KR20220117802A - Electronic device and method for controlling thereof - Google Patents

Abstract

Disclosed are an electronic device and a control method thereof. The electronic device may include: a memory; and a processor, when text is input, inputting information on the text, which is input to a first model trained for determining a semantic role of a sentence component included in the sentence, to determine the semantic role of the sentence component included in the text, inputting a sentence component corresponding to the determined semantic role to a second model trained for outputting a dangerous level of the text based on the semantic role of the sentence component included in the sentence, and performing an operation corresponding to the dangerous level of the acquired text.

Description

ELECTRONIC DEVICE AND METHOD FOR CONTROLLING THEREOF

The present disclosure relates to an electronic device and a control method thereof, and more particularly, to an electronic device that performs an operation corresponding to a risk level of an input text, and a control method thereof.

Recently, artificial intelligence systems have been used in various fields. Unlike the existing rule-based smart system, the artificial intelligence system is a system in which the machine learns, judges, and becomes smarter by itself. The more the artificial intelligence 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 artificial intelligence systems.

Recently, a chatbot using a deep learning-based artificial intelligence system has been developed and is widely used. For example, a customer center chatbot that provides a response to a query when a query regarding a defect or status of the device is input has been widely used.

The customer center chatbot using existing technology recognizes or detects the words included in the input query to determine the level of risk of the situation contained in the query. Meanwhile, a difference in the degree of risk that may be included in a word included in a query may be different according to various contexts. However, there was a limitation that the customer center chatbot using the existing technology could not clearly distinguish the difference in risk that could be contained in a word depending on the context.

The present disclosure has been devised to solve the above problems, and an object of the present disclosure is to determine the semantic range of a sentence component included in a text, and obtain the risk level of the text by using the sentence component corresponding to the determined semantic domain. Disclosed are an electronic device for performing a control operation of a target device suitable for a risk level and a control method thereof.

The electronic device according to an embodiment of the present disclosure inputs information about the input text into a first model trained to determine a semantic role of a sentence component included in a sentence when a text is input into a memory. The sentence component corresponding to the determined semantic range is input to the second model trained to determine the semantic range of the sentence component included in the text, and output a risk level based on the semantic domain of the sentence component included in the text. The processor may include a processor that acquires a risk level of the text and performs an operation corresponding to the acquired risk level of the text.

Meanwhile, the first model may be trained to determine each of the sentence components of the input text as one of an agent, a recipient, and a predicate.

Meanwhile, the second model may be characterized in that it is learned to increase the risk level of the text when the sentence component determined as at least one of the subject and the target means a user or a device.

On the other hand, the second model identifies a word having a similar meaning to a sentence component determined as one of the subject, object, and predicate, and is trained to output the risk level of the text using a weight matched to the identified word. can be characterized.

Meanwhile, the second model may be trained to output one of a plurality of grades classified according to the degree of risk as the risk level of the text.

Meanwhile, the second model may be trained to output the risk level of the text as a numerical value indicating the degree of risk.

Meanwhile, the operation corresponding to the risk level is at least one of shutting down the power of the target device, blocking the network connection of the target device, restarting the target device, and ending a background app running in the target device, and the target device It may be at least one of the electronic device and the other electronic device.

On the other hand, further comprising a communication unit including a circuit, wherein the processor, when the risk level of the text is equal to or greater than a first threshold grade or greater than the first threshold, the text and an operation corresponding to the risk level correspond to the text It is possible to control the communication unit to transmit to a server that manages a device, or to provide a notification message of a situation corresponding to the text and an operation corresponding to the risk level.

Meanwhile, the processor may perform a power-off operation of the target device when the risk level of the text is equal to or greater than a second threshold level or greater than or equal to a second threshold value.

Meanwhile, the processor may input the input text into a syntax analysis model learned to parse, and obtain information about the text including a result of the syntax analysis of the text.

On the other hand, further comprising a microphone, wherein the processor, when a user voice inquiring about the state of the electronic device or another device is inputted through the microphone, the inputted user voice is inputted into an Auto Speech Recognition (ASR) model and the A text corresponding to the user's voice may be obtained.

In a control method according to another embodiment of the present disclosure, when text is input, information on the input text is input into a first model trained to determine a semantic role of a sentence component included in a sentence. determining the semantic range of the sentence component included in the text; inputting the sentence component corresponding to the determined semantic domain to the second model trained to output the risk level based on the semantic domain of the sentence component included in the text The method may include obtaining a risk level of the text and performing an operation corresponding to the obtained text risk level.

Meanwhile, the first model may be trained to determine each of the sentence components of the input text as one of an agent, a recipient, and a predicate.

Meanwhile, the second model may be characterized in that it is learned to increase the risk level of the text when the sentence component determined as at least one of the subject and the target means a user or a device.

On the other hand, the second model identifies a word having a similar meaning to a sentence component determined as one of the subject, object, and predicate, and is trained to output the risk level of the text using a weight matched to the identified word. can be characterized.

Meanwhile, the second model may be trained to output one of a plurality of grades classified according to the degree of risk as the risk level of the text.

On the other hand, the second model may be characterized in that it is learned to output the risk level of the text as a numerical value indicating the degree of risk.

Meanwhile, the operation corresponding to the risk level may be at least one of shutting down the power of the target device, blocking the network connection of the target device, restarting the target device, and ending a background app running in the target device.

