User adaptive conversation apparatus and method based on monitoring of emotional and ethical states
A user adaptive conversation apparatus generating a talk for a conversation based on emotional and ethical states of a user. A voice recognition unit converts a talk of the user in a conversational situation into a natural language script form to generate talk information. An artificial visualization unit generates situation information by recognizing talking situation from a video and generates intention information indicating an intention of the talk. A natural language analysis unit converts the situation information and the intention information into the natural language script form. A natural language analysis unit analyzes the talk information, the intention information, and the situation information. A conversation state tracing unit generates current talk state information representing a meaning of the talk information by interpreting the talk information according to the intention information and the situation information, and determines next talk state information including candidate responses corresponding to the current talk status information.
1. A user adaptive conversation apparatus, comprising:
a voice recognition unit configured to convert a talk of a user in a conversational situation into a natural language script form to generate talk information;
an artificial visualization unit having an artificial neural network and configured to:
recognize, by using the artificial neural network, a facial expression of the user from a video acquired in the conversational situation;
generate situation information indicating a talking situation based on the recognized facial expression of the user;
determine an intention of the talk based on the situation information; and
generate intention information indicating the intention of the talk;
a natural language analysis unit configured to convert the situation information and the intention information into the natural language script form;
a natural language analysis unit configured to perform a natural language analysis for the talk information, the intention information, and the situation information;
a conversation state tracing unit configured to generate current talk state information representing a meaning of the talk information by interpreting the talk information according to the intention information and the situation information, and determine next talk state information that includes a plurality of candidate responses corresponding to the current talk status information;
an emotion tracing unit configured to generate emotion state information indicating an emotional state of the user based on the talk information, the intention information, and the situation information;
an ethic analysis unit configured to generate ethical state information indicating ethics of the conversation based on the talk information, the intention information, and the situation information; and
a multi-modal conversation management unit configured to select one of the plurality of candidate responses according to at least one of the emotion state information and the ethical state information to determine final next talk state information including a selected response.
2. The user adaptive conversation apparatus of claim 1 , further comprising:
a natural language generation unit configured to convert the final talk state information into an output conversation script having the natural language script form; and
an adaptive voice synthesizing unit configured to synthesize a voice signal in which an intonation and conforming to at least one of the emotion state information, the situation information, and the intention information is given to the output conversation script.
3. A user adaptive conversation method, comprising:
converting a talk of a user in a conversational situation into a natural language script form to generate talk information;
recognizing, by using an artificial neural network, a facial expression of the user from a video acquired in the conversational situation;
generating situation information indicating a talking situation based on the recognized facial expression of the user;
determining an intention of the talk based on the situation information;
generating intention information indicating the intention of the talk;
converting the situation information and the intention information into the natural language script form;
performing a natural language analysis for the talk information, the intention information, and the situation information;
generating current talk state information representing a meaning of the talk information by interpreting the talk information according to the intention information and the situation information;
determining next talk state information that includes a plurality of candidate responses corresponding to the current talk status information;
generating emotion state information indicating an emotional state of the user based on the talk information, the intention information, and the situation information;
generating ethical state information indicating ethics of the conversation based on the talk information, the intention information, and the situation information; and
selecting one of the plurality of candidate responses according to at least one of the emotion state information and the ethical state information to determine final next talk state information including a selected response.
4. The user adaptive conversation method of claim 3 , further comprising:
converting the final talk state information into an output conversation script having the natural language script form; and
synthesizing a voice signal in which an intonation and conforming to at least one of the emotion state information, the situation information, and the intention information is given to the output conversation script.