Conversational AI system for interactive characters/agents.
Systems and methods for an interactive communications system capable of generating a response to conversational input are provided. The interactive communications system analyzes the conversational input to determine relevant topics of discussion. The interactive communications system further determines which of the relevant topics of discussion can potentially lead to an unwanted end to a conversation. The interactive communications system redirects the conversation by providing responses to the conversational input that are intended simply to avoid the unwanted end to the conversation.
TECHNICAL FIELD (1) The present disclosure relates generally to interactive communications systems. More particularly, some embodiments of the present disclosure are directed to providing responses conversational inputs designed to engage a human participant in conversation. DESCRIPTION OF THE RELATED ART (2) Chatbots refer to computer programs that can conduct a conversation with a human counterpart or participant. Such computer programs are designed to respond to textual and/or auditory (spoken) inputs in a way that attempts to convincingly simulate an actual human-to-human interaction. Other interactive communication systems use interactive agents that may be automatons, puppets, or human actors. These interactive agents engage a human participant in a conversation and can be used as virtual assistants, for entertainment purposes, as well as for promoting products and/or services. Such interactive communications systems are often provided with pre-scripted responses to some conversational input. The pre-scripted responses can be selected using computer programs focused on achieving a particular end-goal, such as engaging a human participant in an interactive story, and guiding the participant through the interactive story until it ends. BRIEF SUMMARY OF THE DISCLOSURE (3) In accordance with one embodiment, a method comprises receiving conversational input, and determining a likelihood that the conversational input comprises a current conversational topic that will terminat
1. A method, comprising: receiving conversational input associated with a conversation; determining, based on a tree structure comprising a plurality of branches representing a respective plurality of conversational topics, and based on a count of a number of nodes remaining in a first branch of the plurality of branches that corresponds to a current conversational topic of the conversational input, a likelihood that the current conversational topic will terminate earlier than desired, wherein each of the plurality of branches comprises a respective plurality of nodes linked by the tree structure; based on the determined likelihood, seeking out opportunities to segue the current conversational topic to one or more alternative conversational topics represented by one or more of the nodes and one or more branches of the tree structure other than the first branch, wherein seeking out the opportunities comprises: predicting an end to the current conversational topic based on the conversational input and the tree structure; and analyzing the conversational input for cues associated with the one or more alternative conversational topics; selecting a first alternative conversational topic of the one or more alternative conversational topics, the first alternative conversational topic associated with a second branch of the plurality of branches; in response to selecting the first alternative conversational topic, switching from the first branch to the second branch; and selecting, based on the second branch, one or more conversational responses associated with the first alternative conversational topic to the conversational input. ||
16. A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to perform a method comprising: receiving conversational input associated with a conversation; determining, based on a tree structure comprising a plurality of branches representing a respective plurality of conversational topics, and based on a count of a number of nodes remaining in a first branch of the plurality of branches that corresponds to a current conversational topic of the conversational input, a likelihood that the current conversational topic will terminate earlier than desired, wherein each of the plurality of branches comprises a respective plurality of nodes linked by the tree structure; based on the determined likelihood, seeking out opportunities to segue the current conversational topic to one or more alternative conversational topics represented by one or more of the nodes and one or more branches of the tree structure other than the first branch, wherein seeking out the opportunities comprises: predicting an end to the current conversational topic based on the conversational input and the tree structure; and analyzing the conversational input for cues associated with the one or more alternative conversational topics; selecting a first alternative conversational topic of the one or more alternative conversational topics, the first alternative conversational topic associated with a second branch of the plurality of branches; in response to selecting the first alternative conversational topic, switching from the first branch to the second branch; and selecting, based on the second branch, one or more conversational responses associated with the first alternative conversational topic to the conversational input.