KR102696140B1 - Method for providing artificial intelligence model use service and server for performing the same - Google Patents

Abstract

The present invention relates to a method for managing an artificial intelligence model-using service, comprising: a step of obtaining an artificial intelligence model configured to input one type of input data through a service provider device and output predetermined result data, and determining the obtained artificial intelligence model as a serviceable model; a step of providing a service use interface for using a plurality of artificial intelligence models through a service user device; a step of obtaining input data input to one artificial intelligence model through the service use interface; and a step of using the artificial intelligence model to produce result data based on the input data.

Description

Method for providing artificial intelligence model use service and server for performing the same {METHOD FOR PROVIDING ARTIFICIAL INTELLIGENCE MODEL USE SERVICE AND SERVER FOR PERFORMING THE SAME}

The present invention relates to a method for providing an artificial intelligence model utilization service and a server performing the same.

Artificial Intelligence (AI) refers to a technology that realizes human learning ability, reasoning ability, perception ability, and natural language comprehension ability as a computer program. In order to actually service an AI model that uses this AI, factors different from those in AI model development must be considered. For example, users who want to use the service must be able to upload data that matches the input characteristics required by the AI model and easily check the results of applying the model. In addition, the service server that manages the service must allocate resources by considering the memory and GPU usage rate used by the AI model.

In particular, when providing services for various types of models rather than individual models, resources must be considered, and the information and data each model requires from users and the operating process of each model must be standardized and supported. In addition, in order to operate services for a large number of different AI models that are continuously registered on a single server, a technology is needed that allows newly registered models to be serviced easily.

The background technology of the invention has been written to facilitate understanding of the present invention. It should not be understood that the matters described in the background technology of the invention are recognized as prior art.

The present invention relates to a platform that efficiently supports a service provider using an unspecified number of artificial intelligence models when the service provider subscribes to an artificial intelligence model and initiates a service, rather than an optimized service for a specific number of fixed models.

Specifically, the inventors of the present invention recognized that although there exist web pages and applications that provide artificial intelligence models that provide a single result, no platform has been disclosed that allows users to search for artificial intelligence models suitable for their needs and receive services for each field of application of artificial intelligence technology.

Accordingly, the inventors of the present invention developed a method for providing an inference service environment in a simple manner, as there is a demand and supply for artificial intelligence models in the specialized technical field.

In particular, the inventors of the present invention have configured a method to easily register models necessary for sharing and using multiple artificial intelligence models and environmental information for servicing them in a service platform environment.

Furthermore, the inventors of the present invention have configured a method to improve the quality of service and prevent resource shortage due to unnecessary operations by determining the validity of input data entered by a user to use a specific artificial intelligence model.

Furthermore, the inventors of the present invention have configured a method to provide a real-time artificial intelligence model utilization service to service users and to increase the utilization efficiency of the service by scheduling the computational order of the artificial intelligence model according to the platform utilization status.

The tasks of the present invention are not limited to the tasks mentioned above, and other tasks not mentioned will be clearly understood by those skilled in the art from the description below.

In order to solve the above-described problem, an artificial intelligence model utilization service management method according to one embodiment of the present invention is provided. The method is configured to include a step of obtaining an artificial intelligence model configured to input one type of input data through a service provider device and output predetermined result data, and determining the obtained artificial intelligence model as a serviceable model, a step of providing a service utilization interface for using a plurality of artificial intelligence models through a service user device, a step of obtaining input data input to one artificial intelligence model through the service utilization interface, and a step of producing result data based on the input data by using the artificial intelligence model.

According to a feature of the present invention, after the step of acquiring the input data, the method may further include a step of determining the validity of the input data based on the input data matched to the artificial intelligence model.

According to another feature of the present invention, the step of determining the validity of the input data may further include the step of determining suitability of the data format of the acquired input data, and, if the data format is a suitable data format as a result of the determination, the step of determining suitability of the content of the acquired input data.

According to another feature of the present invention, the step of determining suitability for the data format may be a step of determining suitability for at least one of the type, resolution, size, range, and number of frames of the image.

According to another feature of the present invention, the artificial intelligence model is an image analysis model configured to take an image as input and output a preset analysis result, and the step of producing the result data may be a step of inputting a medical image for a target area determined to be suitable for the medical image analysis model and producing an analysis result for the target area.

According to another feature of the present invention, the step of acquiring the artificial intelligence model may further include the step of acquiring data representing the artificial intelligence model service from the service provider device, parsing the acquired data, and acquiring at least one data set among service guide settings, input settings, output settings, and result settings of the artificial intelligence model together with the structure of the artificial intelligence model based on the result of the parsing.

According to another feature of the present invention, the step of acquiring the artificial intelligence model and determining it as a serviceable model may further include the step of providing a service providing interface for implementing an artificial intelligence model service with the service provider device, and the step of extracting at least one data set among service guide settings, input settings, output settings, and result settings of the artificial intelligence model together with the structure of the artificial intelligence model through the service providing interface.

According to another feature of the present invention, the service providing interface may include at least one area among an area for inputting an image or text matching an artificial intelligence model, an area for generating a service providing page for using the artificial intelligence model, and an area for generating a service result page for explaining result data of the artificial intelligence model.

According to another feature of the present invention, the step of transmitting the service use interface may further include the step of generating the service use interface for selecting one of a plurality of artificial intelligence models based on an image or text obtained through the service providing interface.

According to another feature of the present invention, the step of obtaining the input data may further include a step of providing real-time feedback on the input data input to the artificial intelligence model through the service use interface, in response to a request from the service user device.

According to another feature of the present invention, after the step of acquiring the artificial intelligence model, the step of determining whether to provide a service of the artificial intelligence model may further include a step of evaluating the validity of the artificial intelligence model if the determination result is a new artificial intelligence model.

According to another feature of the present invention, the step of evaluating the validity of the artificial intelligence model may further include a step of providing feedback to the service provider device based on layer-by-layer result data produced by inputting test data into the artificial intelligence model.

According to another feature of the present invention, the step of evaluating the validity of the artificial intelligence model may further include the step of calculating computational resources of the artificial intelligence model using test data and the step of storing the calculated computational resources in memory.

According to another feature of the present invention, prior to the step of producing the result data, if input data to be input into at least two or more artificial intelligence models is obtained, the method may further include a step of determining a computation time for the two or more artificial intelligence models based on computational resources of the artificial intelligence models previously stored in the memory and computational characteristics of the artificial intelligence models.

According to another feature of the present invention, after the step of producing the result data, the method may further include a step of displaying the result data through the service use interface or providing the result data to the service user device.

In order to solve the above-described problem, a risk artificial intelligence model utilization service management server according to another embodiment of the present invention is provided. The server includes a communication interface, a memory, and a processor operably connected to the communication interface and the memory, wherein the processor is configured to obtain an artificial intelligence model configured to input any one type of input data through a service provider device and output predetermined result data, and determine the artificial intelligence model as a serviceable model, provide a service utilization interface for using a plurality of artificial intelligence models with a service user device, obtain input data input to any one artificial intelligence model through the service utilization interface, and produce result data based on the input data by using the artificial intelligence model.

Specific details of other embodiments are included in the detailed description and drawings.

The present invention provides an environment in which multiple artificial intelligence models can be supplied and used, thereby increasing users' accessibility to various artificial intelligence models. In particular, the present invention provides a user interface convenient for registering and using artificial intelligence models, thereby standardizing the input/output process of an artificial intelligence model use service, thereby increasing the convenience of use for service providers.

In addition, the present invention enables users without specialized knowledge of the structure of an artificial intelligence model to obtain desired results by using an artificial intelligence model. In particular, by requesting from the user whether input data is suitable or not, and if not suitable, requesting correct input data, the artificial intelligence model utilization service can be provided in real time.

