Abstract: The present invention provides artificial intelligence technology which has machine-learning-based information understanding capability, including metric learning providing improved classification performance, classification of an object considering a semantic relationship, understanding of the meaning of a scene based on the metric learning and the classification, and the like. An electronic device according to one embodiment of the present invention comprises a memory in which at least one instruction is stored, and a processor for executing the stored instruction. Here, the processor extracts feature data from training data of a first class, obtains a feature point by mapping the extracted feature data to an embedding space, and makes an artificial neural network learn in a direction for reducing a distance between the obtained feature point and an anchor point.

Abstract: Provided is a method, performed by a device, of recognizing content, the method including selecting a template to be used for recognizing content displayed on the device when a type of a content service that provides the content to the device is determined; capturing a screen of the device when a control signal is received by the device; determining whether the captured screen corresponds to a gateway page; determining whether the captured screen corresponds to the selected template when the captured screen corresponds to the gateway page; and generating a template corresponding to the content service, based on the captured screen, when the captured screen does not correspond to the selected template.

Abstract: Provided is a method of providing information about content, performed by a mobile device, the method including detecting an electronic device displaying the content, determining whether at least one search application is executed on the mobile device, as the electronic device is detected, transmitting a request for recognition information indicating at least one object related to the content to the electronic device, as the at least one search application is executed, and receiving the requested recognition information from the electronic device.

Abstract: Provided is a method of recognizing text in a terminal, the method including generating first tag information about a kind of language set in a user interface (UI) for inputting text and a location of a cursor at a time point when a text input has started, when the UI for inputting text displayed on the terminal is executed; when a language switch request that requests the terminal to switch the kind of language set in the UI is received, generating second tag information about a kind of switched language and a location of the cursor at a time point of receiving the language switch request; when the text input is finished, storing a screen image of the terminal; and recognizing the text input to the terminal based on at least one piece of tag information and the screen image.

Abstract: Provided is a method, performed by a device, of recognizing content, the method including selecting a template to be used for recognizing content displayed on the device when a type of a content service that provides the content to the device is determined; capturing a screen of the device when a control signal is received by the device; determining whether the captured screen corresponds to a gateway page; determining whether the captured screen corresponds to the selected template when the captured screen corresponds to the gateway page; and generating a template corresponding to the content service, based on the captured screen, when the captured screen does not correspond to the selected template.

Abstract: The present invention provides artificial intelligence technology which has machine-learning-based information understanding capability, including metric learning providing improved classification performance, classification of an object considering a semantic relationship, understanding of the meaning of a scene based on the metric learning and the classification, and the like. An electronic device according to one embodiment of the present invention comprises a memory in which at least one instruction is stored, and a processor for executing the stored instruction. Here, the processor extracts feature data from training data of a first class, obtains a feature point by mapping the extracted feature data to an embedding space, and makes an artificial neural network learn in a direction for reducing a distance between the obtained feature point and an anchor point.

Abstract: Provided is a method of recognizing text in a terminal, the method including generating first tag information about a kind of language set in a user interface (UI) for inputting text and a location of a cursor at a time point when a text input has started, when the UI for inputting text displayed on the terminal is executed; when a language switch request that requests the terminal to switch the kind of language set in the UI is received, generating second tag information about a kind of switched language and a location of the cursor at a time point of receiving the language switch request; when the text input is finished, storing a screen image of the terminal; and recognizing the text input to the terminal based on at least one piece of tag information and the screen image.

Abstract: Provided is a method of providing information about content, performed by a mobile device, the method including detecting an electronic device displaying the content, determining whether at least one search application is executed on the mobile device, as the electronic device is detected, transmitting a request for recognition information indicating at least one object related to the content to the electronic device, as the at least one search application is executed, and receiving the requested recognition information from the electronic device.

Abstract: A device is provided, wherein the device captures a screen the device when a control signal for controlling one or more pieces of content provided to the device is received, and, when a captured screen corresponds to a template screen, extracts a character string including content information from a preset area of the captured screen, compares the extracted character string with one or more pieces of text included in a preset meaning recognition model, detects text corresponding to the content information, and recognizes content displayed on the screen of the device based on the detected text.