KR20220121637A - Electronic device and operating method for the same - Google Patents

Abstract

The present invention relates to an electronic device which generates an AI model. The electronic device includes: a memory which stores one or more instructions; and a processor which executes the one or more instructions stored in the memory, wherein the processor obtains first data having a first unit size, determines a first random noise on the basis of the first unit size, and adds the first random noise to the first data to convert the first data into second data which does not have the first unit size, to obtain third data, and to generate an AI model on the basis of training data including the second data and the third data.

Description

Electronic device and operating method thereof

Various embodiments relate to an electronic device and an operating method thereof, and more particularly, to an electronic device generating an AI model using data sensed by the electronic device and data sensed by an external device, and an operating method thereof.

An artificial intelligence (AI) system is a computer system that implements human-level intelligence, and unlike the existing rule-based smart system, a machine learns, judges, and becomes smarter by itself. The more the AI system is used, the better the recognition rate and the more accurate understanding of user preferences, and the existing rule-based smart systems are gradually being replaced by deep learning-based AI systems.

Artificial intelligence technology consists of machine learning and element technologies using machine learning. Machine learning is an algorithm technology that categorizes/learns the characteristics of input data by itself, and utilizes machine learning algorithms such as deep learning, and consists of technical fields such as linguistic understanding, visual understanding, reasoning/prediction, knowledge expression, and motion control. do.

The element technologies for realizing artificial intelligence technology are linguistic understanding technology that recognizes human language/text, visual understanding technology that recognizes objects as if they were human eyes, and reasoning/prediction technology that logically infers and predicts information by judging information. , it may include at least one of a knowledge expression technology that processes human experience information as knowledge data, an autonomous driving of a vehicle, and a motion control technology that controls the movement of a robot.

Various embodiments may provide an electronic device capable of training a tree structure-based classification model using data having different unit sizes, and an operating method thereof.

An electronic device for generating an AI model according to an embodiment includes a memory storing one or more instructions and a processor executing the one or more instructions stored in the memory, wherein the processor includes: first data having a first unit size , determining a first random noise based on the first unit size, and adding the first random noise to the first data, thereby converting the first data to a second data not having the first unit size. The data may be converted into data, third data may be obtained, and an AI model may be generated based on training data including the second data and the third data.

The processor according to an embodiment may obtain fourth data having a second unit size, determine a second random noise based on the fourth data based on the second unit size, and add the second random noise to the fourth data. By summing random noise, the third data not having the second unit size may be obtained.

The electronic device according to an embodiment may further include a sensing unit configured to sense the first data.

The electronic device according to an embodiment may further include a display, and the processor may control to display a guide screen for sensing the first data on the display.

The electronic device according to an embodiment may further include a communication unit configured to receive the third data from an external device.

The AI model according to an embodiment is a tree structure-based authentication model, and the processor may train the AI model by setting the second data as user data and the third data as data of others.

The processor according to an embodiment may determine whether the data sensed by the electronic device is data of the user of the electronic device using the AI model.

According to an embodiment, a method of operating an electronic device for generating an AI model includes: acquiring first data having a first unit size; determining a first random noise based on the first unit size; converting the first data into second data not having the first unit size by adding the first random noise to first data; obtaining third data; and the second data and the second data 3, based on the training data including the data, generating an AI model.

The electronic device according to an embodiment converts first data including discrete values represented by a multiple of a first unit size and second data including discrete values represented by a multiple of a second unit size to data including continuous values. By transforming and training the AI classification model, it is possible to prevent the AI model from being trained in the direction of classification according to the characteristics of the unit size of the data.

In addition, since the electronic device according to an embodiment trains or updates the AI model based on data having various characteristic information collected from external devices as well as data sensed by the electronic device, the performance (accuracy) of the AI model is improved. can increase

1 is a diagram illustrating a method in which a first device generates an AI model by using training data obtained from external devices, according to an embodiment.
2 is a block diagram illustrating a configuration of a first device and a second device according to an exemplary embodiment.
3 and 4 are diagrams referenced to describe a method of a data processing unit converting data and training an AI model based on the transformed data according to an embodiment.
5 is a flowchart illustrating a method of operating a first device and a second device according to an exemplary embodiment.
6 is a flowchart illustrating a method of operating a first device according to an exemplary embodiment.
7 illustrates an example of a guide screen output by the first device to collect first data according to an embodiment.
FIG. 8 is a diagram referenced to describe a method in which a first device performs authentication using an AI model according to an exemplary embodiment.
9 is a block diagram illustrating a configuration of a first device according to another exemplary embodiment.