In addition, the present invention can improve the efficiency of service use by scheduling the calculation order of an artificial intelligence model according to the number of service users, the calculation progress status of the artificial intelligence model, and the priority.

In addition, the present invention provides a service that can utilize an artificial intelligence model learned based on learning data possessed by experts, such as in the medical field, thereby obtaining accurate analysis results similar to those diagnosed by medical staff.

In particular, the present invention can improve the quality of a service and prevent resource shortage due to unnecessary operations by performing a step of evaluating input data input by a user who wishes to use the service in advance.

The effects according to the present invention are not limited to those exemplified above, and more diverse effects are included in the present invention.

Figure 1 is a schematic diagram for explaining an artificial intelligence model utilization service according to one embodiment of the present invention.
FIG. 2 is a block diagram of an artificial intelligence model utilization service system according to one embodiment of the present invention.
FIG. 3 is a block diagram showing the configuration of an artificial intelligence model utilization service provider device according to one embodiment of the present invention.
Figure 4 is a schematic flowchart of a method for providing an artificial intelligence model utilization service according to one embodiment of the present invention.
FIG. 5 is a block diagram showing the configuration of an artificial intelligence model utilization service management server according to one embodiment of the present invention.
Figure 6 is a schematic flowchart of a method for managing an artificial intelligence model utilization service according to one embodiment of the present invention.
FIGS. 7A to 7C are exemplary diagrams of user interface screens for registering use of an artificial intelligence model according to one embodiment of the present invention.
Figure 8 is a flowchart specifying step S210 illustrated in Figure 6.
Figure 9 is a flowchart specifying step S230 illustrated in Figure 6.
Figure 10 is a block diagram showing the configuration of an artificial intelligence model utilization service user device according to one embodiment of the present invention.
Figure 11 is a schematic flowchart of a method for using an artificial intelligence model utilization service according to one embodiment of the present invention.
Figure 12 is a schematic diagram generally explaining the flow of an artificial intelligence model utilization service system according to one embodiment of the present invention.

The advantages and features of the present invention, and the methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and these embodiments are provided only to make the disclosure of the present invention complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. With respect to the description of the drawings, similar reference numerals may be used for similar components.

In this document, the expressions "have," "can have," "include," or "may include" indicate the presence of a given feature (e.g., a numerical value, function, operation, or component such as a part), but do not exclude the presence of additional features.

In this document, the expressions "A or B," "at least one of A and/or B," or "one or more of A or/and B" can include all possible combinations of the listed items. For example, "A or B," "at least one of A and B," or "at least one of A or B" can all refer to (1) including at least one A, (2) including at least one B, or (3) including both at least one A and at least one B.

The expressions "first," "second," "first," or "second," etc., used in this document can describe various components, regardless of order and/or importance, and are only used to distinguish one component from another, but do not limit the components. For example, a first user device and a second user device can represent different user devices, regardless of order or importance. For example, without departing from the scope of the rights set forth in this document, a first component can be named a second component, and similarly, a second component can also be renamed as a first component.

When it is stated that a component (e.g., a first component) is "(operatively or communicatively) coupled with/to" or "connected to" another component (e.g., a second component), it should be understood that the component can be directly coupled to the other component, or can be connected via another component (e.g., a third component). Conversely, when it is stated that a component (e.g., a first component) is "directly coupled to" or "directly connected to" another component (e.g., a second component), it should be understood that no other component (e.g., a third component) exists between the component and the other component.

The expression "configured to" as used herein can be used interchangeably with, for example, "suitable for," "having the capacity to," "designed to," "adapted to," "made to," or "capable of." The term "configured to" does not necessarily mean something that is "specifically designed to" in terms of hardware. Instead, in some contexts, the expression "a device configured to" can mean that the device is "capable of" doing something in conjunction with other devices or components. For example, the phrase "a processor configured (or set) to perform A, B, and C" can mean a dedicated processor (e.g., an embedded processor) to perform those operations, or a generic-purpose processor (e.g., a CPU or an application processor) that can perform those operations by executing one or more software programs stored in a memory device.

The terms used in this document are only used to describe specific embodiments and may not be intended to limit the scope of other embodiments. The singular expression may include the plural expression unless the context clearly indicates otherwise. The terms used herein, including technical or scientific terms, may have the same meaning as commonly understood by a person of ordinary skill in the art described in this document. Among the terms used in this document, terms defined in general dictionaries may be interpreted as having the same or similar meaning in the context of the related technology, and shall not be interpreted in an ideal or excessively formal meaning unless explicitly defined in this document. In some cases, even if a term is defined in this document, it cannot be interpreted to exclude the embodiments of this document.

The individual features of the various embodiments of the present invention may be partially or wholly combined or combined with each other, and as can be fully understood by those skilled in the art, various technical connections and operations are possible, and each embodiment may be implemented independently of each other or may be implemented together in a related relationship.

For clarity in the interpretation of this specification, the terms used in this specification are defined below.

The term "artificial intelligence model" used in this specification may be a model learned to predict a specified result by taking input data specified by a service provider as input. Specifically, the artificial intelligence model is based on providing an image analysis service, but in a specialized field, for example, the artificial intelligence model may be a medical image analysis model learned to take a medical image as input and output a pre-specified analysis result. Here, the medical image may be a medical image of an individual, and the analysis result may include the occurrence of a disease in a specific area, the success of treatment, the probability of diagnosis, etc. As another example, the artificial intelligence model may be a medical data analysis model learned to take measurement data as input and output a pre-specified analysis result. Here, the measurement data may be data measured or calculated by an individual, and the analysis result may include the occurrence of a disease, the success of treatment, the probability of diagnosis, etc. As another example, the artificial intelligence model may be an analysis model learned to take audio data such as the voice of an individual, text data acquired through an individual, etc. as input and output a pre-specified analysis result.

In various embodiments, the artificial intelligence model may be composed of at least one model or two or more ensemble models of VGG net, R, DenseNet based on CNN (Convolutional Neural Network), FCN (Fully Convolutional Network) with encoder-decoder structure, SegNet, DeconvNet, DeepLAB V3+, DNN (deep neural network) such as U-net, SqueezeNet, Alexnet, ResNet18, MobileNet-v2, GoogLeNet, Resnet50, Resnet101, Inception-v3.

Figure 1 is a schematic diagram for explaining an artificial intelligence model utilization service according to one embodiment of the present invention.

Referring to FIG. 1, the artificial intelligence model utilization service may mean a service that can utilize various types of artificial intelligence models in one space. The artificial intelligence model utilization service management server (200) (hereinafter referred to as the 'service management server (200)') that provides the artificial intelligence model utilization service of the present invention may not directly develop and provide multiple artificial intelligence models, but may receive multiple artificial intelligence models from a service provider and share them. Here, sharing an artificial intelligence model may not mean providing data on the artificial intelligence model to other users as is, but may be understood as sharing the results obtained when a user inputs data he or she has into an artificial intelligence model that performs a specific function. For example, the artificial intelligence model utilization service may be provided to a service user device (300) through a web page (1) provided by the service management server (200). In addition, the artificial intelligence model utilization service may be provided in the form of an application or program that can be installed or executed on a service provider device (100) and a service user device (300).

A web page (11) of an artificial intelligence model utilization service may provide graphic objects (12) representing a plurality of artificial intelligence models available for use in various fields. Here, the graphic objects (2) may be graphic objects generated by a service provider device (100) providing an artificial intelligence model. Specifically, the service provider device (100) may provide data for explaining an artificial intelligence model to a service management server (200), and the service management server (200) may extract images and texts for explaining the artificial intelligence model from the data and generate graphic objects to be provided to a service user device (300). For example, the image may include an image for intuitively explaining an analysis target of the artificial intelligence model, and the text may include text for explaining a type of data input to the artificial intelligence model and an analysis result that can be obtained through the artificial intelligence model.

