system according to an embodiment of the present invention includes a storage unit that stores commands and data required to perform artificial intelligence-based data analysis, and user data for a service app or web from a client. And a communication unit that receives service item data, calculates a menu item with a high probability of selection by the user based on the user data and service item data, requests the operator to change the calculated menu setting value, or changes the setting value Thus, it may include a control unit for reprogramming the service app or web.

Description

Service restoration device linked with the change in service usage by menu chage}

The present invention relates to an artificial intelligence-based service implementation and system operation apparatus.

With the proliferation of the digital economy, a'big data environment' has arrived around us in which a lot of information and data are produced that cannot be determined. Compared to data generated in an analog environment in the past, big data has a large scale, a short generation period, and includes character and image data as well as numerical values.

Machine learning is a representative technique for effectively using such big data. Machine learning is a technology that supports learning by assigning data to computers as humans learn, and gaining new knowledge through it.

However, currently, analysis methods using artificial intelligence are widely used, but reprogramming using information calculated through artificial intelligence techniques is not immediate, and there is an inconvenience that requires operator intervention.

An object of the present invention is to automatically change and provide the display content of a service providing environment according to the result of analysis using artificial intelligence.

Specifically, an object of the present invention is to establish a back end for controlling various services provided from a front end.

The apparatus for implementing an artificial intelligence-based service and operating a system according to an embodiment of the present invention includes a storage unit that stores commands and data required to perform artificial intelligence-based data analysis, and user data for a service app or web from a client. And a communication unit that receives service item data, calculates a menu item with a high probability of selection by the user based on the user data and service item data, requests the operator to change the calculated menu setting value, or changes the setting value Thus, it may include a control unit for reprogramming the service app or web.

The present invention can perform reprogramming by applying an artificial intelligence model suitable for each field of service.

The present invention extracts a menu item determined to have a high probability of selecting or browsing based on the user's usage data as a design variable for reprogramming, and performing a design change operation to highlight it, so that the user's selection or purchase There is an effect to induce.

In addition, since the present invention automatically reprograms the app or the web according to the result value calculated based on artificial intelligence analysis, it is possible to reduce the workload of the programming operator in a simpler way.

1 is an exemplary diagram schematically illustrating a platform according to an embodiment of the present invention.
2 is a block diagram showing a configuration of an apparatus for implementing an artificial intelligence-based service and operating a system according to an embodiment of the present invention.
3 is a block diagram showing the configuration of a data collection unit according to an embodiment of the present invention.
4 is a diagram illustrating a configuration of a data analysis unit according to an embodiment of the present invention.
5 is a diagram showing the configuration of an evaluation unit according to an embodiment of the present invention.
6 is a flowchart illustrating a sequence of a programming support operation according to an embodiment of the present invention.

In the present invention, various modifications may be made and various embodiments may be provided. Specific embodiments will be illustrated in the drawings and described in detail.

However, this is not intended to limit the present invention to a specific embodiment, it is to be understood to include all changes, equivalents, and substitutes included in the spirit and scope of the present invention. In describing each drawing, similar reference numerals have been used for similar elements.

When an element is referred to as being'connected' or'connected' to another element, it is understood that it may be directly connected or connected to the other element, but other elements may exist in the middle. Should be. On the other hand, when a component is referred to as being'directly connected' or'directly connected' to another component, it should be understood that there is no other component in the middle.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present application, terms such as'comprise' or'have' are intended to designate the presence of features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, but one or more other features It is to be understood that the presence or addition of elements or numbers, steps, actions, components, parts, or combinations thereof, does not preclude in advance.

Meanwhile, the method of implementing an artificial intelligence-based service and supporting programming of a system operating apparatus according to the present invention can be applied not only to web services but also to application services. Such application services or web services may be services performed in various electronic devices. At this time, the electronic device is a smartphone, a tablet PC (tablet personal computer), a mobile phone (mobile phone), a video phone, an e-book reader (e-book reader), a desktop PC (desktop PC), a laptop PC (laptop PC). PC), a netbook computer, a workstation, a server, a personal digital assistant (PDA), a portable multimedia player (PMP), an MP3 player, a mobile medical device, a camera, or a wearable device. It can contain one. That is, the application service or web service executed in various electronic devices described above may be reprogrammed according to various embodiments of the present invention.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings.