Meanwhile, in the performing step, when the risk level of the text is equal to or greater than a first threshold grade or greater than or equal to a first threshold, the text and an operation corresponding to the risk level are transmitted to a server managing a device corresponding to the text or providing a notification message of a situation corresponding to the text and an action corresponding to the risk level.

Meanwhile, the performing may include performing a power-off operation of the target device when the risk level of the text is greater than or equal to a second threshold level or greater than or equal to a second threshold, wherein the target device is configured to use the electronic device or It may be at least one of the other electronic devices.

Meanwhile, the step of receiving the input may further include inputting the input text into a syntax analysis model trained to analyze the text, and obtaining information about the text including a result of parsing the text.

On the other hand, when a user's voice inquiring about the state of the electronic device or another device is inputted through the microphone, inputting the inputted user's voice into an Auto Speech Recognition (ASR) model to obtain a text corresponding to the user's voice may further include.

According to the above-described various embodiments, the electronic device may more accurately evaluate a dangerous level of a situation included in the input text and respond quickly.

1 is a block diagram schematically illustrating a configuration of an electronic device according to an embodiment of the present disclosure;
2 is a flowchart illustrating a method of controlling an electronic device according to an embodiment of the present disclosure;
3 is a view for explaining a process in which an electronic device outputs a risk level of text according to an embodiment of the present disclosure;
4 is a view for explaining the configuration and operation of the first model, according to an embodiment of the present disclosure;
5 is a view for explaining the configuration and operation of a second model, according to an embodiment of the present disclosure;
6 is a diagram for explaining a control operation of a target device, according to an embodiment of the present disclosure.
7 is a block diagram illustrating in detail the configuration of the electronic device 100 according to an embodiment of the present disclosure.

Prior to describing the present disclosure in detail, a description will be given of the description of the present specification and drawings.

First, terms used in the present specification and claims have been selected in consideration of functions in various embodiments of the present disclosure. However, these terms may vary depending on the intention or legal or technical interpretation of a person skilled in the art, and the emergence of new technology. Also, some terms are arbitrarily selected by the applicant. These terms may be interpreted in the meaning defined in the present specification, and if there is no specific term definition, it may be interpreted based on the general content of the present specification and common technical knowledge in the art.

Also, the same reference numerals or reference numerals in each drawing attached to this specification indicate parts or components that perform substantially the same functions. For convenience of description and understanding, the same reference numerals or reference numerals are used in different embodiments. That is, even though all components having the same reference number are illustrated in a plurality of drawings, the plurality of drawings do not mean one embodiment.

Various elements and regions in the drawings are schematically drawn. Accordingly, the technical spirit of the present invention is not limited by the relative size or spacing drawn in the accompanying drawings.

In addition, in this specification and claims, terms including an ordinal number, such as "first" and "second", may be used to distinguish between elements. This ordinal number is used to distinguish the same or similar elements from each other, and the meaning of the term should not be construed as limited due to the use of the ordinal number. For example, the components combined with such an ordinal number should not be limited in the order of use or arrangement by the number. If necessary, each ordinal number may be used interchangeably.

In this specification, the singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, terms such as "comprises" or "consisting of" are intended to designate that the features, numbers, steps, operations, components, parts, or combinations thereof described in the specification exist, and are intended to indicate that one or more other It is to be understood that this does not preclude the possibility of addition or presence of features or numbers, steps, operations, components, parts, or combinations thereof.

The expression "configured to (or configured to)" as used in this disclosure depends on the context, for example, "suitable for," "having the capacity to" ," "designed to," "adapted to," "made to," or "capable of." The term “configured (or configured to)” may not necessarily mean only “specifically designed to” in hardware.

Instead, in some circumstances, the expression “a device configured to” may mean that the device is “capable of” with other devices or parts. For example, the phrase “a coprocessor configured (or configured to perform) A, B, and C” may refer to a dedicated processor (eg, an embedded processor), or one or more software programs stored in a memory device, to perform the corresponding operations. By doing so, it may mean a generic-purpose processor (eg, a CPU or an application processor) capable of performing corresponding operations.

In an embodiment of the present disclosure, terms such as “module”, “unit”, “part”, etc. are terms for designating a component that performs at least one function or operation, and such component is hardware or software. It may be implemented or implemented as a combination of hardware and software. In addition, a plurality of "modules", "units", "parts", etc. are integrated into at least one module or chip, except when each needs to be implemented as individual specific hardware, and thus at least one processor. can be implemented as

A component (eg, a first component) is "coupled with/to (operatively or communicatively)" to another component (eg, a second component) When referring to "connected to", it should be understood that the certain element may be directly connected to the other element or may be connected through another element (eg, a third element). On the other hand, when it is said that a component (eg, a first component) is "directly connected" or "directly connected" to another component (eg, a second component), the component and the It may be understood that other components (eg, a third component) do not exist between other components.

Meanwhile, the electronic device 100 may include, for example, at least one of a smart phone, a tablet PC, a desktop PC, a laptop PC, a netbook computer, a workstation, a medical device, a camera, and a wearable device. However, the present invention is not limited thereto, and the electronic device 100 may be implemented as various types of devices such as a display device, a refrigerator, an air conditioner, and a cleaner.

In the present disclosure, the term user may refer to a person who uses an electronic device or a device (eg, an artificial intelligence electronic device) using the electronic device.

Hereinafter, the present disclosure will be described in detail with reference to the drawings.