The service user device (300) can select any one graphic object displayed on the web page (1) of the artificial intelligence model utilization service, and can receive a user interface for utilizing the corresponding artificial intelligence model from the service management server (200). Here, the user interface for utilizing the artificial intelligence model can also be generated by the service provider device (100), and a detailed description will be provided later.

The service management server (200) can input input data acquired from a service user device (300) into an artificial intelligence model to operate the artificial intelligence model, and produce result data. For example, the service management server (200) can display result data on a web page (1) running on a service user device (300).

So far, an artificial intelligence model utilization service according to one embodiment of the present invention has been roughly described. According to the present invention, data analysis using an artificial intelligence model can be performed more easily in fields such as the medical field that require specialized and accurate result data through an artificial intelligence model utilization service.

Below, we will specifically describe each component of the system that provides artificial intelligence model utilization services.

FIG. 2 is a block diagram of an artificial intelligence model utilization service system according to one embodiment of the present invention.

Referring to FIG. 2, the artificial intelligence model utilization service system (1000) may be a system configured to provide a service that can utilize multiple artificial intelligence models in one space.

An artificial intelligence model utilization service system (1000) may include a service provider device (100) that develops and provides an artificial intelligence model, a service management server (200) that manages multiple artificial intelligence models provided by the service provider device (100) and determines result data requested by a service user using the artificial intelligence model, and a service user device (300) that utilizes the artificial intelligence model.

The service provider device (100) may be a device of a service provider that develops an artificial intelligence model and provides various data that can utilize the artificial intelligence model. For example, the service provider device (100) may include a smartphone, a tablet PC (personal computer), a laptop, and a PC.

In various embodiments, the service provider device (100) may provide data representing an artificial intelligence model service. Specifically, the service provider device (100) may provide a file in the form of json, which may generate an interface screen for describing or using an artificial intelligence model, together with identification data of the artificial intelligence model, to the service management server (200). For example, the file in the form of json may include files for an artificial intelligence model registration and use page structure (input config.), an artificial intelligence model description page structure (model config.), and an artificial intelligence model result page structure (result config.). The service provider device (100) may receive a service registration interface for providing a file in the form of json from the service management server (200). The service registration interface may list essential data items (e.g., input/output standard specifications of the artificial intelligence model) required to provide an artificial intelligence model use service, and the service provider device (100) may provide a file in the form of json that it has created to the service management server (200) through the service registration interface.

In various embodiments, the service provider device (100) may input various data for describing the artificial intelligence model through the service provision interface provided from the service management server (200). For example, the service provider device (100) may input structural data of the artificial intelligence model, learning data, available input data, available result data, accuracy of analysis results, etc.

In various embodiments, the service provider device (100) may configure a service screen for using an artificial intelligence model through a service provision interface provided from a service management server (200). Here, the service screen may include an artificial intelligence model introduction page, an artificial intelligence model use page, an artificial intelligence model result page, etc. For example, the service provider device (100) may input a representative image or text describing an artificial intelligence model through the service provision interface, may input the type, range, example screen, etc. of input data that can be input to the artificial intelligence model, and may input the type, etc. of result data that can be obtained through the artificial intelligence model.

The service management server (200) manages multiple artificial intelligence models and may be a device of an intermediary service provider that can mediate between a service provider device (100) and a service user device (300). For example, it may include a general-purpose computer, laptop, data server, web server, etc. that can produce analysis results based on medical images of a service user device (300) using a medical image analysis model provided by a service provider device (100).

In various embodiments, the service management server (200) may obtain data for describing an artificial intelligence model from the service provider device (100) and generate a service use interface to be provided to the service user device (300) based on the data. Specifically, the service management server (200) may obtain data regarding an artificial intelligence model based on a service provision/registration interface or a file in the form of json. In addition, the service management server (200) may generate a service use interface capable of searching and using multiple artificial intelligence models by combining artificial intelligence models obtained from the service provider device (100).

In various embodiments, the service management server (200) may determine the validity of the artificial intelligence model obtained from the service provider device (100) without registering it as an immediately available service. This may be performed when a new artificial intelligence model is provided, and the service management server (200) may determine the validity of the artificial intelligence model using test data provided by the service provider device (100). For example, the service management server (200) may determine the validity of whether the same result data as the result data input to the service provider device (100) is produced, whether the result data is defined as a probability value using SoftMax, whether there are exceptions in the result data, and whether the resource usage such as the calculation time and the memory capacity is appropriate.

In various embodiments, the service management server (200) may provide a web/app application that can mediate between the service provider device (100) and the service user device (300) through an artificial intelligence model. For example, the service management server (200) may provide a web browser that can be accessed by the service provider device (100) and the service user device (300), or may provide an application or program that can be installed on the service provider device (100) and the service user device (300).

In various embodiments, the service management server (200) may obtain input data from the service user device (300), and perform verification on the input data before calculating an artificial intelligence model and producing result data therefor. Specifically, the service management server (200) may determine the validity of the input data obtained based on the input data previously specified by the service provider. The service management server (200) may determine the suitability of the data format of the input data, and thereafter determine the suitability of the content of the obtained input data. Depending on the determination result, the service management server (200) may perform an artificial intelligence model calculation based on the input data, or request new input data from the service user device (300). For example, the service management server (200) may determine whether a medical image requested as input to a medical image analysis model is an appropriate type and whether the target area of the medical image is appropriate.

The service user device (300) may be a device of a service user who uses an artificial intelligence model utilization service. For example, the service user device (300) may include a smartphone, a tablet PC (personal computer), a laptop, and a PC.

In various embodiments, the service user device (300) may input input data for using one artificial intelligence model through a service use interface provided from the service management server (200). That is, the service user device (300) may request an artificial intelligence model operation result based on the input data to the service management server (200). For example, the service user device (300) may request an analysis result of the medical image using an artificial intelligence model together with the medical image to the service management server (200). The service user device (300) may obtain an analysis result using the corresponding artificial intelligence model from the service management server (200). Here, the analysis result may be provided to the service user device (300) through a result analysis page generated by the service provider device (100).

In various embodiments, the service user device (300) may obtain real-time result data depending on the type of artificial intelligence model and the service provision status. For example, the service user device (300) may provide real-time captured images, videos, or recorded audio as input data requested by the service management server (200), and obtain result data therefor from the service management server (200). However, here, if the amount of computation of the artificial intelligence model increases, a delay time may occur, and the service user device (300) may receive a notification about the delay time from the service management server (200).

So far, an artificial intelligence model utilization service system (1000) according to one embodiment of the present invention has been described. According to the present invention, the artificial intelligence model utilization service system (1000) can increase users' accessibility to artificial intelligence models by providing an environment in which multiple artificial intelligence models can be supplied and utilized.

Below, a service provider device that provides an artificial intelligence model utilization service will be described with reference to FIGS. 3 and 4.

FIG. 3 is a block diagram showing the configuration of an artificial intelligence model utilization service provider device according to one embodiment of the present invention.

Referring to FIG. 3, the service provider device (100) may include a memory interface (110), one or more processors (120), and a peripheral interface (130). Various components within the service provider device (100) may be connected by one or more communication buses or signal lines.

The memory interface (110) is connected to the memory (150) and can transmit various data to the processor (120). Here, the memory (150) can include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (for example, an SD or XD memory, etc.), a RAM, an SRAM, a ROM, an EEPROM, a PROM, a network storage, a cloud, and a blockchain database.