Terms used in this specification will be briefly described, and the present invention will be described in detail.

The terms used in the present invention have been selected as currently widely used general terms as possible while considering the functions in the present invention, but these may vary depending on the intention or precedent of a person skilled in the art, the emergence of new technology, and the like. In addition, in a specific case, there is a term arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the description of the corresponding invention. Therefore, the term used in the present invention should be defined based on the meaning of the term and the overall content of the present invention, rather than the name of a simple term.

In the entire specification, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated. In addition, terms such as "...unit" and "module" described in the specification mean a unit that processes at least one function or operation, which may be implemented as hardware or software, or a combination of hardware and software. .

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art to which the present invention pertains can easily implement them. However, the present invention may be embodied in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

In the embodiment of the present specification, the term “user” refers to a person who controls functions or operations of an electronic device, a system, etc., and may include a developer, an administrator, or an installer.

1 is a diagram illustrating a method in which a first device generates an AI model by using training data obtained from external devices, according to an embodiment.

The first device 100 according to an embodiment generates an AI model 50 by using data sensed by the first device 100 and data collected from external devices 200 and 300 as training data. It may be an electronic device.

The first device 100 and the external devices 200 and 300 according to an embodiment are, for example, a mobile phone, a tablet PC, a digital camera, a camcorder, a laptop computer, a tablet PC, a desktop computer, an e-book terminal, and digital broadcasting. Implemented in various electronic devices such as terminals, PDA (Personal Digital Assistants), PMP (Portable Multimedia Player), navigation, MP3 player, camcorder, IPTV (Internet Protocol Television), DTV (Digital Television), wearable device, etc. can be

The first device 100 according to an embodiment may sense data of a first device user (hereinafter, referred to as a 'first user'). For example, the first device 100 may include at least one of a touch sensor, an image sensor (camera), a location sensor (eg, GPS), an acceleration sensor, a gyroscope sensor, a microphone, a geomagnetic sensor, and a temperature/humidity sensor. It may include a sensing unit including a. The first device 100 may sense data about the first user (hereinafter, referred to as 'first data') by using the sensing unit.

1 illustrates an example in which the first device 100 obtains touch information (touch data) of the first user by using a touch sensor, but is not limited thereto, and the first device 100 uses various sensors Thus, various data can be sensed.

In this case, the first data D1 sensed by the first device 100 may have a first unit size U1. For example, data sensed by the first device 100 may include discrete values that appear as multiples of the first unit size U1.

In addition, the external devices 200 and 300 may sense user data by using a sensor, respectively. For example, the second device 200 may sense data for the second device user (hereinafter, referred to as 'second data'), and the third device 300 may Hereinafter, referred to as 'third data') may be sensed.

In this case, the second device 200 and the third device 300 may be devices having different types or versions from the first device 100 , and accordingly, the types and sizes of the sensors are the sensors of the first device 100 . may be different from each other, and the unit size of data sensed by the sensors may also have a unit size different from the first unit size of the first device 100 .

For example, the second data D2 sensed by the second device 200 may have a second unit size U2, and the second data D2 is discontinuous in multiples of the second unit size U2. contains negative values. In addition, the third data D3 sensed by the third device 300 may have a third unit size U3 , and the third data D3 is a discontinuous value that appears as a multiple of the third unit size U3 . include those

The first device 100 according to an embodiment includes first data D1 sensed by the first device 100 and second data D2 collected from the second device 200 and the third device 300 . And by using the third data D3, the AI model 50 for authenticating the first device user may be generated. For example, the first device 100 includes the first data D1 as generic user data, and the second data D2 and the third data D3 collected from the external devices 200 and 300 . By training with imposter data, the AI model 50 can be generated.

The AI model 50 according to an embodiment may be a tree structure-based classification model. Accordingly, the first device 100 uses the trained AI model 50 to determine whether input data (eg, data sensed by the device) is device user data or not, and authenticates the device user. can do.

