Trained model creation method for performing specific function for electronic device, trained model for performing same function, exclusive chip and operation method for the same, and electronic device and system using the same
A trained model creation method for performing a specific function for an electronic device includes preparing big data for training an artificial neural network and specific function performance determination data for determining whether to perform a specific function of an electronic device with respect to the sensing data; and preparing an artificial neural network model, which calculates inputs of the sensing data for the nodes of the input layer in order to output the specific function performance determination data from the nodes of the output layer. The artificial neural network model is trained by repeatedly performing a process of inputting the sensing data included in the prepared big data into the nodes of the input layer and outputting the specific function performance determination data that pairs with the sensing data included in the big data from the nodes of the output layer so as to update the association parameters.
1. An electronic device for performing a specific function according to a trained model created by a machine learning device performing machine learning on an artificial neural network model to generate the trained model, the electronic device comprising:
a sensing data generation unit configured to generate at least one sensing data;
a dedicated artificial intelligence (AI) acceleration processor configured to
generate wake-up data based on the at least one sensing data using the trained model,
wherein the trained model is embedded in the dedicated AI acceleration processor and is configured to output determination data of performing the specific function, the determination data including the wake-up data to switch from a first mode to a second mode using the at least one sensing data as input to the dedicated AI acceleration processor during the first mode, wherein the first mode is one of a stop mode, a sleep mode, or a lock mode of the electronic device, and wherein the second mode is one of a booting mode, an activation mode, or an unlock mode of the electronic device;
a control unit configured to generate a control command based on the wake-up data;
a second function unit including an output unit and a driver, the output unit including a display, the driver being driven based on the control command; and
a power source unit configured to
supply power to the sensing data generation unit and the dedicated AI acceleration processor while supplying no power to the control unit and supplying no power to the second function unit during the first mode, and
supply power to the control unit and the second function unit during the second mode.
2. The electronic device of claim 1 ,
wherein the electronic device is a home appliance including the display of the second function unit, and
wherein the wake-up data includes data to turn on the display.
3. The electronic device of claim 1 ,
wherein the electronic device is a vehicle, and
wherein the wake-up data includes data to start an engine of the vehicle.
4. The electronic device of claim 1 ,
wherein the trained model is generated using the artificial neural network model,
wherein the artificial neural network model includes
nodes of an input layer through which voice data or image data is inputted,
nodes of an output layer through which the determination data of performing the specific function is outputted, and
association parameters between the nodes of the input layer and the nodes of the output layer,
wherein the artificial neural network model outputs the determination data of performing the specific function from the nodes of the output layer in response to an input of the at least one sensing data for the nodes of the input layer, and
wherein the association parameters are updated to train the artificial neural network model by repeatedly performing a process of
inputting sensing data included in big data into the nodes of the input layer and
outputting determination data of performing the specific function included in the big data that pairs with the sensing data included in the big data from the nodes of the output layer.
5. The electronic device of claim 1 ,
wherein the electronic device is a vehicle, and
wherein the determination data of performing the specific function includes data for determining one of
rear window heater operation in the vehicle,
front window defroster operation in the vehicle,
air conditioner or heater operation in the vehicle,
wiper operation in the vehicle,
lights operation in the vehicle,
audio, visual, and navigation (AVN) device operation in the vehicle, and
voice assistant call operation in the vehicle.
6. The electronic device of claim 1 , further comprising:
a first function unit that is an always-on module turned on even when the electronic device is turned off.
7. The electronic device of claim 6 , wherein the first function unit includes a communication unit.
8. The electronic device of claim 6 ,
wherein the first function unit is turned on, and
wherein the second function unit is turned off to reduce power consumption and then is turned on when the control command is received from the control unit.
9. A specific function performing processor embeddable in an electronic device for performing a specific function according to a trained model created by a machine learning device performing machine learning on an artificial neural network model to generate the trained model, the specific function performing processor comprising:
a dedicated artificial intelligence (AI) acceleration processor configured to
generate wake-up data for generating a control command in a control unit based on sensing data generated by a sensing data generator, using the trained model,
wherein the trained model is embedded in the dedicated AI acceleration processor and is configured to output determination data of performing the specific function, the determination data including the wake-up data to switch from a first mode to a second mode using the sensing data as input to the dedicated AI acceleration processor during the first mode, wherein the first mode is one of a stop mode, a sleep mode, or a lock mode of the electronic device, and wherein the second mode is one of a booting mode, an activation mode, or an unlock mode of the electronic device,
wherein power is supplied to the sensing data generation unit and the dedicated AI acceleration processor while no power is supplied to the control unit during the first mode, and
wherein the power is supplied to the control unit during the second mode.
10. The specific function performing processor of claim 9 ,
wherein the trained model is generated using the artificial neural network model,
wherein the artificial neural network model includes
nodes of an input layer through which voice data or image data is inputted,
nodes of an output layer through which the determination data of performing the specific function is outputted, and
association parameters between the nodes of the input layer and the nodes of the output layer,
wherein the artificial neural network model outputs the determination data of performing the specific function from the nodes of the output layer in response to an input of the at least one sensing data for the nodes of the input layer, and
wherein the association parameters are updated to train the artificial neural network model by repeatedly performing a process of
inputting sensing data included in big data into the nodes of the input layer and
outputting determination data of performing the specific function included in the big data that pairs with the sensing data included in the big data from the nodes of the output layer.