In various embodiments, the memory (150) may store at least one of an operating system (151), a communication module (152), a graphical user interface module (GUI) (153), a sensor processing module (154), a telephone module (155), and an application (156). Specifically, the operating system (151) may include instructions for processing basic system services and instructions for performing hardware operations. The communication module (152) may communicate with at least one of other devices, computers, and servers. The graphical user interface module (GUI) (153) may process a graphical user interface. The sensor processing module (154) may process sensor-related functions (e.g., processing voice input received using one or more microphones (192). The telephone module (155) may process telephone-related functions. The application module (156) may perform various functions of a user application, such as electronic messaging, web browsing, media processing, navigation, imaging, and other processing functions. In addition, the service provider device (100) can store one or more software applications (156-1, 156-2) associated with a type of service (e.g., an application for a service provider utilizing an artificial intelligence model) in the memory (150).

In various embodiments, the memory (150) may store various data for providing an artificial intelligence model utilization service. Specifically, the memory (150) may store configuration data of a service screen designed by a service provider. Such configuration data may be stored in the memory (150) when input through a service provision interface. For example, the memory (150) may store images and texts input to an artificial intelligence model introduction page. For another example, the memory (150) may store structural data of an artificial intelligence model input to an artificial intelligence model utilization page, learning data, test data, types of data that can be input for each artificial intelligence model, types of result data that can be provided for each artificial intelligence model, accuracy of analysis results, updatable data, etc. For another example, the memory (150) may store a layout input to an artificial intelligence model result page.

In various embodiments, the memory (150) may store a digital assistant client module (157) (hereinafter, DA client module), and thereby store instructions for performing client-side functions of the digital assistant and various user data (158) (e.g., user-customized vocabulary data, preference data, user's electronic address book, to-do list, shopping list, and other data).

Meanwhile, the DA client module (157) can obtain the user's voice input, text input, touch input, and/or gesture input through various user interfaces (e.g., I/O subsystem (140)) provided in the service provider device (100).

In addition, the DA client module (157) can output data in audiovisual and tactile forms. For example, the DA client module (157) can output data consisting of a combination of at least two or more of voice, sound, notification, text message, menu, graphic, video, animation, and vibration. In addition, the DA client module (157) can communicate with a digital assistant server (not shown) using a communication subsystem (180).

In various embodiments, the DA client module (157) may collect additional information about the surroundings of the service provider device (100) from various sensors, subsystems, and peripheral devices to construct a context associated with the user input. For example, the DA client module (157) may provide context information along with the user input to a digital assistant server to infer the user's intent. Here, the context information that may accompany the user input may include sensor information, such as lighting, ambient noise, ambient temperature, images of the surroundings, videos, etc. As another example, the context information may include physical state of the service provider device (100) (e.g., device orientation, device location, device temperature, power level, speed, acceleration, motion patterns, cellular signal strength, etc.). As yet another example, the context information may include information related to software state of the service provider device (100) (e.g., processes running on the service provider device (100), installed programs, past and present network activity, background services, error logs, resource usage, etc.).

In various embodiments, the memory (150) may include additional or deleted instructions, and further, the service provider device (100) may include additional configurations other than those illustrated in FIG. 3, or may exclude some configurations.

The processor (120) can control the overall operation of the service provider device (100) and execute various commands to implement a service providing interface that can register an artificial intelligence model utilization service by driving an application or program stored in the memory (150).

The processor (120) may correspond to a computational device such as a CPU (Central Processing Unit) or an AP (Application Processor). In addition, the processor (120) may be implemented in the form of an integrated chip (IC) such as a SoC (System on Chip) in which various computational devices such as an NPU (Neural Processing Unit) are integrated.

In various embodiments, the processor (120) may register an artificial intelligence model so that the functions of the artificial intelligence model can be shared with other users.

In relation to this, FIG. 4 is a schematic flowchart of a method for providing an artificial intelligence model utilization service according to one embodiment of the present invention.

Referring to FIG. 4, the processor (120) can obtain a service providing interface for registering an artificial intelligence model from the service management server (200) (S110). Specifically, the service providing interface can include input windows for defining graphic objects included in a page to be produced by a service provider, and the processor (120) can output a service providing interface including these input windows. For example, the service providing interface can be an interface for configuring an artificial intelligence model registration page, an introduction page, a usage page, a result page, a service history management page, etc. As another example, the service providing interface can include an area for inputting a title, a subtitle, a header, media, and notes.

After step S110, the processor (120) may provide input data for defining an artificial intelligence model through a service providing interface. For example, the processor (120) may obtain an image or text input to a main page introducing an artificial intelligence model. For another example, the processor (120) may obtain structural data representing the number of layers of the artificial intelligence model, an input-output pipeline, and may obtain learning data, test data, types of data that can be input for each artificial intelligence model, types of result data that can be provided for each artificial intelligence model, accuracy of analysis results, etc. For another example, the processor (120) may obtain a layout input to an artificial intelligence model result page. Here, the layout may mean a location, size, etc. where the result data is displayed.

After step S120, the processor (120) can update the artificial intelligence model based on the input data provided from the service management server (200) (S130). Specifically, the processor (120) can obtain the input data input into the artificial intelligence model by the service management server (200) and update the artificial intelligence model based on the input data. In addition, the processor (120) can obtain a list of service user devices (300) using the artificial intelligence model from the service management server (200). This list can be utilized by the service provider to manage service users.

In various embodiments, in addition to obtaining an interface for defining each of a main page, a usage page, and a result page from the service management server (200), the processor (120) may obtain an interface for providing data representing an artificial intelligence model service. Specifically, the processor (120) may provide the service management server (200) with a file in the form of json that can generate an interface screen for describing or using an artificial intelligence model together with identification data of the artificial intelligence model. For example, the file in the form of json may include files for an artificial intelligence model registration and usage page structure (input config.), an artificial intelligence model description page structure (model config.), and an artificial intelligence model result page structure (result config.). The processor (120) may receive a service registration interface for providing a file in the form of json from the service management server (200). The service registration interface may list essential data items (e.g., input/output standard specifications of the artificial intelligence model) required to provide an artificial intelligence model utilization service, and the processor (120) may provide a file in the form of JSON that it has created to the service management server (200) through the service registration interface.

The peripheral interface (130) can be connected to various sensors, subsystems, and peripheral devices to provide data so that the service provider device (100) can perform various functions. Here, it can be understood that the function performed by the service provider device (100) is performed by the processor (120).

The peripheral interface (130) can receive data from a motion sensor (160), a light sensor (light sensor) (161), and a proximity sensor (162), through which the service provider device (100) can perform orientation, light, and proximity detection functions, etc. For another example, the peripheral interface (130) can receive data from other sensors (163) (positioning system-GPS receiver, temperature sensor, biometric sensor), through which the service provider device (100) can perform functions related to the other sensors (163).

In various embodiments, the service provider device (100) may include a camera subsystem (170) connected to a peripheral interface (130) and an optical sensor (171) connected thereto, which enables the service provider device (100) to perform various photographing functions, such as taking pictures and recording video clips.

In various embodiments, the service provider device (100) may include a communication subsystem (180) coupled with a peripheral interface (130). The communication subsystem (180) may be comprised of one or more wired/wireless networks and may include various communication ports, radio frequency transceivers, and optical transceivers.

In various embodiments, the service provider device (100) includes an audio subsystem (190) coupled to the peripheral interface (130), the audio subsystem (190) including one or more speakers (191) and one or more microphones (192), such that the service provider device (100) can perform voice-operated functions, such as voice recognition, voice replication, digital recording, and telephony.

In various embodiments, the service provider device (100) may include an I/O subsystem (140) coupled with the peripheral interface (130). For example, the I/O subsystem (140) may control a touch screen (143) included in the service provider device (100) via a touch screen controller (141). As an example, the touch screen controller (141) may detect a user's contact and movement or cessation of contact and movement using any one of a number of touch sensing technologies, such as capacitive, resistive, infrared, surface acoustic wave technology, proximity sensor array, etc. In another example, the I/O subsystem (140) may control other input/control devices (144) included in the service provider device (100) via other input controller(s) (142). As an example, the other input controller(s) (142) may control one or more buttons, rocker switches, thumb-wheels, infrared ports, USB ports, and pointer devices such as a stylus.