1 is an exemplary diagram schematically illustrating a platform according to an embodiment of the present invention.

As shown in Figure 1, the platform connects the client and the worker, and can provide various services in the middle. In one embodiment, the platform may receive a request for data work from a client, and match the workers corresponding to the request to allocate the data work of the client to the worker. In the process of allocating data tasks, the platform can provide an optimized UX/UI tool to workers using an artificial intelligence model.

In another embodiment, the platform may build a system to increase the security of data transmitted between a client and an operator. For example, the platform may encrypt data provided by the client and provide it to the worker, and encrypt the work created by the worker and provide it to the client.

In addition to this, the platform can use an artificial intelligence model to inspect the work created by the worker and request rework from the worker or deliver it to the client.

2 is a block diagram showing a configuration of an apparatus for implementing an artificial intelligence-based service and operating a system according to an embodiment of the present invention.

The device 100 according to an embodiment of the present invention may be implemented in the form of a server. In addition, the device 100 may include a communication unit 110, a storage unit 120, and a control unit 130 as illustrated in FIG. 2. The control unit 130 may include a data collection unit 131, a data analysis unit 132, and an evaluation unit 133.

First, the communication unit 110 may use a network to transmit and receive data between a user device and a server, and the type of the network is not particularly limited. The network may be, for example, an Internet Protocol (IP) network that provides a service for transmitting and receiving large amounts of data through an Internet Protocol (IP) or an All IP network integrating different IP networks. In addition, the network includes a wired network, a wireless broadband (Wibro) network, a mobile communication network including WCDMA, a high speed downlink packet access (HSDPA) network and a mobile communication network including a long term evolution (LTE) network, and LTE advanced (LTE-A). ), a mobile communication network including 5G (Five Generation), a satellite communication network, and a Wi-Fi network, or at least one or more of them may be combined.

The communication unit 110 according to an embodiment of the present invention may communicate with a device of a client or a device of a worker. Specifically, the communication unit 110 receives user data (user's questionnaire information, personal information, user's search word, etc.) related to a service provided through an application (hereinafter referred to as an application) or a website (hereinafter referred to as a web) from the client's device (server). Can receive. In addition, the communication unit 110 may transmit the source code in which the set value of at least one menu is changed to the operator's device according to the result of data analysis. According to various embodiments of the present disclosure, the communication unit 110 may transmit a list of menu items that are requested to be changed according to a result of data analysis to a worker's device.

In addition, the communication unit 110 may communicate with other web servers or one device for various purposes.

The storage unit 120 may include, for example, an internal memory or an external memory. The internal memory includes, for example, volatile memory (e.g., dynamic RAM (DRAM), static RAM (SRAM), synchronous dynamic RAM (SDRAM), etc.)), non-volatile memory (e.g., OTPROM (one time programmable ROM), programmable ROM (PROM), erasable and programmable ROM (EPROM), electrically erasable and programmable ROM (EEPROM), mask ROM, flash ROM, flash memory (such as NAND flash or NOR flash), hard drive, Alternatively, it may include at least one of a solid state drive (SSD).

External memory is a flash drive, for example, compact flash (CF), secure digital (SD), micro secure digital (Micro-SD), mini secure digital (mini-SD), extreme digital (XD), It may further include a multi-media card (MMC) or a memory stick. The external memory may be functionally and/or physically connected to the electronic device through various interfaces.

The storage unit 120 according to an embodiment of the present invention may store user data and service item data received from a client side or an operator side. In the case where the artificial intelligence-based service implementation and system operation device 100 is the same as the client server, the storage unit 120 may be set to collect related data in response to a service web or a backend of the app, and accordingly, the storage unit 120 User data and service item data can be collected and stored.

The storage unit 120 may also store commands and data required to perform artificial intelligence-based data analysis.

In addition, the storage unit 120 may store commands and data required for overall operations according to an embodiment of the present invention.

The controller 130 may also be referred to as a processor, a controller, a microcontroller, a microprocessor, a microcomputer, or the like. Meanwhile, the control unit may be implemented by hardware, firmware, software, or a combination thereof.

