- Number
- 20250225986
- Published
- 2025-07-10
- Filed
- 2024-01-04
- Assignee
- Disney Enterprises, Inc.
- Inventors
- Kennedy; James R. et al.
- CPC
- G10L15/222; G06N3/006; G06N3/0455; G10L13/10; G10L15/1807; G10L15/22
- Verdict
- Low Notable software
- Source
- Google Patents · FreePatentsOnline
The keeper's note
Interruption-response behavior technique for AI interactive character.
Abstract
A system includes a hardware processor, a memory storing software code, and a machine learning (ML) model trained to detect an interruption to a conversation. The system detects, during a conversational turn by an artificial intelligence (AI) character in interaction with a human and/or another AI character, sound produced by the human and/or the other AI character, classifies, using the ML model, the sound as an interruption or irrelevant to the interaction. When the sound is irrelevant to the interaction, the conversational turn of the AI character continues. When the sound is an interruption, the system identifies a response strategy for continuing the interaction, and executes the response strategy including at least one of: (i) retention of the conversational turn, (ii) relinquishment of the conversational turn, or (iii) a negotiation, with the human and/or the other AI character, to determine the retention or the relinquishment of the conversational turn.
Background
BACKGROUND
A characteristic feature of most human interactions is spontaneity, including how individuals in conversation react to interruptions to their speech. In order for a non-human social agent, such as one embodied as an artificial intelligence (AI) interactive character, for example, to engage in conversation in a naturalistic human-like way, it is desirable to provide the non-human social agent with the ability to respond appropriately when interrupted.
Conventional solutions for providing speech for a non-human social agent may be unable to detect an interruption to that speech, or lack sophistication in their response when one is detected. For example, when the interruption goes undetected, it is simply ignored and the social agent continues its speech, i.e., “talking over” the interruption. Alternatively, a conventional response to an interruption to speech by a non-human social agent that is detected includes stopping the speech entirely when the interruption occurs. These responses may be appropriate in certain circumstances but not others. Unfortunately, both of those conversational scenarios may result in an awkward and unnatural experience for a human interacting with the social agent, and particularly in the case in which the interruption goes undetected and the social agent talks over the interruption, which may undesirably make the human feel ignored. Consequently, there is a need in the art for a solution enabling a non-human social agent to
Claims
1. A system comprising: a hardware processor; a system memory storing a software code; and a machine learning (ML) model trained to detect an interruption to a conversation by a participant in the conversation; the hardware processor configured to execute the software code to: detect, during a conversational turn by a first artificial intelligence (AI) character in an interaction with at least one of a human or a second AI character, sound produced by the at least one of the human or the second AI character; classify, using the ML model, the sound as one of an interruption to the interaction by the at least one of the human or the second AI character, or as irrelevant to the interaction; when the sound is classified as irrelevant to the interaction: continue, using the first AI character, the conversational turn of the first AI character; when the sound is classified as the interruption to the interaction: identify a response strategy for continuing the interaction; and execute the identified response strategy including at least one of: (i) retention, by the first AI character, of the conversational turn, (ii) relinquishment, by the first AI character, of the conversational turn, or (iii) a negotiation, with the at least one of the human or the second AI character, to determine the retention or the relinquishment of the conversational turn by the first AI character. ||
8. A method for use by a system including a hardware processor and a system memory, the system memory storing a software code and a machine learning (ML) model trained to detect an interruption to a conversation by a participant in the conversation, the method comprising: detecting, by the software code executed by the hardware processor, during a conversational turn by a first artificial intelligence (AI) character in an interaction with at least one of a human or a second AI character, sound produced by the at least one of the human or the second AI character; classifying, by the software code executed by the hardware processor and using the ML model, the sound as one of an interruption to the interaction by the at least one of the human or the second AI character, or as irrelevant to the interaction; when the sound is classified as irrelevant to the interaction: continuing, by the software code executed by the hardware processor and using the first AI character, the conversational turn of the first AI character; when the sound is classified as the interruption: identifying, by the software code executed by the hardware processor, a response strategy for continuing the interaction; and executing, by the software code executed by the hardware processor, the identified response strategy including at least one of: (i) retention, by the first AI character, of the conversational turn, (ii) relinquishment, by the first AI character, of the conversational turn, or (iii) a negotiation, with the at least one of the human or the second AI character, to determine the retention or the relinquishment of the conversational turn by the first AI character. ||
15. A computer-readable non-transitory medium having stored thereon instructions, which when executed by a hardware processor, instantiate a method comprising: detecting during a conversational turn by a first artificial intelligence (AI) character in an interaction with at least one of a human or a second AI character, sound produced by the at least one of the human or the second AI character; classifying using a machine learning (ML) model trained to detect an interruption to a conversation by a participant in the conversation, the sound as one of an interruption by the at least one of the human or the second AI character, or as irrelevant to the interaction; when the sound is classified as irrelevant to the interaction: continuing, using the first AI character, the conversational turn of the first AI character; when the sound is classified as the interruption: identifying a response strategy for continuing the interaction; and executing the identified response strategy including at least one of: (i) retention, by the first AI character, of the conversational turn, (ii) relinquishment, by the first AI character, of the conversational turn, or (iii) a negotiation, with the at least one of the human or the second AI character, to determine the retention or the relinquishment of the conversational turn by the first AI character.