So far, a service provider device (100) according to one embodiment of the present invention has been described. According to the present invention, the service provider device (100) does not simply provide configuration data to enable the use of an artificial intelligence model, but directly configures a service page that can describe and use an artificial intelligence model, thereby enabling the service provider device to more actively promote an artificial intelligence model that it has developed, and to provide information to service users that can enable the use of an artificial intelligence model more efficiently.

Hereinafter, with reference to FIGS. 5 and 6, a service management server (200) that mediates a service provider device (100) and a service user device (300) and manages the overall artificial intelligence model utilization service will be described.

FIG. 5 is a block diagram showing the configuration of an artificial intelligence model utilization service management server according to one embodiment of the present invention.

Referring to FIG. 5, the service providing server (200) may include a communication interface (210), a memory (220), an I/O interface (230), and a processor (240), and each component may communicate with each other through one or more communication buses or signal lines.

The communication interface (210) can be connected to the service provider device (100) and the service user device (300) via a wired/wireless communication network to exchange data. For example, the communication interface (210) can transmit a service provision interface to the service provider device (100) and receive data necessary for using an artificial intelligence model from the service provider device (100). Here, the data necessary for using an artificial intelligence model can include structural data of the artificial intelligence model, learning data, test data (or sample data), types of data that can be input for each artificial intelligence model, types of result data that can be provided for each artificial intelligence model, etc. As another example, the communication interface (210) can transmit a service use interface to the service user device (300) and receive input data to be input to a corresponding artificial intelligence model together with selection information for a certain artificial intelligence model.

Meanwhile, the communication interface (210) that enables transmission and reception of such data includes a communication port (211) and a wireless circuit (212), and the wired communication port (211) may include one or more wired interfaces, for example, Ethernet, a universal serial bus (USB), FireWire, etc. In addition, the wireless circuit (212) may transmit and receive data with an external device via an RF signal or an optical signal. In addition, the wireless communication may use at least one of a plurality of communication standards, protocols, and technologies, for example, GSM, EDGE, CDMA, TDMA, Bluetooth, Wi-Fi, VoIP, Wi-MAX, or any other suitable communication protocol.

The memory (220) can store various data used in the service providing server (200). For example, the memory (220) can store configuration data of a service providing interface and a service using interface, and can additionally store data acquired through each interface. For another example, the memory (220) can store a plurality of artificial intelligence models and learning data used to learn the same, types of input data, types of result data, etc. For another example, the memory (220) can store identification data of a service provider device (100) and a service user device (300).

In various embodiments, the memory (220) may include a volatile or nonvolatile storage medium capable of storing various data, commands, and information. For example, the memory (220) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., an SD or XD memory, etc.), a RAM, an SRAM, a ROM, an EEPROM, a PROM, a network storage storage, a cloud, and a blockchain database.

In various embodiments, the memory (220) may store configurations of at least one of an operating system (221), a communication module (222), a user interface module (223), and one or more applications (224).

An operating system (221) (e.g., embedded operating systems such as LINUX, UNIX, MAC OS, WINDOWS, VxWorks, etc.) may include various software components and drivers to control and manage general system operations (e.g., memory management, storage device control, power management, etc.) and may support communication between various hardware, firmware, and software components.

The communication module (223) can support communication with other devices through the communication interface (210). The communication module (220) can include various software components for processing data received by the wired communication port (211) or wireless circuit (212) of the communication interface (210).

The user interface module (223) can receive a user's request or input from a keyboard, touch screen, microphone, etc. through an I/O interface (230) and provide a user interface on the display.

The application (224) may include a program or module configured to be executed by one or more processors (240). Here, the application for providing multiple artificial intelligence model management services may be implemented on a server farm.

The I/O interface (230) can connect at least one of input/output devices (not shown) of the service providing server (200), such as a display, a keyboard, a touch screen, and a microphone, to the user interface module (223). The I/O interface (230) can receive user input (e.g., voice input, keyboard input, touch input, etc.) together with the user interface module (223) and process a command according to the received input.

The processor (240) is connected to a communication interface (210), a memory (220), and an I/O interface (230) to control the overall operation of the service providing server (200), and can perform various commands for managing an artificial intelligence model utilization service through an application or program stored in the memory (220).

The processor (240) may correspond to a computational device such as a CPU (Central Processing Unit) or an AP (Application Processor). In addition, the processor (240) may be implemented in the form of an integrated chip (IC) such as a SoC (System on Chip) in which various computational devices are integrated. Alternatively, the processor (240) may include a module for calculating an artificial neural network model such as an NPU (Neural Processing Unit).

In various embodiments, the processor (240) can mediate and manage artificial intelligence model utilization services, as described below with reference to FIG. 6.

Figure 6 is a schematic flowchart of a method for managing an artificial intelligence model utilization service according to one embodiment of the present invention.

Referring to FIG. 6, the processor (240) may obtain an artificial intelligence model configured to input a certain type of input data and output designated result data through the service provider device (100) and determine it as a serviceable model (S210). Specifically, the processor (240) may obtain data representing an artificial intelligence model service from the service provider device (100) and may parse the obtained data. For example, the processor (240) may obtain a file in the form of json representing an artificial intelligence model registration and use page structure (input config.), an artificial intelligence model description page structure (model config.), and an artificial intelligence model result page structure (result config.). Here, the file in the form of json may be obtained through a service registration interface provided by the processor (240) to the service provider device (100). At this time, the service registration interface may include essential data items (e.g., input/output standard specifications of the artificial intelligence model) required to provide an artificial intelligence model use service. The processor (240) may obtain at least one data set among the service guide setting, input setting, output setting, and result setting of the artificial intelligence model along with the structure of the artificial intelligence model based on the result of parsing it. For example, the processor (240) may obtain a data set including data corresponding to each of the title, subtitle, header, media, and notes items based on the result of parsing it.

In this way, the processor (240) can obtain all kinds of data that can provide an artificial intelligence model operation service according to a user request and determine this as a serviceable model. In other words, the processor (240) can determine the obtained artificial intelligence model as a model that can be registered in the service providing interface.

In various embodiments, the processor (240) may provide a service providing interface for obtaining an artificial intelligence model as a service provider device (100). The processor (240) may extract at least one data set of a service guide setting, an input setting, an output setting, and a result setting of the artificial intelligence model together with the structure of the artificial intelligence model through the service providing interface. To this end, the service providing interface may include, for example, at least one area among an area for inputting an image or text matching the artificial intelligence model, an area for generating a service providing page for using the artificial intelligence model, and an area for generating a service result page for explaining the result data of the artificial intelligence model. Here, the area for inputting an image or text may be included in the service providing interface to generate a main page introducing the artificial intelligence model.

In relation to this, FIGS. 7a to 7c are exemplary diagrams of user interface screens for registering use of an artificial intelligence model according to one embodiment of the present invention.

Referring to FIGS. 7a and 7b, the processor (240) may provide the following service provision interface screen (11)(11') to the service provider device (100) in order to generate a service use interface that uses an artificial intelligence model service. Here, the underlined portion is an item that is not included in the actual service use interface, and for example, the service provision interface screen (11) may include an area (12) for entering a subtitle, a header, and notes, and a graphic object (13) for selecting an artificial intelligence model. In addition, the service provision interface screen (11') may include a graphic object (14) representing input data to be requested from the service user. In this way, the processor (240) may extract a data set for guide setting and input setting through the service provision interface, and may provide the service provider device (100) with a service provision interface that may extract a data set for output setting and result setting as in FIG. 7c. For example, the service provision interface screen (15) may include an area (16) for entering a subtitle, header, and media, and the processor (240) may extract a text or image data set to be displayed on the service use interface, such as “image result” and “the frontal lobe of the brain is weak.”