In the case of implementation by firmware or software, an embodiment of the present invention may be implemented in the form of a module, procedure, or function that performs the functions or operations described above. The software code may be stored in a memory and driven by the control unit. The memory may be located inside or outside the user terminal and the server, and may exchange data with the controller through various known means.

The controller 130 according to an embodiment of the present invention may perform a role of calculating a variable required for designing a backend and/or a frontend based on artificial intelligence in web or app programming. Furthermore, the controller 130 may calculate a variable required for backend and/or frontend design and then apply the calculated variable to change the design. Specifically, the control unit 130 may calculate a menu item that is highly likely to be selected or interested by the user through data analysis based on artificial intelligence, and transmits a request to change the calculated menu setting value to the operator, or You can perform an operation of reprogramming by changing.

The control unit 130 may include a data collection unit 131, a data analysis unit 132, and an evaluation unit 133.

The data collection unit 131 may collect data required to calculate a menu item (design variable) with a high probability to be selected by a user. In this case, the data collection unit 131 requests editing of programming according to an embodiment of the present invention. It is possible to perform an operation of collecting data from a client server. In this case, the requester server may collectively refer to a server that provides a service through, for example, an app or the web, and may transmit predetermined user data and service-related data to the interworked operating device 100. According to an embodiment, the requester server may include a server of a portal web site that provides services such as shopping and search.

According to various embodiments, the requester server may be configured in the same device as the device 100 for implementing an artificial intelligence-based service and system operation, and based on artificial intelligence based on user response information collected while providing daily services through a web or an app. Data analysis can also be performed.

The data collection unit 131 may include a user data collection unit 131a, a service item data collection unit 131b, and a developer data collection unit 131c. A more detailed description of the data collection unit 131 will be described with reference to FIG. 3.

The data analysis unit 132 may perform an operation for calculating a programming design variable based on data collected by the data collection unit 131. Specifically, the data analysis unit 132 performs an operation of extracting data to be input for calculating a design variable during reprogramming and an operation of calculating a result value by putting the extracted data items into a specific artificial intelligence algorithm model. I can. In this case, the artificial intelligence algorithm may be configured such that a result value calculated from an output layer as a result of execution includes a type of language related to programming, an item on an interface to designate an emphasis effect, and the like. However, the present invention is not limited thereto, and more various items of programming design parameters may be designated as a result of the output layer calculation.

The data analysis unit 132 may include a data input unit 132a, a model selection unit 132b, and a design variable calculation unit 132c. A detailed description of the data analysis unit 132 will be described with reference to FIG. 4.

The evaluation unit 133 may perform an operation of evaluating a programming result to which an output calculated as a result of the operation of the data analysis unit 132 is applied. In this case, the method of evaluation may be to evaluate the performance (eg, change in usage rate, change in the number of payments, etc.) according to the application of the change. The evaluation unit may be configured to include a change amount measurement unit 133a. A detailed description of the evaluation unit 133 will be described with reference to FIG. 5.

3 is a block diagram showing the configuration of a data collection unit according to an embodiment of the present invention.

As shown in FIG. 3, the data collection unit 131 according to an embodiment of the present invention may include a user data collection unit 131a, a service item data collection unit 131b, and a developer data collection unit 131c. The data collected by the data collection unit 131 may refer to data received through the communication unit 110 from a separate server interlocked with each other, or may refer to data in the device 100 stored in the storage unit 120.

The user data collection unit 131a may collect data related to a user who uses a service. In one embodiment, the user using the service may be a worker-side user. The user data collection unit 131a may collect information related to a user's response to a service item to be provided. Information related to the user's response may be collected by classifying static information and dynamic information according to an embodiment.

For example, the user data collection unit 131a may collect information such as a user's main choice for a service item and questionnaire information previously input from the user. In this case, the static information may include personal information of the user (eg, gender, year of birth, residential area, etc.), information directly input by the user, such as questionnaire information, or information that does not change after one-time setting. On the other hand, dynamic information among user data may include information related to a user's choice. For example, the user's main click menu information, the user's main complaint, the user's search history, and the like may be included in the dynamic information item.