The first device 100 according to an embodiment classifies first data represented by a multiple of the first unit size as user data, and second data or third data represented by a multiple of the second unit size or third unit size as other data. When the AI model 50 is trained to do this, the AI model 50 can be trained by the unit size characteristic of the data, and can be trained to classify the data according to the unit size of the data. For example, the AI model 50 trained as described above is user data if the input data has a first unit size (expressed as a multiple of the first unit size), and the input data has a unit size other than the first unit size. If it has (appears as a multiple of a unit size other than the first unit size), it can be classified as data other than the user.

Accordingly, when a person other than the user of the first device 100 touches the first device 100 , touch data having a first unit size (represented as a multiple of the first unit size) is input, and the above and The AI model trained together classifies the input touch data as user data. Accordingly, there is a problem in that the first device user cannot be authenticated using the AI model trained as described above.

Also, the AI model trained to classify the first user as user data in the first device 100 cannot be used to authenticate the first user in another device having a unit size different from the first unit size. For example, when another device has a second unit size, when the first user touches another device, touch data having a second unit size is input, and the AI model trained in the first device 100 is There is a problem in that the input touch data input by the first user is classified as data other than the user data.

Therefore, in order to prevent the data input to the AI model from being classified by the unit size characteristic, the first device 100 according to an embodiment trains the AI model using the converted data so as not to have a unit size. can

The first device 100 according to an embodiment converts the first data D1 based on the first unit size U1, and the first device 100 or the external devices 200 and 300 The data D2 may be converted based on the second unit size U2 , and the third data D3 may be converted based on the third unit size U3 . The first device 100 may train the AI model by using the converted first to third data as training data. Accordingly, the AI model may be trained to classify the data as user data or other data based on the characteristic information of each of the first to third data. For example, the characteristic information of the touch data may include coordinate information of a touch point, touch intensity information, touched area information, and the like.

The trained AI model may classify the input data into user data or other data based on characteristic information of the input data. Hereinafter, with reference to the drawings, it will be described in detail.

2 is a block diagram illustrating a configuration of a first device and a second device according to an exemplary embodiment.

Referring to FIG. 2 , the first device 100 and the second device 200 according to an embodiment may be different types or different versions of electronic devices, and data sensed in each may have different unit sizes. can have For example, data sensed by the first device 100 may have a first unit size, and data sensed by the second device 200 may have a second unit size.

The first device 100 according to an embodiment includes a sensing unit 110 , a data processing unit 120 , a data storage unit 130 , a communication unit 140 , a data collection unit 150 , and an AI model generation unit 160 . may include.

The sensing unit 110 according to an embodiment may sense a state of the first device 100 or a state around the first device 100 , and transmit the sensed data to the data processing unit 120 .

The sensing unit 110 may include at least one of an image sensor, a touch sensor, an acceleration sensor, a location sensor (eg, GPS), a temperature/humidity sensor, a geomagnetic sensor, a gyroscope sensor, an infrared sensor, a barometric pressure sensor, and a proximity sensor. It may include one, but is not limited thereto. The sensing unit 110 may sense touch information, voice information, location information, step distance information, and the like (first data) of the first device user (first user). In this case, the sensed first data may have a first unit size and include discrete values that appear as multiples of the first unit size.

The data processing unit 120 may acquire characteristic information of the sensed first data. For example, when more than a preset number of touch data is collected, the data processing unit 120 may analyze the collected touch data to obtain characteristic information on the touch data. In this case, the characteristic information on the touch data may include coordinate information of the touch point, touch intensity information, touched area information, and the like. In this case, the characteristic information of the touch data may also have a first unit size and include discrete values that appear as multiples of the first unit size.

Also, the data processing unit 120 may convert the characteristic information of the touch data based on the first unit size. For example, the data processing unit 120 may generate the first converted data by adding the random noise determined based on the first unit size to the characteristic information of the first data. In this case, the first conversion data may include continuous values.

The data storage unit 130 according to an embodiment may store the first converted data.

In addition, the data storage unit 130 according to an embodiment may store data collected from an external device and characteristic information about the data in advance as a database. In this case, the external device may be a device of a different type and different version from that of the first device 100 .

The data collection unit 150 according to an embodiment may control overall operations related to collection of training data used to generate or update an AI model. For example, the data collection unit 150 may determine a time point at which the collection of training data is required, and may control to collect the training data from external devices at the determined time point or periodically. Alternatively, the data collection unit 150 may control the data collected from the external devices to be properly processed by the data processing unit 120 for training data or to be stored in the data storage unit 130 .