1 is a block diagram schematically illustrating a configuration of an electronic device 100 according to an embodiment of the present disclosure. 1 , the electronic device 100 may include a memory 110 and a processor 120 . However, the configuration shown in FIG. 1 is an exemplary diagram for implementing embodiments of the present disclosure, and appropriate hardware and software configurations at a level obvious to those skilled in the art may be additionally included in the electronic device 100 .

The memory 110 may store data related to at least one other component of the electronic device 100 or at least one instruction. The command means one action statement that can be directly executed by the processor 120 in a programming language, and is a minimum unit for the execution or operation of a program. The memory 110 is accessed by the processor 120 , and reading/writing/modification/deletion/update of data by the processor 120 may be performed.

In the present disclosure, the term memory may be implemented as a memory 110 , a ROM (not shown), a RAM (not shown) in the processor 120 , or a memory separate from the processor 120 . In this case, the memory 110 may be implemented in the form of a memory embedded in the electronic device 100 or may be implemented in the form of a memory detachable from the electronic device 100 according to the purpose of data storage. For example, data for driving the electronic device 100 is stored in a memory embedded in the electronic device 100 , and data for an extended function of the electronic device 100 is detachable from the electronic device 100 . It can be stored in any available memory.

Data required for at least one of a first model, a second model, an Auto Speech Recognition (ASR) model, and a syntax analysis model to perform various operations may be stored in the memory 110 . A description of each model will be provided in the following section.

The processor 120 may be electrically connected to the memory 110 to control overall operations and functions of the electronic device 100 . The processor 120 may include one or a plurality of processors. At this time, one or more processors are general-purpose processors such as CPU (Central Processing Unit), AP (Application Processor), DSP (Digital Signal Processor), etc., and dedicated graphics such as GPU (Graphic Processing Unit) and VPU (Vision Processing Unit). It can be a processor or a processor dedicated to artificial intelligence, such as a Neural Processing Unit (NPU). When one or more processors are AI-only processors, the AI-only processor may be designed with a hardware structure specialized for processing a specific AI model.

In addition, the processor 120 may be implemented as a system on chip (SoC), large scale integration (LSI), or a field programmable gate array (FPGA) having a built-in processing algorithm. In addition, the processor 120 may perform various functions by executing computer executable instructions stored in a memory.

The processor 120 may receive text input from the user. The text input by the processor 120 may include text inquiring about the status or defect of the electronic device 100 or another device.

In an embodiment, the processor 120 may receive text input from the user through a virtual keyboard UI displayed on the touch screen.

In another embodiment, when a voice inquiring about the status or defect of the electronic device 100 or another device is input through the microphone 160 , the processor 120 inputs the input voice into the ASR model to respond to the voice. text can be obtained. Here, the ASR model (or STT (Speech-to-Text) model) refers to an artificial intelligence model trained to recognize an input voice and output a text corresponding to the recognized voice.

The processor 120 may input information about the inputted text into the first model to determine the semantic range of sentence components included in the text. The information on the text may include a result of sentence parsing of the text. The processor 120 may obtain a syntax analysis result by inputting text into a syntax analysis model trained to perform a syntax analysis operation. Here, the syntax analysis may be performed based on general knowledge, and the general knowledge may be formed of at least one of a thesaurus, a synonym, a sub-concept, and a super-concept included in the thesaurus. The processor 120 may obtain a more accurate syntax analysis result by inputting synonyms, synonyms, sub-concepts, and super-concepts together with text into the syntax analysis model trained to perform the syntax analysis operation.

The first model refers to an artificial intelligence model trained to determine the semantic domain of a sentence component included in a sentence. The semantic domain refers to the semantic role of an action or a noun phrase described by a predicate in a sentence, and may include, for example, an agent, a recipient, and a predicate. That is, when information on text is input, the first model may be trained to determine each sentence component of the text as one of a subject, an object, and a predicate. The configuration and operation of the first model will be described in detail with reference to FIGS. 3 and 4 .

The processor 120 may acquire the risk level of the text by inputting a sentence component corresponding to the semantic domain determined by the first model into the second model. Here, the second model refers to an artificial intelligence model trained to output a risk level based on the semantic domain of a sentence component included in the sentence.

The risk level of the text refers to a numerical value or grade indicating the degree of danger or urgency implied in the situation that the text means. A high numerical value corresponding to the risk level of the text or a high grade corresponding to the risk level of the text may mean that the degree of danger or urgency implied in the situation in the text is large.

That is, the second model may be trained to determine, as the risk level of the text, one of a plurality of risk classes classified according to the degree of risk or the degree of urgency. As another example, the second model may be trained to output the risk level of the text as a numerical value representing the degree of risk.

As an embodiment, when the text inquiring about the state or defect of the electronic device 100 or other device includes a sentence component indicating a user (or, a person) or a specific device, it is embedded in the text. There is a high possibility that the risk or urgency of the situation is high. Accordingly, when the sentence component determined as at least one of the subject and the target means a user or a specific device, the second model may increase the risk level of the text by a preset value or by a preset grade.

In another embodiment, the second model may be trained to identify a word having a similar meaning to a sentence component determined as one of a subject, an object, and a predicate, and output the risk level of the text using a weight matched to the identified word. have. Specifically, the second model may be trained to identify words similar to sentence components using a language database such as a dictionary (eg, thesaurus) in the learning step.