The service item data collection unit 131b may collect information related to a service provided to a user. For example, in the case of a shopping mall app, the service item data may include information related to each product. In addition, the service item data may include information on menu items (eg, best items, new products, sale items, language change, etc.) displayed on the web or app. Also, the service item data may be collected separately for static information and dynamic information. For example, the static information of the service item data may mean information that does not change according to an artificial intelligence-based data analysis result according to an embodiment of the present invention. Likewise, dynamic information may mean information that is variable according to the result of data analysis based on artificial intelligence.

In addition, the service item data collection unit 131b may collect information related to a service field. In this case, the service item data collection unit 131b may collect information related to the service field through not only the server providing the service but also other external web servers. Alternatively, the service item data collection unit 131b may classify a service item based on previously stored information, and calculate and apply a service field characteristic value corresponding to the item. In this case, the service item data collection unit 131b may determine a service field based on text, images, and menu items constituting the app or web. In addition, the service field determination operation may be performed based on a preset service surrender classification method.

According to various embodiments of the present disclosure, the data collection unit 131 may include a developer data collection unit 131c. The developer data collection unit 131c may collect information on items set by the developer at the time of initial programming. According to various embodiments of the present disclosure, information may be separately stored at the time of programming in a specific source code that can be changed based on an artificial intelligence analysis result by an initial developer. Accordingly, the developer data collection unit 131c may collect source code information separately designated as a variable region at the time of programming.

According to various embodiments, the developer data collection unit 131c may receive source code information for a specific page from a client or operator, extract items set as numerical values from them as changeable items, and collect the extracted information as developer data. I can. For example, the developer data collection unit 131c is numerical information from code data such as "<iframe width="850" height="500" src=//www.youtove.com/embed/~></iframe>" 850" and "500" can be extracted as variable items. According to various embodiments of the present disclosure, the developer data collection unit 131c may transmit a signal inquiring whether to set the data extracted as a variable item as a variable item to a worker side and obtain approval information from the worker side.

According to various embodiments of the present disclosure, the developer data collection unit 131c may read the entire source code of an app or web and perform an operation of extracting a variable item therefrom.

* Figure 4 is a diagram showing the configuration of a data analysis unit according to an embodiment of the present invention.

The data analysis unit 132 may calculate an item to be changed among programming contents and a change contents for the collected data based on artificial intelligence. At this time, the calculated content can be delivered to the operator or the changed content is automatically applied to support compilation and service.

As illustrated in FIG. 4, the data analysis unit 132 according to an embodiment of the present invention may include a data input unit 132a, a model selection unit 132b, a set classification unit 133c, and a design variable calculation unit 133d.

The data input unit 132a may perform an operation of injecting information required for performing data analysis based on artificial intelligence into an artificial intelligence algorithm as a variable. The data input unit 132a may extract data to be input from among information collected by the data collection unit 131. In addition, the data input unit 132a is input to an artificial intelligence algorithm based on service item information, which is information collected by the data collection unit 131, or a calculated service field characteristic value (e.g., a data item calculation standard to be input) according to various embodiments. You can extract data items to do. For example, the data input unit 132a is a user data item corresponding to the'News' service field, and is the user's gender, age, user's search word, main selected media company, and sequence data (reading order, for example, in the case of a specific user, You can have a pattern of reading the news first and then reading the sports field), you can select news that have been viewed for more than a preset time. In addition, the data input unit 132a is service item data corresponding to the'news' service field, such as'press','category','headline word','issue time','length of content','whether images are included', etc. You can choose.

According to various embodiments of the present disclosure, when input data is selected, the data input unit 132a may classify a user's search record for a weekend, a page view record, a search record for a weekday, and a page view record. Similarly, the data input unit 132a may classify a user's search record and page browsing record for each day of the week and select the input data.

The model selection unit 132b may select a machine learning model. According to an embodiment, the model selection unit 132b may apply a preset specific machine learning model without performing a separate selection operation. According to various embodiments, the model selection unit 132b may select a machine learning model to be used for data analysis from among a plurality of machine learning models.

The model selection unit 132b may primarily select an appropriate machine learning algorithm based on input data selected by the data input unit 132a.

