Method and electronic device for predicting a response from context with a language model
An electronic device and a method for predicting a response are provided. The electronic device includes a display and a processor configured to receive at least one message, identify at least one contextual category of the at least one message, predict at least one response for the at least one message from a language model based on the at least one contextual category, and control the display to display the at least one predicted response.
1. An electronic device for automatically predicting a response, the electronic device comprising:
a display; and
a processor configured to:
receive, from at least one external device connected to the electronic device,
a first message and a second message;
control the display to display a user interface (UI) including the first message and the second message;
identify, based on a semantic modelling, a first dialog act category of the first message;
obtain at least one response for the first message based on the identified first dialog act category using a trained language model;
control the display to display the at least one obtained response for the first message on the UI;
in response to a touch input on the UI by a user for selecting the second message from among the displayed first message and the displayed second message being received while the at least one obtained response for the first message is displayed, identify, based on a semantic modelling, a dialog act category of the second message from among the displayed first message and the displayed second message;
obtain at least one response message for the second message from among the displayed plurality of messages based on the identified dialog act category of the selected message using trained language model; and
control the display to display the at least one obtained response message for the second message on the UI.
2. The electronic device of claim 1 , wherein the dialog act category of the second message is automatically identified based on at least one context indicative.
3. The electronic device claim 1 , wherein the second message comprises one of a topic selected from a written communication and a topic formed based on at least one input field available in an application.
4. The electronic device of claim 3 , wherein, the second message is an input topic from a first application, the processor is further configured to:
identify at least one contextual event associated with a second application, and
obtain at least one response message for the input topic from the first application based on the at least one contextual event.
5. The electronic device of claim 4 , wherein the at least one contextual event is a time bound event.
6. The electronic device of claim 4 , wherein the at least one contextual event associated with the second application is determined based on at least one context indicative associated with the input topic from the first application, and
wherein the at least one context indicative is determined based on at least one of content available in the input topic, context of the first application, user activities, and events defined.
7. The electronic device of claim 1 , wherein the second message and the at least one obtained response message are displayed within a notification area.
8. A method for automatically predicting a response, the method comprising:
receiving, by an information manager from at least one external device connected to an electronic device, a first message and a second message at the electronic device;
controlling a display of the electronic device to display a user interface (UI) including the first message and the second message;
identifying, based on a semantic modelling, a first dialog act category of the first message;
obtaining at least one response for the first message based on the identified first dialog act category using a trained language model;
controlling the display to display the at least one obtained response for the first message on the UI;
in response to a touch input on the UI by a user selecting the second message from among the displayed first message and the displayed second message being received, identifying, by a contextual category detector including a semantic modelling manager, a dialog act category of the second message from among the displayed first message and the displayed second message;
obtaining, by a response predictor, at least one response message for the second message from among the displayed plurality of messages based on the identified dialog act category of the second message using the trained language model; and
causing, by the response predictor, to display the at least one obtained response message for the second message on the UI.
9. The method of claim 8 , wherein the dialog act category of the second message is automatically identified based on at least one context indicative.
10. The method of claim 8 , wherein the second message comprises one of a topic selected from a written communication and a topic formed based at least one input field available in an application.