For example, it is assumed that a specific word not learned by the second model is included in the text input to the electronic device 100 . The second model may identify a word having a similar meaning to a specific word, and output a risk level of the text using a weight matched to the identified word. That is, the second model may infer the meaning of the unlearned word by using the pre-learned language database even if some words are not learned.

Meanwhile, the processor 120 may identify whether the risk level of the text is greater than or equal to a threshold grade or greater than a threshold. If the risk level of the text is greater than or equal to the threshold grade or greater than the threshold, it may mean that the risk or urgency of the situation in the text is very high.

The critical grade means a grade set by a user among a plurality of risk grades classified according to the degree of risk, and may of course be changed. For example, it is assumed that multiple risk classes are classified into Extreme, High, Mid, Low, and None in the order of the degree of danger. At this time, the critical grade may be determined as a high grade by the user, but of course it may be changed to Extreme or Mid.

The threshold value means a value preset by an experiment or research, of course, but may be changed by a user.

The processor 120 may perform an operation corresponding to the risk level of the text. The action corresponding to the risk level may be at least one of shutting down the power of the target device, blocking the network connection of the target device, restarting the target device, and terminating a background app running in the target device. Here, the target device may be at least one of the electronic device 100 or another electronic device.

If the risk level of the text is greater than or equal to the first threshold grade or is identified as greater than or equal to the first threshold, the processor 120 transmits an operation corresponding to the text and the risk level to a server managing a device corresponding to the text. control or provide a notification message of an action corresponding to a situation and a risk level corresponding to the text. Here, if the risk level of the text is less than the first threshold grade or is identified as less than the first threshold value, the processor 120 may provide only a notification of a dangerous situation to the user without performing a separate additional control operation.

The processor 120 may control the target device to perform only a minimum function as the level of the risk level increases and the value increases. Specifically, the processor 120 may perform an operation for overall system recovery, such as shutting down or restarting the power of the target device. For example, the processor 120 performs a power-off operation of the target device when the risk level of the text is equal to or greater than a second threshold level, which is a level higher than the first threshold level, or is equal to or greater than a second threshold value, which is a value greater than the first threshold level. can do.

As another embodiment of the present invention, information on a dangerous situation from a user's query may be identified, and a target device control operation corresponding to the dangerous situation may be performed.

Specifically, when the internal temperature of the target device is equal to or greater than the first threshold temperature, the processor 120 may terminate all background apps of the target device. When the internal temperature of the target device is equal to or greater than the second threshold temperature higher than the first threshold temperature, the processor 120 may control the target device to restart the target device. When the internal temperature of the target device is equal to or greater than the third threshold temperature higher than the second threshold temperature, the processor 120 may turn off the power of the target device.

Also, if it is identified that the volume of the target device expands by more than a preset volume, the processor 120 may shut off the power of the target device.

Here, the device corresponding to the text may mean a device indicated by a sentence component determined as a subject or an object. That is, when it is identified that the risk level of the text is greater than or equal to the threshold grade or the threshold, the processor 120 may perform an operation corresponding to a dangerous or urgent situation meaning the text.

For example, it is assumed that the processor 120 receives a text indicating that the battery of the smart phone is inflated and determines that the risk level of the text is greater than or equal to a threshold grade. The processor 120 may control the communication unit 130 to transmit text to a server managing the smart phone. Accordingly, the server administrator can make it possible to urgently respond to a situation occurring in the smart phone.

In addition, the processor 120 may provide a notification message of a situation corresponding to the text. For example, the processor 120 may control the display 140 to display a UI including pre-stored manual information related to the battery of the smart phone or a coping method. As another example, the processor 120 may control the display 140 to display information (eg, how to behave when the battery is inflated) provided from the server managing the smart phone.

As another example, the processor 120 may control the speaker 150 to output an emergency notification sound or message indicating that the situation corresponding to the text is an emergency situation. As another example, the processor 120 may control the speaker 150 to output information received from the server managing the smart phone in the form of voice. As another example, the processor 120 may make a call to the number of a manager of a server that manages a pre-registered smart phone. The present invention is not limited thereto, and an operation corresponding to the risk level of the text may be implemented in various ways.

Through the above-described embodiment, the electronic device 100 may quickly respond to an emergency situation of various devices by acquiring the risk level of the input text, and the user may be provided with information to cope with the emergency situation.

Meanwhile, if it is identified that the risk level of the text is less than the threshold grade or the threshold, the processor 120 may store a log file indicating that the text is input in the memory 110 .

Meanwhile, the function related to artificial intelligence according to the present disclosure is operated through the processor 120 and the memory 110 . One or more processors 120 control to process input data according to a predefined operation rule or artificial intelligence model stored in the memory 110 .

The predefined action rule or artificial intelligence model is characterized in that it is created through learning. Here, being made through learning means that a basic artificial intelligence model is learned using a plurality of learning data by a learning algorithm, so that a predefined action rule or artificial intelligence model set to perform a desired characteristic (or purpose) is created means burden. Such learning may be performed in the device itself on which artificial intelligence according to the present disclosure is performed, or may be performed through a separate server and/or system.

Examples of the learning algorithm include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning.

The artificial intelligence model includes a plurality of artificial neural networks, and the artificial neural network may be composed of a plurality of layers. Each of the plurality of neural network layers has a plurality of weight values, and a neural network operation is performed through an operation between an operation result of a previous layer and a plurality of weights. The plurality of weights of the plurality of neural network layers may be optimized by the learning result of the artificial intelligence model. For example, a plurality of weights may be updated so that a loss value or a cost value obtained from the artificial intelligence model during the learning process is reduced or minimized.

