Electronic device for processing utterance, operating method thereof, and storage medium

A method for processing an utterance in an electronic device, the method comprising:

• acquiring utterance data corresponding to a voice of a user through a microphone of the electronic device;
• acquiring first input text by speech recognition applied to the utterance data;
• identifying a named entity based on the first input text;
• determining, based on the utterance data and using the named entity and a first intent determination model, an intent;
• detecting that the intent determination has resulted in a failure;
• responsive to the failure, identifying whether a length of the named entity of the utterance data is less than a predetermined length;
• identifying the first input text as a text not including a predicate, based on identifying that the length of the named entity of the utterance data is less than the predetermined length, wherein the length is measured in characters or words;
• and
• generating one or more second intent determination models based on the first input text not including the predicate.

A non-transitory storage medium storing instructions, wherein the instructions are configured to, when executed by a processor of an electronic device, cause the electronic device to perform at least one operation, the at least one operation comprising:

• acquiring utterance data corresponding to a voice of a user through a microphone of the electronic device;
• acquiring first input text by applying speech recognition on the utterance data;
• identifying a named entity based on the first input text;
• determining, based on the utterance data and using the named entity and a first intent determination model, an intent;
• detecting that the intent determination has resulted in a failure;
• responsive to the failure, identifying whether a length of the named entity of the utterance data is less than a predetermined length;
• identifying the first input text as a text not including a predicate, based on identifying that the length of the named entity of the utterance data is less than the predetermined length, wherein the length is measured in characters or words;
• and
• based on the first input text not including the predicate, generating one or more second intent determination models based on the first input text.