Machine learning algorithms are typically Ridge Regression, Logistic Regression, Generalized Linear Model, Random Forest, Gradient Boosting Model, and Neural Network ( Neural Network), etc. exist.

Ridge regression is a basic linear model and provides additional parameters to handle extremes or outliers.

Logistic regression analysis is a basic linear model, and it is effective to use when the target variable, that is, the target to be predicted, shows a binomial distribution.

The generalized linear model is a basic linear model, and it is effective to use when the target variable shows a possion distribution map.

Random forest is a model consisting of a number of decision trees, and each decision tree is developed individually in consideration of the relationship between input variables and target variables, and has more flexible characteristics than the upper linear model.

The change stimulus model is also called the Generalized Boosting Model or GBM, and is often regarded as a model with higher accuracy than the Random Forest because there is a difference in that each decision tree is developed by considering the accuracy of the decision tree that has been created or an ensemble model that has similar tendencies to the Random Forest. It is also possible.

Neural networks are flexible models that can predict virtually any target, and can encompass both linear and non-linear patterns by adjusting parameters. For this reason, parameter tuning needs to be very fine in neural networks.

Thereafter, the model selection unit 132b may select a model among selected machine learning algorithms. In this case, each model may have different training data set. In this case, each model may be classified according to a collection period of the learned data. For example, the first model may be a model trained based on data collected for a period of one month, and the second model may be set as a model trained based on data collected for a period of one week. In addition, according to various embodiments, each model may be classified according to whether or not the weekend of the collection day of the learned data. For example, model A may be a model trained based on data collected during weekdays, and model B may be set as a model trained based on data collected during weekends. Similarly, each model can be trained based on data for each day of the week, and the type of each model can be classified according to the day of the week classification. In addition, the model selection unit 132b may select a corresponding algorithm model according to whether the data analysis operation is performed on a weekday or a weekday/weekend.

Alternatively, the model selection unit 132b may select a model according to characteristics of each service field, which may be determined based on the degree of fluidity of the user's interest. According to an embodiment, the more fluid the user's interest fluidity is, the higher the score may be given. Liquidity figures for specific service sectors can be set by the operator. And based on this, the model selection unit 132b may select an artificial intelligence model that has a longer data collection period or a longer learning period in a service field having a large user's liquidity value.

The design variable calculation unit 133c may calculate data corresponding to the design variable for reprogramming by inserting the variable into the selected artificial intelligence model and driving the algorithm.

First, the design variable calculation unit 133c expresses the user data (eg, search word list) collected by the user data collection unit 131a and the service item data (eg news article) collected by the service item data collection unit 131b on a virtual space. It is possible to perform an embedding operation that converts into a vector of as high a dimension as possible. And by merging them all, you can create an input vector. This may mean a process of converting data in a form not suitable for machine learning into data of a suitable type recognizable by a machine learning algorithm.

The design variable calculation unit 133c may additionally perform an operation of normalizing the scale of the variable when the type of the input variable is numeric.

Thereafter, the design variable calculation unit 133c may perform an operation according to the artificial intelligence algorithm after the embedding and input vector generation operation selected in the selected machine learning algorithm and the selected algorithm model. In this case, the operation may be performed using, for example, a Relu function.

As the calculation operation according to such an artificial intelligence algorithm is performed, the design variable calculation unit 133c is configured for each service item (e.g., news, shopping) determined to have a high correlation with the user's usage pattern according to the calculation result of the machine learning algorithm. , Items such as video) menus (eg, on-screen menus such as politics, society, entertainment, sports, etc. for news services) or items (eg, links to specific articles in the case of news services) may be determined.

In addition, the design variable calculation unit 133c is the result information calculated based on the artificial intelligence algorithm. You can check whether it includes possible items, etc.). In this case, whether or not the variable item may be determined according to whether or not variable item information collected and calculated by the developer data collection unit 131c exists. When a variable item is included, the design variable calculation unit 133c may transmit related information to a worker to change information on a corresponding menu or item. Alternatively, according to various embodiments, the design variable calculation unit 133c may perform reprogramming by directly changing at least one setting value of a variable item of a menu or item calculated as a result of an artificial intelligence operation. Although not shown in the drawing, the design variable calculation unit 133c may include a source code editing unit (not shown) that edits a source code to directly change a variable item of the menu or item.