The artificial neural network may include a deep neural network (DNN), for example, a Convolutional Neural Network (CNN), a Deep Neural Network (DNN), a Recurrent Neural Network (RNN), a Restricted Boltzmann Machine (RBM), There may be a Deep Belief Network (DBN), a Bidirectional Recurrent Deep Neural Network (BRDNN), or a Deep Q-Networks, but is not limited to the above-described example.

2 is a flowchart illustrating a method of controlling the electronic device 100 according to an embodiment of the present disclosure.

The electronic device 100 may receive a text input (S210). As an embodiment, the electronic device 100 may receive a text including a content to inquire about a state or a defect of the electronic device 100 or another device from a user.

As another embodiment, the electronic device 100 may receive a user voice inquiring about the state or defect of the electronic device 100 or another device. The electronic device 100 may obtain a text corresponding to the user's voice by inputting the inputted user's voice into the ASR model.

The electronic device 100 may input information about the input text into the first model to determine the semantic range of a sentence component included in the text ( S520 ). The first model may mean an artificial intelligence model trained to determine the semantic domain of a sentence component included in a sentence.

Specifically, the electronic device 100 may obtain information about the text including the syntax analysis result by inputting the input text into the syntax analysis model. The electronic device 100 may input information about the text into the first model and determine each of the sentence components included in the text as one of a subject, an object, and a predicate.

The electronic device 100 may obtain the risk level of the text by inputting the sentence component corresponding to the determined semantic domain into the second model ( S230 ).

The second model refers to an artificial intelligence model trained to output a risk level based on a semantic domain of a sentence component included in a sentence. The second model may output one of a plurality of risk classes classified according to the degree of risk as the risk level of the text. As another example, the second model may be trained to output the risk level of the text as a numerical value indicating the degree of risk.

The electronic device 100 may perform an operation corresponding to the risk level of the acquired text (S240). Here, the operation corresponding to the risk level may be at least one of shutting down the power of the target device, blocking the network connection of the target device, restarting the target device, and terminating a background app running in the target device. Here, the target device may be at least one of the electronic device 100 or another electronic device.

Specifically, the electronic device 100 may identify whether the risk level of the acquired text is greater than or equal to a first threshold level or greater than or equal to a second threshold. When it is identified that the risk level of the text is greater than or equal to the threshold grade or the threshold, the electronic device 100 transmits an action corresponding to the text and the risk level to a server managing the device corresponding to the text, or a situation and risk level corresponding to the text It is possible to provide a notification message of an operation corresponding to . When the danger level of the text is equal to or greater than the second threshold level or equal to or greater than the second threshold value, the electronic device 100 may perform a power-off operation of the target device.

In this case, the device corresponding to the text may mean a device in which one sentence component of the text determined as a subject or an object means. In addition, the notification message of the situation corresponding to the text may include a message in which the electronic device 100 can cope with the situation corresponding to the pre-stored text or a notification sound for notifying the situation corresponding to the text. As another example, the notification message of a situation corresponding to the text may include information received from a server managing a device corresponding to the text.

Meanwhile, when it is identified that the risk level of the text is less than the threshold level or the threshold value, the electronic device 100 may store a log file indicating that the text has been input.

3 is a diagram for describing a process in which the electronic device 100 acquires a risk level of a text using each model, according to an embodiment of the present disclosure. As shown in FIG. 3 , each model 20 , 40 , and 60 may be connected in a pipeline structure.

However, this is only an embodiment, and each model 20 , 40 , and 60 may be implemented as a component of a risk level determination model that is one artificial intelligence model. In this case, the risk level determination model is a model trained to output the risk level 70 of the text using the input text 10 , and may be implemented in an end-to-end structure.

Also, each model 20 , 40 , and 60 may be embedded in the electronic device 100 , but at least one of the models 20 , 40 , and 60 may be included in the server.

The electronic device 100 may obtain information about the text 10 including the syntax analysis result 30 by inputting the text 10 into the classification analysis model 20 . For example, it is assumed that the text (X Charger inflated my phone battery) 10 is a text inquiring about a situation in which the X charger inflates the battery. The syntax analysis result may be output as shown in Table 1 below.

ROOT
(S
(NP (NNP X) (NNP Charger))
(VP (VBD inflated)
(NP (PRP$ my) (NN phone) (NN battery)))
(..))))

The electronic device 100 may input information about the text including the syntax analysis result 30 into the first model 40 to determine the semantic range of the sentence component included in the text ( 50 ). For example, X Charger may be determined as a subject, inflate may be determined as a predicate, and my phone battery may be determined as a target. Input to (60) can output a risk level (70) of the text. The second model 60 may output a risk level 70 indicating whether it is dangerous or urgent for the subject 'X charger' to perform the operation 'inflate' on the target 'my phone battery'. Each individual meaning of the words 'X charger', 'my phone battery', and 'inflate' included in the text 10 may not have a meaning of a dangerous or urgent situation. However, when each word is combined, the meaning of a situation in which the charger swells the battery is derived. Situations from which meaning is derived are likely to be dangerous or urgent. When the risk level of the text 10 is determined by detecting the degree of risk for each word, there is a high possibility that a situation included in the text is misunderstood as not dangerous.