The communication unit 140 according to an embodiment may transmit/receive data or signals to and from an external device (eg, the second device 200 ).

The communication unit 140 may include, but is not limited to, a short-range wireless communication unit, a mobile communication unit, and the like, in response to the performance and structure of the first device 100 .

Short-range wireless communication unit, Bluetooth communication unit, BLE (Bluetooth Low Energy) communication unit, near field communication unit (Near Field Communication unit), WLAN (Wi-Fi) communication unit, Zigbee communication unit, infrared (IrDA, infrared) It may include a data association) communication unit, a Wi-Fi Direct (WFD) communication unit, an ultra wideband (UWB) communication unit, an Ant+ communication unit, a microwave (uWave) communication unit, and the like, but is not limited thereto.

The mobile communication unit transmits/receives a radio signal to and from at least one of a base station, an external terminal, and a server on a mobile communication network. Here, the wireless signal may include various types of data according to transmission/reception of a voice call signal, a video call signal, or a text/multimedia message.

The communication unit 140 according to an embodiment may transmit a data request to the second device 200 and may receive data from the second device 200 . For example, the communication unit 140 may receive the second device user's data (second data) sensed by the second device 200 , characteristic information about the second data, or unit size information of the second data. can

When the second data received by the communication unit 140 or the characteristic information on the second data received from the second device 200 includes discrete values having a second unit size, the data processing unit 120 performs the second The data or characteristic information of the second data may be converted based on the second unit size. The data processing unit 120 may generate the second converted data by adding the random noise determined based on the second unit size to the characteristic information of the second data. The second converted data may be stored in the data storage unit 130 .

The AI model generation unit 160 may generate an AI model based on the training data stored in the data storage unit 130 . In this case, the AI model may be an authentication model for authenticating a user, but is not limited thereto. The AI model according to an embodiment may be a tree structure-based classification model. The AI model generator 160 may train the AI model by using the first transformation data and the second transformation data.

Accordingly, the first device 100 uses the trained AI model to determine whether input data (eg, data sensed by the device) is data of the first device user, and the first device user can be authenticated.

The second device 200 according to an embodiment may include a sensing unit 210 , a data processing unit 220 , a data storage unit 230 , and a communication unit 240 .

The sensing unit 210 may sense a state of the second device 200 or a state around the second device 200 , and transmit the sensed second data to the data processing unit 220 . The sensing unit 210 may include at least one of an image sensor, a touch sensor, an acceleration sensor, a location sensor (eg, GPS), a temperature/humidity sensor, a geomagnetic sensor, a gyroscope sensor, and a microphone, However, the present invention is not limited thereto. Also, the sensing unit 210 may sense touch information, voice information, location information, and step distance information of the second device user (second user). In this case, the sensed second data may have a second unit size and may appear as a multiple of the second unit size.

The data processing unit 220 may obtain characteristic information of the second data by analyzing the second data sensed and collected by the sensing unit 210 . Since the method of acquiring the characteristic information of data has been described in detail in the data processing unit 120 of the first device 100 , a detailed description thereof will be omitted. In this case, the characteristic information of the second data may have a second unit size, and may include discrete values that appear as multiples of the second unit size.

Also, the data processing unit 220 may convert the characteristic information of the second data based on the second unit size. For example, the data processing unit 220 may generate the second converted data by adding the random noise determined based on the second unit size to the characteristic information of the second data. In this case, the second converted data may have continuous values.

The data storage unit 230 may store the second converted data. Alternatively, when the data processing unit 220 does not convert the characteristic information of the second data, the data storage 230 may store the characteristic information of the second data including discontinuous values.

The communication unit 240 according to an embodiment may transmit/receive data or signals to and from an external device (eg, the first device 100 ).

The communication unit 240 may receive a data sharing request from the first device 100 , and may transmit the second converted data stored in the data storage unit 230 to the first device 100 . Alternatively, the communication unit 240 may transmit characteristic information of the second data or information on the second unit size to the first device 100 .

3 and 4 are diagrams referenced to describe a method of a data processing unit converting data and training an AI model based on the transformed data according to an embodiment.