5 is a diagram showing the configuration of an evaluation unit according to an embodiment of the present invention.

As shown in FIG. 5, the evaluation unit 133 according to an embodiment of the present invention may include a change amount measurement unit 133a.

The change amount measurement unit 133a may measure a change amount of the user's usage after the change occurs as a change of a display menu of a web or an app is applied as a result of the operation of the data analysis unit 132 of the present invention. At this time, the change amount measuring unit 133a is based on the information collected by the user data collecting unit 131a of the data collecting unit 131 of the present invention, based on the change content and the amount of change (e.g., Changes in number of reading, change in reading time, etc.) can be calculated. The change amount measurement unit 133a can extract items with reduced usage and reading time, and controls applied changes to items with reduced usage and reading time to be restored after a preset time (transfer or restoration Can be reprogrammed by reflecting. As described above, the change amount measurement unit 133a provides an attention to the related product when the user's search record for the product temporarily increases for a specific period of time and then the interest in the product decreases after purchase. It can be determined that the necessity of the effect of changing the display menu (eg, increasing the text size, increasing the menu screen size, etc.) for concentrating the values is unnecessary, and accordingly, the changed setting value can be restored to the original state.

6 is a flowchart illustrating a sequence of a programming support operation according to an embodiment of the present invention.

As shown in FIG. 6, the control unit 130 according to an embodiment of the present invention may perform operation 605 of collecting user data and service item data. The user data may refer to information on a user's search word, gender, and web pages viewed. In addition, the service item data may be information related to a service provided by a client, information about a menu name provided by a service app or a web.

Thereafter, the controller 130 may perform operation 610 of collecting developer data. The developer data is information related to the source code of the service app or the web, and may mean specific source code area information set to be changeable, or changeable numerical information.

The controller 130 may then perform operation 615 of extracting variable data to be input to the artificial intelligence algorithm. The control unit 130 may use the user data and service item data as variable data to be injected into the artificial intelligence algorithm.

Thereafter, the controller 130 may perform operation 620 of selecting a model of an artificial intelligence algorithm. Operation 620 may include an operation of selecting a type of machine learning algorithm and an operation of selecting a model of the selected machine learning algorithm. For example, when the ANN is selected as the machine learning algorithm, one model may be selected from among a plurality of algorithm models such as a first ANN and a second ANN classified based on a learning period.

The controller 130 may then perform operation 625 of calculating a menu item to be changed by calculating variable data using an artificial intelligence algorithm. Specifically, operation 625 may mean determining the similarity between user data and service item data, and calculating service item data having a high similarity. In addition, the service item data calculated to have a high similarity to the user data may mean a menu item having a high user selection probability. The controller 130 may perform operation 625 of controlling to change a display setting of a menu calculated after that. According to this operation sequence, the control unit 130 according to an embodiment of the present invention performs setting changes such as expanding the size of a menu with a high probability of selection by the user or changing the color of the headline of a news article with a high probability of selection by the user can do.

Meanwhile, in the various embodiments described above, a method of changing various settings in the front end by reprogramming an app or web using an artificial intelligence model by the device 100 has been described. In general, in order for the artificial intelligence model to operate smoothly There is a disadvantage in that it requires a large storage space and a high-performance GPU.

Accordingly, the artificial intelligence model operated according to another embodiment of the present disclosure may be compressed and used as necessary. For example, the artificial intelligence model may be compressed using at least one of Pruning, Quantization, Decomposition, and Knowledge Distillation. Pruning is one of the compression techniques for removing meaningless weights or weights that do not significantly affect the output value among the weights of the artificial intelligence model. Quantization is one of the compression techniques that quantize each weight into preset bits. Decomposition is one of compression techniques to reduce the size of weights by linearly decomposing a weight matrix or tensor, which is a set of weights. Knowledge Distillation is one of the compression techniques that creates and trains a Student model smaller than the original model by using the original model as the Teacher model. In this case, the Student model may be generated through the aforementioned Pruning, Decomposition, or Quantization.