Accordingly, the second model outputs the risk level of the text by using the sentence component of the text corresponding to the determined semantic range, thereby more accurately representing the degree of danger or the degree of urgency contained in the context in which the text is meant.

On the other hand, the second model may increase the risk level of the text when the subject or object means a user or a device. For example, 'X charger' and 'My phone battery', which are sentence components determined as subject and subject, refer to a specific device, and the second model may increase the risk level of the text by a preset grade or numerical value.

The electronic device 100 may perform an operation corresponding to the risk level 70 of the text. Since the operation corresponding to the risk level has been described above, the overlapping description will be omitted.

4 is a view for explaining the configuration and operation of the first model according to an embodiment of the present disclosure.

As shown in FIG. 4 , the first model may include an encoder layer 410 and a decoder layer 430 . The encoder layer 410 may output code information (output data) 420 by extracting data necessary to determine a semantic range of a sentence component included in text from a result of the syntax analysis (input data). That is, the code information 420 may mean information in which information for determining a semantic range of a sentence component included in a text is compressed.

The decoder layer 430 may use the code information 420 to output data (output data) for determining a semantic range of a sentence component included in the text.

As shown in FIG. 4 , the data for determining the semantic range of the sentence component output from the decoder layer 430 and the result of parsing may have a one-to-one matching relationship. For example, there may be a one-to-one matching relationship between the third line ((NP (NNP X) (NNP Charger))) of the classification analysis result and the third component (Device_Agent) of the output data that can determine the semantic domain of the sentence component. . Accordingly, the X Charger among the sentence components of the text may be determined as the subject in the semantic domain.

5 is a diagram for explaining the configuration and operation of a second model according to an embodiment of the present disclosure.

As an embodiment, when a sentence component corresponding to the semantic domain determined by the first model is input, the second model may output a risk level of the text. For example, when sentence components (Agent: X Charger, Predicate: Inflate, Recipient: My phone battery) corresponding to the determined semantic domain are input, the second model may output a risk level of the text.

As another embodiment, when the data 510 in which the sentence component corresponding to the semantic range determined by the first model and the syntax analysis result are combined is input, the second model may output the risk level of the text. That is, the electronic device 100 combines the syntax analysis result 30 shown in FIG. 3 and the sentence component 50 corresponding to the determined semantic domain, and inputs the combined data 510 into the second model to provide text of risk level can be obtained. The combined data 510 may be implemented as shown in Table 2 below.

ROOT / - /None
S / - / None
NP / X Charger / Device_Agent
VP / Inflate / Predicate
NP / My phone battery / Device_Recipient
. / - / None

Meanwhile, the second model may be trained to identify a word having a similar meaning to a sentence component determined as one of a subject, an object, and a predicate, and output a risk level of the text using a weight matched to the identified word. Specifically, the second model may be trained to identify words similar to sentence components using a language database 540 composed of a dictionary (eg, thesaurus) including synonyms. For example, It is assumed that the second model is not trained using training data including the word 'inflate'. The second model may identify a word (eg, blow up) having a similar meaning to 'inflate', which is a sentence component determined as a predicate using the language database 540 . The second model may output the risk level of the text using a weight matched to the identified 'blow up'. That is, the second model may output the risk level of the text by using the weight corresponding to the sentence component determined as the subject and the target and the weight matched to 'blow up'. According to the above-described embodiment, the second model is In the learning stage, the meaning of the unlearned word can be identified using a database including synonyms such as a dictionary. Accordingly, it is possible to reduce the amount of data, training time, and cost for training the second model.

Meanwhile, as an embodiment, the second model may classify (or determine) the risk level of the text into one of a plurality of risk classes using a classification method. An output layer of the second model may include a softmax layer. The soft max layer refers to a layer that performs a soft max function so that the sum of probabilities of correct answers becomes 1 for all possible results as predicted values of input data.

For example, suppose a case is classified into 5 risk classes (Extreme, High, Mid, Low, None) according to the degree of risk. When the combined data 510 is input, the second model may output a probability 520 that each of the five risk grades is a grade corresponding to the risk level of the text using the soft max layer. Since the probability that the risk level of the text is Extreme among the plurality of risk grades is the highest as 90%, the electronic device 100 may identify the risk level of the text as Extreme.

As another embodiment of the present disclosure, the second model may output the risk level of the text as a numerical value 530 indicating the degree of risk by using a regression method. The higher the numerical value indicating the degree of risk, the higher the degree of risk or urgency of the situation contained in the text.

The electronic device 100 may identify the degree of risk or urgency of a situation contained in the text by identifying whether the risk level of the text output from the second model is equal to or greater than a threshold grade or a threshold value. When the risk level of the text is a threshold grade or a threshold value or more, the electronic device 100 may transmit the text to a server managing a device corresponding to the text or provide a notification message of a situation corresponding to the text.

6 is a diagram for explaining a control operation of a target device, according to an embodiment of the present disclosure. When a user query 610 (eg, the smartphone generates a lot of heat) is input in a voice or text format, the processor 120 may obtain information 620 about the risk level of the user query 610 . When the acquired risk level is the first threshold grade or the first threshold value or more, the processor 120 may perform a control operation of the target device corresponding to the risk level. In addition, if the acquired risk level is a second threshold grade, which is a grade higher than the first threshold grade, or if the acquired risk level is a second threshold value that is a value greater than the first threshold value, a dangerous situation (eg, mobile phone fever) and danger to the user A UI 630 for performing an operation corresponding to the level (eg, a power-off operation) may be provided.