Referring to FIG. 3 , the characteristic information (X) of the first data has discrete values that appear as multiples of the first unit size (U1). In addition, the characteristic information (Y) of the second data has discrete values that appear as multiples of the second unit size (U2), and the characteristic information (Z) of the third data (D3) is the third unit size (U3). ) has discrete values that appear as multiples of

In this case, the characteristic information (X) of the first data is classified as user data, and the characteristic information (Y) of the second data and the characteristic information (Z) of the third data (D3) is imposter data. When the tree structure-based first AI classification model is trained to classify as ) is classified as user data, and data 415 having a unit size different from the first unit size is classified as other data. Accordingly, user authentication cannot be accurately performed using the first AI classification model.

Meanwhile, as shown in FIG. 3 , the first device 100 according to an exemplary embodiment has a first random noise (−) determined based on the first unit size (U1) in the characteristic information (X) of the first data By summing U1/2≤N1≤U1/2), the first converted data may be generated. For example, the first converted data X' may be expressed as X-(U1/2)≤X'≤X+(U1/2). In this case, the value of the first random noise N1 may be determined as various values. In addition, since the value of the first random noise N1 is not greater than the characteristic information (X) of the first data, it does not significantly affect the training result of the classification model.

In addition, the first device 100 or the second device 200 determines the second random noise (-U2/2≤N2≤N2≤) determined based on the second unit size (U2) in the characteristic information (Y) of the second data. By summing U2/2), the second converted data Y' may be generated. For example, the second converted data Y′ may be expressed as Y−(U2/2)≦Y′≦Y+(U2/2). In this case, the value of the second random noise N2 may be determined as various values. In addition, since the value of the second random noise N2 is not large compared to the characteristic information (Y) for the second data, it does not significantly affect the training result of the classification model.

In addition, the first device 100 or the third device 300 generates a third random noise (-U3/2≤N3≤) determined based on the third unit size U3 in the characteristic information Z for the third data. U3/2) can be added to generate third transformed data. For example, the third converted data Z' may be expressed as Z-(U3/2)≤Z'≤Z+(U3/2). In this case, the value of the third random noise N3 may be determined as various values. In addition, since the value of the third random noise N3 is not large compared to the characteristic information Z of the third data, it does not significantly affect the training result of the classification model.

The first to third converted data (X', Y', Z') according to an embodiment are represented as continuous values, and accordingly, the first to third converted data (X', Y', Z') are It has no unit size. Therefore, when training the tree structure-based AI classification model using the first to third transformation data (X', Y', Z'), the AI classification model uses characteristics other than the unit size of the data. , can be trained to classify data. Accordingly, as shown in the second graph 420 of FIG. 4 , the second AI classification model does not classify the data based on the unit size, but other characteristics of the data (eg, the average of the data, the standard Data may be classified based on deviation, variance, maximum value, minimum value, and distribution information). Accordingly, based on the characteristic information of the input data, it is possible to classify whether the corresponding data is user data or not. For example, the second AI classification model classifies the input data as user data when the characteristic information of the input data is located in the first area A1, and when the characteristic information of the input data is located in the second area A2 , can be classified as non-user data.

5 is a flowchart illustrating a method of operating a first device and a second device according to an exemplary embodiment.

Referring to FIG. 5 , the second device 200 may sense second data ( S510 ). For example, the second device 200 may sense touch information, voice information, location information, step distance information, and the like (second data) of the second device user. In this case, the sensed second data has a second unit size and may appear as a multiple of the second unit size.

The second device 200 may extract characteristic information of the second data (S520). For example, when the second data is user's touch data, the second device 200 extracts characteristic information including coordinate information of a touch point, touch intensity information, touched area information, and the like from the sensed touch data. can do. In this case, the characteristic information of the second data may also appear as a multiple of the second unit size.

The second device 200 may convert characteristic information of the second data based on the second unit size ( S530 ). For example, the second device 200 may generate second converted data by determining random noise based on the second unit size and adding the determined random noise to characteristic information of the second data.

The second device 200 may transmit the second converted data to the first device 100 (S540).

Meanwhile, in step 540 ( S540 ) of FIG. 5 , it has been described that the second converted data is directly transmitted from the second device 200 to the first device 100 , but the present invention is not limited thereto. The second converted data may be transmitted to the server, and the second converted data may be transmitted from the external server to the first device 100 .

The first device 100 may store the second converted data received from the second device 200 (S550).