In another embodiment, the device 100 measures resources consumed to operate the artificial intelligence model according to the present disclosure, and obtains the accuracy of the result output from the artificial intelligence model (for example, it can be obtained through user feedback). can do. The device 100 may input resources and accuracy as input data into another artificial intelligence model to obtain a value for a specific level of accuracy as output data. That is, the reprogramming according to the present invention is to improve the user's satisfaction, and it is not only impossible to have 100% accuracy, but also the utility is poor. Therefore, the device 100 outputs a value for a specific level of accuracy based on the value of the resource required for the artificial intelligence model to operate (for example, the ratio of data storage space, processor share, etc.) and accuracy. I can. Device 100 can compress the artificial intelligence model to satisfy the output data (a certain level of accuracy). Through the above method, the device 100 can efficiently manage resources.

Although the present invention has been described in detail with reference to the above-described examples, those skilled in the art can make modifications, changes, and modifications to these examples without departing from the scope of the present invention. In short, in order to achieve the intended effect of the present invention, it is not necessary to separately include all functional blocks shown in the drawings or to follow all the sequences shown in the drawings as shown in the order shown. Note that it may fall within the range.

110: communication department
120: storage
130: control unit
131: data collection unit
132: data analysis unit
133: evaluation unit
131a: user data collection unit
131b: service item data collection unit
131c: Developer data collection unit
132a: data analysis unit
132b: model selector
132c: design variable calculation unit
133a: change amount measurement unit

Claims (1)

A storage unit for storing commands and data required for performing artificial intelligence-based data analysis;
A communication unit for receiving user data and service item data for a service app or web from a client side;
Based on the user data and service item data, a menu item with a high probability of selection by the user is calculated, and a request for change of the calculated menu setting value is requested from the developer, or in the design of a backend or frontend constituting a service app or web. Including; a control unit for reprogramming by changing and applying the required setting value,
The control unit
A data collection unit that collects data required to calculate a menu item with a high probability of selection by a user;
A data analysis unit for calculating an item to be a variable upon reprogramming by inputting the data collected by the data collection unit into an artificial intelligence algorithm model; And
Including; an evaluation unit that evaluates the amount of change in the user utilization rate after applying the variable calculated by the data analysis unit,
The data collection unit
Based on the AI analysis result set by the developer at the time of initial programming, information on a specific source code that can be variable or numerical information in the source code for a specific page is extracted as a variable item, and the numerical data extracted as a variable item is a variable item Includes; a developer data collection unit that queries the developer whether or not to set as,
The data analysis unit
A data input unit for extracting variable data to be input as a variable in the artificial intelligence algorithm model from the user data and service item information collected by the data collection unit, and extracting based on a service field characteristic value;
A model selection unit that selects an artificial intelligence algorithm model based on the type of variable data selected by the data input unit, and selects an artificial intelligence algorithm model that has a longer learning data collection period in a service field having a greater user interest fluidity;
By inserting a variable into the artificial intelligence algorithm model and performing an algorithmic operation, an item corresponding to the design variable is extracted, but according to the operation result, a menu or item for each service item that is judged to have a correlation with the user's usage pattern is greater than or equal to the reference value Includes; a design variable calculation unit,
The design variable calculation unit
It is checked whether the menu for each service item, which is the result information calculated based on artificial intelligence, includes a variable item, and if it includes a variable item, at least one set value for the variable item of the menu for each service item is changed. Including; source code editing unit for reprogramming,
The control unit
The artificial intelligence algorithm model is compressed based on the ratio of the artificial intelligence algorithm model in the storage unit and the processor share of the artificial intelligence algorithm model, but the compression of the artificial intelligence algorithm model is a method of Pruning, Quantization, Decomposition, and Knowledge Distillation. Is compressed using
The variable item is
Including resizable items, reorderable items, and repositionable items,
The user data item is
Contains the user's gender, age, search term and sequence data,
The evaluation unit
And a change amount measuring unit that measures the amount of change in use after the change is applied as a change in the daily display menu of the web or app is applied, and controls to restore the change after a preset time for the item whose amount is reduced. A service restoration device that is linked with the amount of change in service use according to the menu change.

Images