7 is a block diagram illustrating in detail the configuration of the electronic device 100 according to an embodiment of the present disclosure. 7 , the electronic device 100 includes a memory 110 , a processor 120 , a communication unit 130 , a display 140 , a speaker 150 , a microphone 160 , an input unit 170 , and a sensor. (180) may be included. Since the memory 110 and the processor 120 have been described in detail with reference to FIG. 1 , redundant descriptions will be omitted.

The communication unit 130 may communicate with an external device. In this case, the communication connection of the communication unit 130 with the external device may include communication through a third device (eg, a repeater, a hub, an access point, a server, or a gateway).

The communication unit 130 may receive a user voice input through the microphone 160 wirelessly connected to the electronic device 100 . The communication unit 130 may transmit the text to a server (or a server manager's device) that manages a device corresponding to the text input to the electronic device 100 . The communication unit 130 may receive information for coping with a situation corresponding to the text from the server.

Meanwhile, the communication unit 130 may include various communication modules to communicate with an external device. For example, the communication unit 130 may include a wireless communication module, for example, LTE, LTE Advance (LTE-A), CDMA (code division multiple access), WCDMA (wideband CDMA), UMTS (universal mobile telecommunications) system), a cellular communication module using at least one of Wireless Broadband (WiBro), Global System for Mobile Communications (GSM), ^5th generation (5G), and the like.

As another example, the wireless communication module may include, for example, at least one of wireless fidelity (WiFi), Bluetooth, Bluetooth low energy (BLE), and Zigbee.

The display 140 may display various information under the control of the processor 120 . In particular, the display 140 may display text input by the user (or text corresponding to the user's voice). The display 140 may display a UI including information for coping with a situation corresponding to the text. As another example, the display 140 may display an indicator or message indicating that the situation corresponding to the text is dangerous or urgent.

The display 140 may be implemented as a touch screen together with a touch panel, or may be implemented as a flexible display.

The speaker 150 is configured to output various types of audio data on which various processing tasks such as decoding, amplification, and noise filtering have been performed by the audio processing unit, as well as various notification sounds or voice messages. The speaker 150 may output information for coping with a situation corresponding to the text in the form of a voice. As another example, the speaker 150 may output a notification sound indicating that the situation corresponding to the text is dangerous or urgent.

Meanwhile, a configuration for outputting audio may be implemented as a speaker, but this is only an exemplary embodiment and may be implemented as an output terminal capable of outputting audio data.

The microphone 160 may receive a user's voice. In this case, the microphone 110 may receive a trigger voice (or wake-up voice) requesting the start of recognition of the ASR model, and receive specific information (eg, information about the state of the electronic device 100 or other devices, etc.) It may receive a requesting user query. In particular, the microphone 160 may be provided inside the electronic device 100 , but may be provided outside and electrically connected to the electronic device 100 . As another example, the microphone 160 may be provided outside the electronic device 100 to be connected to the electronic device 100 through wireless communication.

The input unit 170 may receive a user input for controlling the electronic device 100 . The input unit 170 may receive text input by the user. As another example, the input unit 170 may receive a user command for determining a critical level among a plurality of risk levels. As another example, the input unit 170 may receive a user command for changing the threshold value.

In particular, the input unit 170 may include a touch panel for receiving a user touch input using a user's hand or a stylus pen, a button for receiving a user manipulation, and the like. In addition, the input unit 170 may be implemented as another input device (eg, a keyboard, a mouse, a motion input unit, etc.).

The sensor 180 may detect various state information of the electronic device 100 . For example, the sensor 180 may include a motion sensor (eg, a gyro sensor, an acceleration sensor, etc.) capable of detecting motion information of the electronic device 100, and a sensor (eg, GPS) capable of detecting location information. (Global Positioning System) sensor), and a sensor capable of detecting the user's presence (eg, a camera, a UWB sensor, an IR sensor, a proximity sensor, an optical sensor, etc.) and the like. In addition, the sensor 180 may further include an image sensor for photographing the outside of the electronic device 100 .

Meanwhile, various embodiments of the present disclosure may be implemented as software including instructions stored in a machine-readable storage medium readable by a machine (eg, a computer). A device that calls an instruction and can operate according to the called instruction, and may include the server cloud according to the disclosed embodiments.When the instruction is executed by a processor, the processor directly or other configuration under the control of the processor A function corresponding to the command can be performed using the elements.

Instructions may include code generated or executed by a compiler or interpreter. The device-readable storage medium may be provided in the form of a non-transitory storage medium. Here, 'non-transitory storage medium' does not include a signal and only means that it is tangible and does not distinguish that data is semi-permanently or temporarily stored in the storage medium. For example, the 'non-transitory storage medium' may include a buffer in which data is temporarily stored.