The first device 100 may sense the first data (S560). For example, the first device 100 may sense touch information, voice information, location information, step distance information, and the like (first data) of the first device user. In this case, the sensed first data has a first unit size and may appear as a multiple of the first unit size.

The first device 100 may extract characteristic information of the first data (S570). For example, when the first data is the user's touch data, the first device 100 extracts characteristic information including coordinate information of the touch point, touch intensity information, touched area information, etc. from the sensed touch data. can do. In this case, the characteristic information of the first data may also appear as a multiple of the first unit size.

The first device 100 may convert characteristic information of the first data based on the first unit size ( S580 ). For example, the first device 100 may generate the first converted data by determining random noise based on the first unit size and adding the determined random noise to characteristic information of the first data.

The first device 100 may store the first converted data, and may generate an AI model by using the first converted data and the second converted data (S590). In this case, the AI model may be an authentication model for authenticating a user, but is not limited thereto. The AI model according to an embodiment may be a tree structure-based classification model.

6 is a flowchart illustrating a method of operating a first device according to an exemplary embodiment.

Referring to FIG. 6 , the first device 100 according to an embodiment may acquire first data having a first unit size and characteristic information of the first data ( S610 ). For example, the first device 100 may sense touch information, voice information, location information, walking distance information, and the like (first data) of the user of the first device user. In this case, the first data may appear as a multiple of the first unit size.

The first device 100 may extract characteristic information of the sensed first data. For example, when the first data is the user's touch data, coordinate information of the touch point and the touch intensity from the sensed touch data. It is possible to extract characteristic information including information, touched area information, and the like. However, the present invention is not limited thereto. In this case, the characteristic information of the first data may also appear as a multiple of the first unit size.

The first device 100 may convert the characteristic information of the first data based on the first unit size ( S620 ). Since step 620 ( S620 ) corresponds to step 580 ( S580 ) of FIG. 5 , the same description will be omitted.

The first device 100 may receive characteristic information of the second data having a second unit size (S630). For example, the second data may be user data sensed by the second device, and characteristic information of the second data may be displayed as a multiple of the second unit size.

The first device 100 may convert characteristic information of the second data based on the second unit size ( S640 ). For example, the first device 100 may generate second converted data by determining random noise based on the second unit size and adding the determined random noise to characteristic information of the second data. In this case, the second unit size information may be received from the second device 200 or may be extracted through characteristic information of the second data.

The first device 100 may generate an AI model based on the first converted data and the second converted data (S650). Since step 650 ( S650 ) corresponds to step 590 ( S590 ) of FIG. 5 , the same description will be omitted.

The first device 100 may use the generated AI model to authenticate the first device user ( S660 ).

7 illustrates an example of a guide screen output by the first device to collect first data according to an embodiment.

Referring to FIG. 7 , the first device 100 may collect data (first data) of the first device user, and may train an AI authentication model by using the collected first data as user data. In order to increase the performance (accuracy) of the AI authentication model according to an embodiment, it is necessary to train the AI authentication model using data having various characteristic information.

Accordingly, the first device 100 may collect first data in various environments. When collecting the touch data of the first device user, the first device 100 may display a guide screen 710 as shown in FIG. 7 to collect the touch data of the first device user in various environments. .

For example, the first device 100 provides a message to input text while walking slowly, a message to sit and input text, a message to input text using one hand, a message to input text using two hands, etc. may be displayed on the guide screen 710 . However, the present invention is not limited thereto, and a message guiding the user to collect the first data in various situations may be displayed.

Also, the first device 100 may store the sensed touch data together with other types of sensing data for each situation. For example, in a situation where text is input while walking slowly, the sensed touch data and the acceleration data sensed by the acceleration sensor may be stored together. However, the present invention is not limited thereto.

The first device 100 according to an embodiment may convert the sensed first data based on the first unit size, and train the AI authentication model by using the converted data as user data.

FIG. 8 is a diagram referenced to describe a method in which a first device performs authentication using an AI model according to an exemplary embodiment.

Referring to FIG. 8 , the first device 100 according to an embodiment converts first converted data obtained by converting touch data sensed by the first device based on a first unit size, and touch data sensed by the second device. The AI model generated by using the second converted data converted based on the second unit size and the third converted data data converted based on the third unit size of the touch data sensed by the third device as training data can be stored have. In this case, the AI model may be a model for determining whether the user is the first device user (the first user). For example, as shown in FIG. 8 , the first device 100 may sense touch data, and the AI model may receive the sensed touch data as input data. The AI model may determine whether the input touch data is the first user's touch data or not, and may output the result data.

