Electronic device and learning model determination method for learning of electronic device
An electronic device is provided. The electronic device includes a touch sensor, a processor, and a memory. The processor may determine a touch input from a user as at least one of a force-touch input or a long-touch input, based on received touch data, determine whether a result of determining the touch data matches an intention of the user, store data that does not match the intention of the user as a result of determination among the touch data in the memory, and determine a type of an artificial intelligence (AI)-based pre-learning model to be used in the electronic device, based on touch input accuracy and the data that does not match the intention of the user.
1. An electronic device comprising:
a touch sensor;
a processor configured to be operatively connected to the touch sensor; and
a memory,
wherein the processor is configured to, based on execution of instructions stored in the memory:
receive touch data corresponding to a touch input from a user by using the touch sensor,
determine the touch input from the user as at least one of a force-touch input or a long-touch input, based on the received touch data,
determine whether a result of determining the touch data matches an intention of the user,
store data that does not match the intention of the user as a result of determination among the touch data in the memory, and
determine a type of an AI-based pre-learning model to be used in the electronic device, based on touch input accuracy and the data that does not match the intention of the user.
2. The electronic device of claim 1 ,
wherein the touch input accuracy comprises a ratio by which the result of determining the touch input matches the intention of the user,
wherein the type of the AI-based pre-learning model comprises:
a first learning model in which a ratio of a trainable layer exceeds a first level,
a second learning model in which the ratio of the trainable layer is less than the first level and exceeds a second level, and
a third learning model in which the ratio of the trainable layer is less than the second level, and
wherein the first level has a relatively greater value than the second level.
3. The electronic device of claim 2 , wherein the processor is further configured to determine to use the first learning model, based on the touch input accuracy being less than a preset level and a variance value of the data stored in the memory being less than a preset level.
4. The electronic device of claim 2 , wherein the processor is further configured to determine to use the second learning model, based on the touch input accuracy exceeding a preset level.
5. The electronic device of claim 2 , wherein the processor is further configured to determine to use the third learning model, based on the touch input accuracy being less than a preset level and a variance value of the data stored in the memory exceeding a preset level.
6. The electronic device of claim 1 , wherein the processor is further configured to learn the touch data of the user, based on the determined pre-learning model.
7. The electronic device of claim 6 , wherein the processor is further configured to update an algorithm for determining the force-touch input and the long-touch input, based on a learning result.
8. The electronic device of claim 1 , wherein the processor is further configured to:
display a guide screen requesting the force-touch input and the long-touch input to the user;
generate first data comprising a strength and duration of the force-touch input, based on a user input to store the first data in the memory;
generate second data comprising a strength and duration of the long-touch input, based on a user input to store the second data in the memory; and
determine whether the result of determining the touch data matches the intention of the user, based on the first data and the second data.
9. The electronic device of claim 1 , wherein the processor is further configured to:
display an interface corresponding to the force-touch input on a display, based on determining a type of the touch input as the force-touch input;
determine that the force-touch input matches the intention of the user, based on the user touching the interface; and
determine that the force-touch input does not match the intention of the user, based on the user terminating the interface within a predetermined time instead of touching the interface.
10. The electronic device of claim 1 , wherein the processor is further configured to:
display an interface corresponding to the long-touch input on a display, based on determining a type of the touch input as the long-touch input;
determine that the long-touch input matches the intention of the user, based on the user touching the interface; and
determine that the long-touch input does not match the intention of the user, based on the user terminating the interface within a predetermined time without touching the interface.
11. A learning model determination method of an electronic device, the method comprising:
receiving touch data corresponding to a touch input from a user by using a touch sensor;
determining the touch input from the user as at least one of a force-touch input or a long-touch input, based on the received touch data;
determining whether a result of determining the touch data matches an intention of the user;
storing data that does not match the intention of the user as a result of determination among the touch data in a memory; and
determining a type of an AI-based pre-learning model to be used in the electronic device, based on touch input accuracy and the data that does not match the intention of the user.
12. The method of claim 11 ,
wherein the touch input accuracy comprises a ratio by which the result of determining the touch input matches the intention of the user,
wherein the AI-based pre-learning model comprises:
a first learning model in which a ratio of a trainable layer exceeds a first level,
a second learning model in which the ratio of the trainable layer is less than the first level and exceeds a second level, and
a third learning model in which the ratio of the trainable layer is less than the second level, and
wherein the first level has a relatively greater value than the second level.
13. The method of claim 12 , wherein the determining of the type of the AI-based pre-learning model to be used in the electronic device further comprises determining to use the first learning model, based on the touch input accuracy being less than a preset level and a variance value of the data stored in the memory being less than a preset level.
14. The method of claim 12 , wherein the determining of the type of the AI-based pre-learning model to be used in the electronic device further comprises determining to use the second learning model, based on the touch input accuracy exceeding a preset level.
15. The method of claim 12 , wherein the determining of the type of the AI-based pre-learning model to be used in the electronic device further comprises determining to use the third learning model, based on the touch input accuracy being less than a preset level and a variance value of the data stored in the memory exceeding a preset level.
16. The method of claim 11 , further comprising:
learning the touch data of the user, based on the determined pre-learning model.
17. The method of claim 16 , further comprising:
updating an algorithm for determining the force-touch input and the long-touch input, based on a learning result.
18. The method of claim 11 , wherein the determining of whether the result of determining the touch data matches the intention of the user further comprises:
displaying a guide screen requesting the force-touch input and the long-touch input to the user;
generating first data comprising a strength and duration of the force-touch input, based on a user input to store the first data in the memory;
generating second data comprising a strength and duration of the long-touch input, based on a user input to store the second data in the memory; and
determining whether the result of determining the touch data matches the intention of the user, based on the first data and the second data.
19. The method of claim 11 , wherein the determining of whether the result of determining the touch data matches the intention of the user further comprises:
displaying an interface corresponding to the force-touch input on a display, based on determining a type of the touch input as the force-touch input;
determining that the force-touch input matches the intention of the user, based on the user touching the interface; and
determining that the force-touch input does not match the intention of the user, based on the user terminating the interface within a predetermined time instead of touching the interface.
20. A non-transitory computer readable recording medium storing instructions, which when executed, perform operations of:
receiving touch data corresponding to a touch input from a user by using a touch sensor;
determining the touch input from the user as at least one of a force-touch input or a long-touch input, based on the received touch data;
determining whether a result of determining the touch data matches an intention of the user;
storing data that does not match the intention of the user as a result of determination among the touch data in a memory; and
determining a type of an AI-based pre-learning model to be used in an electronic device, based on touch input accuracy and the data that does not match the intention of the user.