In other words, the processor (240) can obtain at least one of the service guide setting, input setting, output setting, and result setting of the artificial intelligence model from the service provider device (100). In addition, the input setting may include an image or text for briefly explaining the artificial intelligence model and a range of data that the artificial intelligence model can operate on, and the result setting may include a range of data that can be learned and output by the artificial intelligence model.

In various embodiments, the processor (240) may generate a service usage interface for selecting one of multiple artificial intelligence models based on an image or text obtained from the service provider device (100). In other words, the processor (240) may collect titles, phrases, images, etc. introducing artificial intelligence models provided by the service provider devices (100) and generate a single interface for using multiple artificial intelligence models.

In various embodiments, the processor (240) may perform an evaluation of whether the acquired artificial intelligence model is valid. Here, the evaluation of validity may be an evaluation of whether the artificial intelligence model is malicious or not, or whether the artificial intelligence model is actually implementable.

In relation to this, FIG. 8 is a flowchart specifying step S210 illustrated in FIG. 6.

Referring to FIG. 8, the processor (240) can implement an artificial intelligence model based on data acquired from the service provider device (100) and store the model in the memory (220), and determine whether an error exists in the code of the artificial intelligence model (S210-1). For example, the processor (240) can implement the artificial intelligence model using a standardized deep learning data conversion format such as ONNX (Open Neural Network Exchange) or NNEF (Neural Network Exchange Format). If an error exists, the processor (240) can request data for implementing the artificial intelligence model from the service provider device (100) again (S210-11. Example). That is, the processor (240) can provide feedback to the service provider device (100) notifying that an error exists in the artificial intelligence model.

Conversely, if there is no error, the processor (240) can determine whether the acquired artificial intelligence model provides a service (S210-2). In other words, the processor (240) can determine whether there is a history of the artificial intelligence model being provided previously. For example, the processor (240) can determine the history of provision based on the identification data of the artificial intelligence model.

If the service of the artificial intelligence model has been provided previously, the processor (240) may understand that the artificial intelligence model updates the structure or learning data of the model. Accordingly, the processor (240) may re-register the artificial intelligence model and store the resource corresponding to the artificial intelligence model in the memory (220) (S210-5, example).

Conversely, if the service of the artificial intelligence model has not been provided before, the processor (240) can input test data into the artificial intelligence model to obtain test result data (S210-3, No). Here, the processor (240) can optimize the operation of the artificial intelligence model and shorten the operation speed, for example, by using TensorRT, depending on the service provision situation of the artificial intelligence model.

Thereafter, the processor (240) can determine whether the test result data is the same as the result data of the artificial intelligence model provided by the service provider device (100) (S210-4). In addition, the processor (240) can determine whether the layer-by-layer result data produced by inputting the test data into the artificial intelligence model is the same as the layer-by-layer result data provided by the service provider device (100).

If the two result data are not identical, the processor (240) may request data or test data for implementing the artificial intelligence model from the service provider device (100) again (S210-11. No). That is, the processor (240) may provide feedback to the service provider device (100) informing that the structure of the artificial intelligence model is inappropriate.

On the contrary, if the two result data are the same, the processor (240) can calculate the resource of the artificial intelligence model and register it as a serviceable artificial intelligence model in the memory (220) (S210-5, example). Specifically, the processor (240) can calculate the computational resources including the unit calculation amount used, the memory capacity, and the disk space capacity as it calculates the artificial intelligence model. The processor (240) can store the calculated computational resources in the memory (220).

Referring back to FIG. 6, after step S210, the processor (240) may provide a service use interface for using multiple artificial intelligence models to the service user device (300) (S220). Here, the service use interface may be generated by the processor (240) as mentioned above, and may be a main page for selecting an artificial intelligence model. The processor (240) may provide a service use interface classified by application field of the artificial intelligence model at the request of the service user device (300).

In various embodiments, the processor (240) may provide a service use interface that defines input data that a user must input to use the artificial intelligence model with a service user device (300) when one artificial intelligence model is selected. That is, the processor (240) may provide a service use interface that represents a use page, and the range of result data that can be provided as a result of the artificial intelligence model's operation may be defined together with the page.

After step S220, the processor (240) can obtain input data input to one artificial intelligence model through the service use interface (S230). The processor (240) can first determine whether the input data is valid before performing the operation of the artificial intelligence model, thereby improving the quality of the service. Specifically, the processor (240) can determine the validity of the input data based on the input data specified for the artificial intelligence model. In addition, the processor (240) can provide real-time feedback on the input data input to the artificial intelligence model through the service use interface at the request of the service user device (300). For example, the real-time feedback can be feedback notifying the inappropriateness of the data format or content of the input data. The processor (240) can help the service user solve a problem in real time through the artificial intelligence model use service of the present invention by providing feedback on the input data in real time. In addition, if the time of providing the feedback is delayed more than the preset time, the processor (240) can notify the service user device (100) of the expected minimum feedback time.

In relation to this, FIG. 9 is a flowchart specifying step S230 illustrated in FIG. 6.

Referring to FIG. 9, the processor (240) can determine whether the data format of the acquired input data is suitable (S230-1). Specifically, the processor (240) can determine whether the data type (e.g., voice, photo, video DICOM, TXT, CSV, excel, MRI, CT, etc.), size, resolution, range, and quantity (e.g., number of frames) match the data format that can be input to the artificial intelligence model. For example, the processor (240) can determine whether the data format is suitable for the artificial intelligence model based on the header information of the acquired medical image DICOM.

If the data format is not suitable, the processor (240) can inform the service user device (100) that the input data is not suitable (S230-11, No). Conversely, if the data format is suitable, the processor (240) can determine whether the content of the acquired input data is suitable (S230-2, Yes). In other words, the processor (240) can determine whether the content of the input data is suitable. For example, if the artificial intelligence model is a medical image analysis model configured to input a medical image and output a preset analysis result, the processor (240) can determine whether the target part of the input data is suitable.

If the content is inappropriate, the processor (240) can inform the service user device (100) that the input data is inappropriate (S230-11, No). Conversely, if the content is appropriate, the processor (240) can input the corresponding input data into the artificial intelligence model and determine the computation time of the artificial intelligence model before that (S230-3, Yes). Specifically, when a service for use of two or more artificial intelligence models is provided, the processor (240) can schedule the computation time by considering the memory (or computational resource) of the artificial intelligence model registered in step S210-5 and the current memory of the service management server (200). If the processor (240) obtains input data to be input to at least two or more artificial intelligence models, the computation time for the two or more artificial intelligence models can be determined based on the computational resource of the artificial intelligence model and the computational characteristics of the artificial intelligence model previously stored in the memory (220). Here, the computational resources can be calculated based on the acquired input data, and the computational characteristics can include resource information (time zone resources) of the current processor (240), urgency, real-time nature of the service, and the resulting capacity of the artificial intelligence model. Accordingly, for example, the computational time point can be determined in proportion to the urgency, delay time, and memory (computational resource) characteristics of the following [Mathematical Formula 1].

[Mathematical Formula 1]

f(p) ∝ f(urgency)g(1/duration)h(1/memory)

After step S230, the processor (240) can use an artificial intelligence model to produce result data based on input data (S240). For example, the processor (240) can input a medical image for a target area that is determined to be suitable for the judgment result and produce an analysis result for the target area before a medical image analysis model configured to input a medical image and output a preset analysis result.

In various embodiments, the processor (240) may display result data through a service use interface or provide result data to a service user device (300) through a communication interface (210). For example, if the expected result of the operation of the artificial intelligence model is longer than a preset time, the processor (240) may transmit and notify result data to the service user device (300).