When the received touch data is determined as the first user's touch data, the first device 100 may unlock the first device, and it is determined that the received touch data is not the first user's touch data. In this case, the first device 100 may maintain the locked state of the first device.

In FIG. 8 , a method of user authentication based on touch data has been illustrated and described, but is not limited thereto, and the first device 100 uses various data to generate an AI model, and uses the generated AI model , user authentication can be performed.

9 is a block diagram illustrating a configuration of a first device according to another exemplary embodiment. The first device 900 of FIG. 9 may be an embodiment of the first device 100 of FIG. 1 .

Referring to FIG. 9 , the first device 900 according to an embodiment includes a sensing unit 910 , a communication unit 920 , a processor 930 , an A/V input unit 1240 , an output unit 1250 , and a memory ( 1260 ) and a user input unit 970 .

The sensing unit 910 includes not only a sensor (eg, a fingerprint recognition sensor, etc.) that senses the user's biometric information, but also a sensor that detects a state of the first device 900 or a state around the first device 900 . may include. Also, the sensing unit 910 may transmit information sensed by the sensor to the processor 930 . The sensing unit 910 of FIG. 9 may correspond to the sensing unit 110 of FIG. 2 .

Since the communication unit 920 has been described in detail in FIG. 2 as a configuration corresponding to the communication unit 140 of FIG. 2 , a detailed description thereof will be omitted.

The processor 930 according to an embodiment may control overall operation of the first device 900 . Also, the processor 930 may control other components included in the first device 900 to perform a predetermined operation.

The processor 930 according to an embodiment may execute one or more programs stored in the memory 960 . The processor 930 may include a single core, a dual core, a triple core, a quad core, and multiple cores thereof. Also, the processor 930 may include a plurality of processors.

The memory 960 according to an embodiment may store various data, programs, or applications for driving and controlling the first device 900 .

Also, a program stored in the memory 960 may include one or more instructions. A program (one or more instructions) or an application stored in the memory 960 may be executed by the processor 930 .

The processor 930 according to an embodiment may correspond to at least one of the data processing unit 120 , the data collection unit 150 , and the AI model generation unit 160 shown and described in FIG. 2 . The memory 960 may correspond to the data storage unit 130 of FIG. 2 .

The memory 960 may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg, SD or XD memory), and a RAM. (RAM, Random Access Memory) SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, magnetic disk , may include at least one type of storage medium among optical disks.

The A/V (Audio/Video) input unit 940 is for inputting an audio signal or a video signal, and may include a camera 941 , a microphone 942 , and the like. The camera 941 may obtain an image frame such as a still image or a moving image through an image sensor in a video call mode or a photographing mode. The image captured through the image sensor may be processed through the processor 930 or a separate image processing unit (not shown).

The image frame processed by the camera 941 may be stored in the memory 960 or transmitted to the outside through the communication unit 920 . Two or more cameras 941 may be provided according to the configuration of the first device 900 .

The microphone 942 receives an external sound signal and processes it as electrical voice data. For example, the microphone 942 may receive an acoustic signal from an external device or a speaker. The microphone 942 may use various noise removal algorithms for removing noise generated in the process of receiving an external sound signal.

The output unit 950 is for outputting an audio signal, a video signal, or a vibration signal, and may include a display unit 951 , a sound output unit 952 , a vibration motor 953 , and the like.

The sound output unit 952 outputs audio data received from the communication unit 920 or stored in the memory 960 . Also, the sound output unit 952 outputs a sound signal related to a function (eg, a call signal reception sound, a message reception sound, and a notification sound) performed by the first device 900 . The sound output unit 1252 may include a speaker, a buzzer, and the like.

The vibrator 953 may output a vibration signal. For example, the vibrator 953 may output a vibration signal corresponding to output of audio data or video data (eg, a call signal reception sound, a message reception sound, etc.). Also, the vibrator 1253 may output a vibration signal when a touch is input to the touch screen.