According to an embodiment, the method according to various embodiments disclosed in the present disclosure may be provided by being included in a computer program product. Computer program products may be traded between sellers and buyers as commodities. The computer program product may be distributed in the form of a machine-readable storage medium (eg, compact disc read only memory (CD-ROM)) or online through an application store (eg, Play Store™). In the case of online distribution, at least a portion of the computer program product (eg, a downloadable app) is at least temporarily stored in a storage medium such as a memory of a server of a manufacturer, a server of an application store, or a relay server, or temporarily can be created with

Each of the components (eg, a module or a program) according to various embodiments may be composed of a singular or a plurality of entities, and some sub-components of the aforementioned sub-components may be omitted, or other sub-components may be various. It may be further included in the embodiment. Alternatively or additionally, some components (eg, a module or a program) may be integrated into a single entity to perform the same or similar functions performed by each corresponding component prior to integration. According to various embodiments, operations performed by a module, program, or other component may be sequentially, parallel, repetitively or heuristically executed, or at least some operations may be executed in a different order, omitted, or other operations may be added. can

110: memory 120: processor

Claims (22)

In an electronic device,
Memory;
When text is input, information on the input text is input into the first model trained to determine the semantic role of the sentence component included in the sentence, and the semantic role of the sentence component included in the text is determined, and ,
obtaining the risk level of the text by inputting the sentence component corresponding to the determined semantic range into the second model trained to output the risk level based on the semantic range of the sentence component included in the sentence;
and a processor configured to perform an operation corresponding to the risk level of the acquired text. According to claim 1,
The first model is
The electronic device, characterized in that it has been learned to determine each of the sentence components of the input text as one of an agent, a recipient, and a predicate. 3. The method of claim 2,
The second model is
The electronic device according to claim 1 , wherein when the sentence component determined as at least one of the subject and the target means a user or a device, it is learned to increase the risk level of the text. 3. The method of claim 2,
The second model is
The electronic device is characterized in that it is learned to identify a word having a similar meaning to a sentence component determined as one of the subject, object, and predicate, and output the risk level of the text using a weight matched to the identified word. According to claim 1,
The second model is
An electronic device, characterized in that it is learned to output one of a plurality of grades classified according to the degree of risk as the risk level of the text. According to claim 1,
The second model is
The electronic device, characterized in that it is learned to output the risk level of the text as a numerical value indicating the degree of risk. According to claim 1,
The operation corresponding to the risk level is,
At least one of shutting down the power of the target device, blocking the network connection of the target device, restarting the target device, and closing a background app running on the target device;
The target device is at least one of the electronic device and the other electronic device. 8. The method of claim 7,
A communication unit including a circuit; further comprising,
The processor is
When the risk level of the text is equal to or greater than the first threshold grade or greater than or equal to the first threshold, the communication unit is controlled to transmit the text and an operation corresponding to the risk level to a server managing a device corresponding to the text or the text An electronic device that provides a notification message of a situation corresponding to , and an operation corresponding to the risk level. 8. The method of claim 7,
The processor is
and when the danger level of the text is equal to or greater than a second threshold level or greater than or equal to a second threshold value, performing an operation of shutting down the power of the target device. According to claim 1,
The processor is
An electronic device for obtaining information about the text including a result of parsing the text by inputting the input text into a syntax analysis model trained to parse the text. According to claim 1,
a microphone; further comprising:
The processor is
When a user's voice inquiring about the state of the electronic device or another device is inputted through the microphone, the inputted user's voice is input to an Auto Speech Recognition (ASR) model to obtain a text corresponding to the user's voice. A method for controlling an electronic device, comprising:
When text is input, information on the input text is input to the first model trained to determine the semantic role of the sentence component included in the sentence to determine the semantic role of the sentence component included in the text step;
obtaining a risk level of the text by inputting a sentence component corresponding to the determined semantic range into a second model trained to output a risk level based on the semantic range of the sentence component included in the sentence; and
A control method comprising a; performing an operation corresponding to the risk level of the acquired text. 13. The method of claim 12,
The first model is
A control method, characterized in that it has been learned to determine each of the sentence components of the input text as one of an agent, a recipient, and a predicate. 14. The method of claim 13,
The second model is
When the sentence component determined as at least one of the subject and the target means a user or a device, the control method is characterized in that it is learned to increase the risk level of the text. 14. The method of claim 13,
The second model is
A control method, characterized in that it is learned to identify a word having a similar meaning to a sentence component determined as one of the subject, object, and predicate, and output the risk level of the text using a weight matched to the identified word. 13. The method of claim 12,
The second model is
A control method, characterized in that it is learned to output one of a plurality of grades classified according to the degree of risk as the risk level of the text. 13. The method of claim 12,
The second model is
A control method, characterized in that it is learned to output the risk level of the text as a numerical value indicating the degree of risk. 13. The method of claim 12,
The operation corresponding to the risk level is
At least one of shutting down the power of the target device, blocking the network connection of the target device, restarting the target device, and closing a background app running on the target device;
The target device is at least one of the electronic device and the other electronic device. 19. The method of claim 18,
The performing step is,
When the risk level of the text is greater than or equal to the first threshold grade or greater than the first threshold, the text and an operation corresponding to the risk level are transmitted to a server managing a device corresponding to the text or a situation corresponding to the text; and providing a notification message of an operation corresponding to the risk level. 19. The method of claim 18,
The performing step is,
and performing a power-off operation of the target device when the risk level of the text is greater than or equal to a second threshold level or greater than or equal to a second threshold. 13. The method of claim 12,
The step of receiving the input is
The method further comprising: inputting the input text into a syntax analysis model trained to parse the text to obtain information about the text including a result of the analysis of the text. 13. The method of claim 12,
When a user's voice inquiring about the status of the electronic device or other device is input through the microphone, inputting the inputted user's voice into an Auto Speech Recognition (ASR) model to obtain a text corresponding to the user's voice; control method further comprising.

Images