So far, a service management server (200) according to one embodiment of the present invention has been described. According to the present invention, a service provider can easily recruit service users through an interface provided by the service management server (200), and a service user can easily obtain the results he or she wants by using an artificial intelligence model. In particular, since the service management server (200) determines the validity of input data input to an artificial intelligence model, the load on a server that must compute multiple artificial intelligence models can be reduced.

Below, a service user device that uses an artificial intelligence model utilization service will be described with reference to FIGS. 10 and 11.

Figure 10 is a block diagram showing the configuration of an artificial intelligence model utilization service user device according to one embodiment of the present invention.

Referring to FIG. 10, a service user device (300) may include a memory interface (310), one or more processors (320), and a peripheral interface (330).

The service user device (300) can use the artificial intelligence model through the interface provided by the service management server (200), and can also become a service provider if it possesses the artificial intelligence model. Accordingly, the service user device (300) can have the same configuration as the service provider device (100).

In Fig. 10, description of the overlapping configuration with Fig. 3 will be omitted, and a method of using the artificial intelligence model utilization service performed by the processor (320) will be described with reference to Fig. 11.

Figure 11 is a schematic flowchart of a method for using an artificial intelligence model utilization service according to one embodiment of the present invention.

Referring to FIG. 11, the processor (320) may obtain a service use interface for using an artificial intelligence model from the service management server (200) (S310). For example, the service use interface may sequentially provide an interface for selecting one of a plurality of artificial intelligence models, an interface for checking input data that can be input to the selected artificial intelligence model, and an interface for checking result data obtained through the artificial intelligence model. Here, the interface for selecting one artificial intelligence model may include a graphic object for searching the artificial intelligence model. When the processor (320) receives a search request or filtering request for a keyword from a service user, it may provide a search result or filtering result of an artificial intelligence model corresponding to the keyword.

After step S310, the processor (320) can select one artificial intelligence model through the service use interface and provide input data to the service management server (200) (S320). The processor (320) can transmit input data acquired from the service user to the service management server (200) and receive a validity determination result for the input data from the service management server (200). For example, the processor (320) can receive a determination result for the suitability of the data format of the input data and a determination result for the suitability of the content of the input data from the service management server (200). If the result of the validity determination determines that the input data is invalid, the processor (320) can generate an alarm requesting the input data again through the service use interface. After step S320, the processor (320) can acquire result data based on the input data using the artificial intelligence model from the service management server (200) (S330). The processor (320) can visually display the result data through the service use interface. The processor (320) may, at the request of a service user, convert and provide result data so that it can be stored in a separate storage space, or display the usage history of the artificial intelligence model.

In various embodiments, the processor (320) may interact with the service management server (200) in real time to utilize the artificial intelligence model, depending on the type of the artificial intelligence model and the service provision status. For example, the processor (320) may provide real-time captured images, videos, or recorded audio as input data requested by the service management server (200) through the service use interface, and may obtain result data therefor from the service management server (200). However, if the amount of computation of the artificial intelligence model increases, a delay time may occur, and the service user device (300) may receive a notification about the delay time from the service management server (200).

So far, the service user device (300) of the present invention has been described. According to the present invention, through the artificial intelligence model utilization service provided by the service management server (200), even users without specialized knowledge of the structure of the artificial intelligence model can be provided with desired results by utilizing the artificial intelligence model.

Figure 12 is a schematic diagram generally explaining the flow of an artificial intelligence model utilization service system according to one embodiment of the present invention.

Referring to Fig. 12, the service management server (200) can obtain input data from a service user (User) in real time and evaluate the suitability of the input data. In addition, the resource of the artificial intelligence model that can provide the service can be evaluated to determine the operation order and timing of multiple artificial intelligence models.

In various embodiments, the service management server (200) may include a plurality of inference servers for computing artificial intelligence models. The service management server (200) may determine the computation time and order of artificial intelligence models in each of the plurality of inference servers. Specifically, the service management server (200) may schedule resources by considering server resource information, current resource information, memory used by each artificial intelligence model, and computation time according to input data. For example, the service management server (200) may schedule the computation order of a plurality of artificial intelligence models, or schedule the computation order according to service usage requests of a plurality of service user devices (300) for one artificial intelligence model.

In various embodiments, the service management server (200) may perform calculations of an artificial intelligence model according to a scheduling order on each of a plurality of inference servers. Specifically, the inference server may preprocess input data obtained from a service user device (300) into a form that can be input into the input data of an artificial intelligence model, and may obtain result data corresponding to an analysis result through the calculation. If a calculation request for the same one artificial intelligence model is received from a plurality of service user devices (100), the service management server (200) may continuously perform calculations on the artificial intelligence model based on a plurality of input data obtained from a plurality of service user devices (300).

The service management server (200) can store the inference results according to the operation in its own database (corresponding to the memory (220)) and provide them to the service user.

Although the embodiments of the present invention have been described in more detail with reference to the attached drawings, the present invention is not necessarily limited to these embodiments, and various modifications may be made without departing from the technical idea of the present invention. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention, but to explain, and the scope of the technical idea of the present invention is not limited by these embodiments. Therefore, it should be understood that the embodiments described above are exemplary in all aspects and not restrictive. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within a scope equivalent thereto should be interpreted as being included in the scope of the rights of the present invention.

1000: Artificial Intelligence Model Utilization Service System
100: Service Provider Device
300: Service User Device
110, 310: Memory interface 120, 320: Processor
130, 330: Peripheral Interface 140, 340: I/O Subsystem
141, 341: Touch screen controller 142, 342: Other input controller
143, 343: Touch screen
144, 344: Other input control devices
150, 350: Memory 151, 351: Operating System
152, 352: Communication module 153, 353: GUI module
154, 354: Sensor processing module 155, 355: Phone module
156, 356: Applications
156-1, 156-2, 356-1, 356-2: Applications
157, 357: Digital Assistant Client Module
158, 358: User data
160, 360: Motion sensor 161, 361: Light sensor
162, 362: Proximity sensor 163. 363: Other sensors
170, 370: Camera subsystem 171, 371: Optical sensor
180, 380: Communication Subsystem
190, 390: Audio Subsystem
191, 391: Speaker 192, 392: Microphone
200: Service Management Server
210: Communication Interface
211: Wired communication port 212: Wireless circuit
220: Memory
221: Operating System 222: Communication Module
223: User Interface Module 224: Application
230: I/O interface 240: Processor

Claims (30)