The user input unit 970 means a means for a user to input data for controlling the electronic device 900 . For example, the user input unit 970 includes a key pad, a dome switch, and a touch pad (contact capacitive method, pressure resistance film method, infrared sensing method, surface ultrasonic conduction method, integral type). There may be a tension measurement method, a piezo effect method, etc.), a jog wheel, a jog switch, and the like, but is not limited thereto.

The operating method of the first device according to an embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

In addition, the first device or the method of operating the first device according to the disclosed embodiments may be provided as included in a computer program product. Computer program products may be traded between sellers and buyers as commodities.

The computer program product may include a S/W program and a computer-readable storage medium in which the S/W program is stored. For example, computer program products may include products (eg, downloadable apps) in the form of S/W programs distributed electronically through manufacturers of electronic devices or electronic markets (eg, Google Play Store, App Store). have. For electronic distribution, at least a portion of the S/W program may be stored in a storage medium or may be temporarily generated. In this case, the storage medium may be a server of a manufacturer, a server of an electronic market, or a storage medium of a relay server temporarily storing a SW program.

The computer program product, in a system consisting of a server and a client device, may include a storage medium of the server or a storage medium of the client device. Alternatively, when there is a third device (eg, a smart phone) that is communicatively connected to the server or the client device, the computer program product may include a storage medium of the third device. Alternatively, the computer program product may include the S/W program itself transmitted from the server to the client device or the third device, or transmitted from the third device to the client device.

In this case, one of the server, the client device and the third device may execute the computer program product to perform the method according to the disclosed embodiments. Alternatively, two or more of a server, a client device, and a third device may execute a computer program product to distribute the method according to the disclosed embodiments.

For example, a server (eg, a cloud server or an artificial intelligence server) may execute a computer program product stored in the server to control a client device communicatively connected with the server to perform the method according to the disclosed embodiments.

Although the embodiments have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present invention as defined in the following claims are also included in the scope of the present invention. belongs to

Claims (15)

An electronic device for generating an AI model, comprising:
a memory storing one or more instructions; and
a processor executing the one or more instructions stored in the memory;
The processor is
Obtaining first data having a first unit size,
converting the first data into second data not having the first unit size by determining a first random noise based on the first unit size and adding the first random noise to the first data; ,
obtain third data;
An electronic device for generating an AI model based on training data including the second data and the third data. According to claim 1,
The processor is
Obtaining fourth data having a second unit size, determining a second random noise based on the second unit size, and adding the second random noise to the fourth data, thereby not having the second unit size. to obtain the third data, not the electronic device. According to claim 1,
The electronic device is
The electronic device further comprising a sensing unit for sensing the first data. 4. The method of claim 3,
The electronic device is
further comprising a display,
The processor is
An electronic device that controls to display a guide screen for sensing the first data on the display. According to claim 1,
The electronic device is
The electronic device of claim 1, further comprising a communication unit configured to receive the third data from an external device. According to claim 1,
The AI model is a tree structure-based authentication model,
The processor is
An electronic device for training the AI model by setting the second data as user data and the third data as data of others. 7. The method of claim 6,
The processor is
An electronic device for determining whether data sensed by the electronic device is data of a user of the electronic device by using the AI model. In the operating method of an electronic device for generating an AI model,
acquiring first data having a first unit size;
determining a first random noise based on the first unit size;
converting the first data into second data not having the first unit size by adding the first random noise to the first data;
obtaining third data; and
and generating an AI model based on training data including the second data and the third data. 9. The method of claim 8,
Obtaining the third data includes:
obtaining fourth data having a second unit size; and
determining a second random noise based on the second unit size;
and acquiring the third data not having the second unit size by adding the second random noise to the fourth data. 9. The method of claim 8,
The step of obtaining the first data includes:
The method of operating an electronic device, comprising the step of sensing the first data through a sensing unit. 11. The method of claim 10,
The step of obtaining the first data includes:
and displaying a guide screen for sensing the first data on the display. 9. The method of claim 8,
Obtaining the third data includes:
and receiving the third data from an external device. 9. The method of claim 8,
The AI model is a tree structure-based authentication model,
The step of generating the AI model includes:
and training the AI model by setting the second data as user data and the third data as data of others. 14. The method of claim 13,
The method of operation is
and determining whether the data sensed by the electronic device is data of a user of the electronic device by using the AI model. One or more computer-readable recording media in which a program for performing the method of claim 8 is stored.

Images