Electronic device and method for controlling thereof
An electronic device and a control method thereof are provided. The electronic device includes a memory storing instructions; and a processor configured to execute the instructions to: based on a text being input, determine semantic roles of sentence components included in the text by inputting information on the text to a first model trained to determine the semantic roles of the sentence components; obtain a risk level of the text by inputting the sentence components with the determined semantic roles of the sentence components to a second model trained to output the risk level of the text based on the semantic roles of the sentence components; and perform an operation corresponding to the obtained risk level of the text.
1. An electronic device comprising:
a memory storing instructions; and
at least one processor configured to execute the instructions to:
obtain an inquiry from a user of a target device, the inquiry being formatted as a text;
determine semantic roles of sentence components in the text by inputting information on the text to a first model trained to determine the semantic roles of the sentence components, wherein the first model is trained to determine each of the sentence components of the text as one of an agent, a recipient, and a predicate;
input the sentence components to a second model trained with a reduced amount of training data to:
output a risk level of the text based on the semantic roles of the sentence components; and
based on a sentence component determined as at least one of the agent and the recipient being the user or the target device, increase the risk level of the text;
identify an unknown sentence component of the sentence components that is not recognized by the second model;
obtain a synonym of the unknown sentence component from a database;
obtain the risk level of the text by inputting the semantic roles and sentence components, with the synonym in place of the unknown sentence component, to the second model;
identify a dangerous situation based on the risk level of the text being equal to or higher than a threshold; and
based on the dangerous situation being identified, perform an operation corresponding to the risk level of the text among a plurality of operations, the plurality of operations comprising performing a power-off of the target device, blocking network connection of the target device, re-starting of the target device, and turning off a background application which is running at the target device.
2. The electronic device according to claim 1 , wherein the second model is trained to output the risk level of the text by using a weight value matching to the synonym.
3. The electronic device according to claim 1 , wherein the second model is trained to output one of a plurality of grades corresponding to a degree of risk as the risk level of the text.
4. The electronic device according to claim 1 , wherein the second model is trained to output the risk level of the text as a value representing a degree of risk.
5. The electronic device according to claim 1 , wherein the target device is at least one of the electronic device or another electronic device.
6. The electronic device according to claim 1 , further comprising:
a communicator comprising circuitry,
wherein the at least one processor is further configured to execute the instructions to, based on the risk level of the text being equal to or higher than a threshold grade or a threshold value:
control the communicator to transmit the text and the operation corresponding to the risk level to a server managing the target device, or
provide an alert message regarding a situation corresponding to the text and the operation corresponding to the risk level.
7. The electronic device according to claim 1 , wherein the at least one processor is further configured to execute the instructions to, based on the risk level of the text being equal to or higher than a second threshold grade or a second threshold value, perform a power-off operation of the target device.
8. The electronic device according to claim 1 , wherein the at least one processor is further configured to execute the instructions to obtain the information on the text comprising a result of sentence parsing the text by inputting the text to a sentence parsing model trained to perform the sentence parsing.
9. The electronic device according to claim 1 , further comprising:
a microphone,
wherein the at least one processor is further configured to execute the instructions to, based on a user voice input inquiring for a state of the target device being received via the microphone, inputting the user voice input to an auto speech recognition (ASR) model to obtain the text.
10. A method for controlling an electronic device, the method comprising, using at least one processor:
obtain an inquiry from a user of a target device, the inquiry being formatted as a text;
determining semantic roles of sentence components in the text by inputting information on the text to a first model trained to determine the semantic roles of the sentence components, wherein the first model is trained to determine each of the sentence components of the text as one of an agent, a recipient, and a predicate;
inputting the sentence components to a second model trained with a reduced amount of training data to:
output a risk level of the text based on the semantic roles of the sentence components; and
based on a sentence component determined as at least one of the agent and the recipient being the user or the target device, increase the risk level of the text;
identifying an unknown sentence component of the sentence components that is not recognized by the second model;
obtaining a synonym of the unknown sentence component from a database;
obtaining the risk level of the text by inputting the semantic roles and sentence components, with the synonym in place of the unknown sentence component, to the second model;
identifying a dangerous situation based on the risk level of the text being equal to or higher than a threshold; and
based on the dangerous situation being identified, performing an operation corresponding to the risk level of the text among a plurality of operations, the plurality of operations comprising performing a power-off of the target device, blocking network connection of the target device, re-starting of the target device, and turning off a background application which is running at the target device.
11. The method according to claim 10 , wherein the second model is trained to output the risk level of the text by using a weight value matching to the synonym.