A platform management method for an artificial intelligence model utilization service performed by a processor of an artificial intelligence model utilization service management server,
A step for providing a service providing interface for service registration of an artificial intelligence model configured to input a type of input data to a service provider device and output result data through the above processor;
A step of obtaining a data set related to service guide settings, input settings, output settings, and result settings that can define an artificial intelligence model utilization service, together with structural data representing the number of layers and input/output pipelines that can define the structure of an artificial intelligence model from the service providing interface through the above processor;
A step of implementing an artificial intelligence model capable of taking input data as input and outputting result data based on the structural data through the above processor;
A step of determining whether there is an error in the structural data of the above-described artificial intelligence model;
If the error judgment result shows that no error exists, a step of determining the artificial intelligence model as a serviceable model through the processor and storing the computational resources of the determined model;
A step of generating a service usage interface that can utilize an artificial intelligence model in which the error does not exist and computational resources are stored by using the data set through the above processor;
A step of providing a service use interface that can utilize multiple artificial intelligence models that can be provided as a service to a service user device through the above processor;
A step of obtaining input data to be input into one artificial intelligence model selected by a service user from the service use interface through the above processor;
A step of sequentially determining the validity of the input data through the above processor, whether the data format and content of the input data are appropriate;
Through the above processor, based on the judgment result, a step of inputting suitable input data into an artificial intelligence model selected by the service user, and confirming whether a service using multiple artificial intelligence models is provided from multiple service users;
In case a service for using the plurality of artificial intelligence models is provided, a step of calculating the computational resources that can be currently applied to each of the plurality of artificial intelligence models by using the stored computational resources and the plurality of input data through the processor;
A step of determining the computation time of each of the plurality of artificial intelligence models by using the computed computational resources through the processor; and
A method for managing an artificial intelligence model utilization service, comprising: a step of producing a plurality of result data based on the plurality of input data according to the operation time through the processor; delete In the first paragraph,
The step of determining the validity of the above input data is:
A step of determining suitability of the acquired input data for the data format through the above processor; and
A method for managing an artificial intelligence model-using service, further comprising: a step of determining, through the processor, suitability of the acquired input data for the content if the data format is a suitable data format as a result of the determination; In the third paragraph,
The above artificial intelligence model is,
It is an image analysis model configured to input an image and output a preset analysis result.
The steps for producing the above result data are:
A method for managing an artificial intelligence model-using service, the method comprising: inputting an image for a target area suitable for the judgment result of the image analysis model through the above processor; and producing an analysis result for the target area. In paragraph 4,
The steps for determining suitability for the above data format are:
A method for managing an artificial intelligence model-using service, the method comprising: determining suitability of at least one of the type, resolution, size, range, and number of frames of the image through the processor. In the first paragraph,
The steps for obtaining the above data set are:
A method for managing an artificial intelligence model utilization service, further comprising: a step of parsing data in the form of JSON representing an artificial intelligence model service from the service provision interface through the processor; In the first paragraph,
The steps for obtaining the above data set are:
A method for managing an artificial intelligence model use service, further comprising: a step of extracting the data set representing an artificial intelligence model service from the service providing interface, which includes at least one area among an area for inputting an image or text matching an artificial intelligence model, an area for generating a service provision page for using an artificial intelligence model, and an area for generating a service result page for explaining the result data of the artificial intelligence model, through the processor; delete In Article 7,
The steps for creating the above service usage interface are:
A method for managing an artificial intelligence model utilization service, the method comprising: generating a service utilization interface for selecting one of a plurality of artificial intelligence models based on an image or text obtained from the service provision interface through the processor. In the first paragraph,
The step of obtaining the above input data is:
A method for managing an artificial intelligence model use service, further comprising: providing real-time feedback on the input data input to the artificial intelligence model through the service use interface at the request of the service user device through the processor; In the first paragraph,
After the step of determining whether the above error exists,
A step of determining whether to provide a service of the artificial intelligence model through the above processor; and
A method for managing an artificial intelligence model utilization service, further comprising: a step of evaluating the validity of the artificial intelligence model through the processor if the artificial intelligence model is a new artificial intelligence model as a result of the judgment; In Article 11,
The step of evaluating the validity of the above artificial intelligence model is:
A method for managing an artificial intelligence model use service, further comprising a step of providing feedback to the service provider device based on layer-by-layer result data produced by inputting test data into the artificial intelligence model through the processor. In Article 11,
The step of evaluating the validity of the above artificial intelligence model is:
A step of calculating the computational resources of the artificial intelligence model using test data through the above processor; and
A method for managing an artificial intelligence model utilization service, further comprising: a step of determining whether the calculated computational resources are appropriate based on the memory capacity of the server through the processor; In the first paragraph,
Before the step of producing the above result data,
If you have acquired input data that is input to at least two artificial intelligence models,
A method for managing an artificial intelligence model utilization service, further comprising: a step of determining a computation time for two or more artificial intelligence models based on computational resources of artificial intelligence models stored in a memory operably connected to the processor and computational characteristics of the artificial intelligence models through the processor; In the first paragraph,
After the step of producing the above result data,
A method for managing an artificial intelligence model-using service, further comprising: a step of displaying the result data through the service use interface or providing the result data to the service user device via the processor; communication interface;
memory;
a processor operably connected to the communication interface and the memory;
The above processor,
Provide a service providing interface for service registration of an artificial intelligence model configured to input a type of input data as input and output result data as output by a service provider device, and obtain a data set related to at least one of service guide settings, input settings, output settings, and result settings that can define an artificial intelligence model usage service together with structure data representing the number of layers and input/output pipelines that can define the structure of the artificial intelligence model from the service providing interface, implement an artificial intelligence model that can input input data as input and output result data based on the structure data, determine whether an error exists in the structure data of the implemented artificial intelligence model, and if no error exists as a result of the error determination, determine the artificial intelligence model as a serviceable model, and store computational resources of the determined model, and generate a service usage interface that can use an artificial intelligence model in which the error does not exist and computational resources are stored by using the data set, and provide the service usage interface that can use a plurality of artificial intelligence models that can be serviced by a service user device, obtain input data input to one artificial intelligence model selected by a service user from the service usage interface, and sequentially determine the validity of the input data as to whether the data format and content of the input data are appropriate, and, based on the determination result, provide appropriate input data as the service. An artificial intelligence model utilization service management server configured to input an artificial intelligence model selected by a user, and to check whether a plurality of artificial intelligence model utilization services are provided from a plurality of service users, and if a utilization service for the plurality of artificial intelligence models is provided, to calculate computational resources that can be currently charged to each of the plurality of artificial intelligence models using the stored computational resources and the plurality of input data, and to determine computational points in time for each of the plurality of artificial intelligence models using the calculated computational resources, and to produce a plurality of result data based on the plurality of input data according to the computational points in time. delete In Article 16,
The above processor,
Determine the suitability of the acquired input data for the data format,
An artificial intelligence model utilization service management server further configured to determine suitability of the acquired input data for the content if the data format is determined to be a suitable data format. In Article 18,
The above artificial intelligence model is,
It is an image analysis model configured to input an image and output a preset analysis result.
The above processor,
An artificial intelligence model utilization service management server configured to input an image for a target area suitable for the judgment result of the above image analysis model and to produce an analysis result for the target area. In Article 19,
The above processor,
An artificial intelligence model utilization service management server configured to determine suitability for at least one of the following items: type, resolution, size, range, and number of frames of the above image. In Article 16,
The above processor,
An artificial intelligence model utilization service management server further configured to parse data in JSON format representing an artificial intelligence model service from the above service provision interface. In Article 16,
The above processor,
An artificial intelligence model utilization service management server further configured to extract the data set representing the artificial intelligence model service from the service provision interface including at least one area among an area for inputting an image or text matching the artificial intelligence model, an area for generating a service provision page for utilizing the artificial intelligence model, and an area for generating a service result page for explaining the result data of the artificial intelligence model. delete In Article 22,
The above processor,
An artificial intelligence model utilization service management server configured to generate a service utilization interface for selecting one of a plurality of artificial intelligence models based on an image or text obtained from the service provision interface. In Article 16,
The above processor,
An artificial intelligence model usage service management server further configured to provide real-time feedback on the input data input to the artificial intelligence model through the service usage interface at the request of the service user device. In Article 16,
The above processor,
An artificial intelligence model utilization service management server further configured to determine whether the artificial intelligence model provides a service, and if the determination result is a new artificial intelligence model, to evaluate the validity of the artificial intelligence model. In Article 26,
The above processor,
An artificial intelligence model utilization service management server further configured to provide feedback to the service provider device based on layer-by-layer result data produced by inputting test data into the artificial intelligence model. In Article 26,
The above processor,
Calculate the computational resources of the artificial intelligence model using test data,
An artificial intelligence model utilization service management server further configured to determine whether the calculated computational resources are appropriate based on the memory capacity of the above server. In Article 16,
The above processor,
If you have acquired input data that is input to at least two artificial intelligence models,
An artificial intelligence model utilization service management server further configured to determine a computation time for two or more artificial intelligence models based on computational resources of artificial intelligence models previously stored in the memory and computational characteristics of the artificial intelligence models. In Article 16,
The above processor,
An artificial intelligence model utilization service management server further configured to display the result data through the service utilization interface or provide the result data to the service user device.

Images