Memories for virtual characters

Number: 12718481
Published: 2026-08-25
Filed: 2024-05-22
Assignee: Disney Enterprises, INC.
Inventors: Doggett; Erika Varis, Landwehr; Fabian Jonas, Weber; Romann Matthew
CPC: G06F40/56; G06T17/00; G06N3/006; G06N20/00; G06T13/40
Verdict: Medium Notable software
First reported: 2026-W36 (2026-09-05)
Source: Google Patents · FreePatentsOnline

The keeper's note

Gives a virtual character persistent memories of past interactions so later encounters with a guest reference them.

Abstract

One embodiment of the present invention sets forth a technique for synthesizing an interaction with a virtual character. The technique includes matching a first message from a user to a first set of memories associated with the virtual character and determining at least a portion of the first set of memories based on a plurality of factors associated with the first set of memories. The technique also includes inputting a first prompt that includes (i) one or more instructions associated with the virtual character, (ii) the at least a portion of the first set of memories, and (iii) the first message into a machine learning model. The technique further includes generating, via execution of the machine learning model based on the first prompt, a first response by the virtual character to the first message, and causing the first response to be outputted to the user.

Background

BACKGROUND Field of the Various Embodiments (1) Embodiments of the present disclosure relate generally to machine learning and generative models and, more specifically, to memories for virtual characters. DESCRIPTION OF THE RELATED ART (2) Virtual characters have become an important part of many interactive media experiences, such as (but not limited to) video games, virtual reality, interactive robots, and/or chatbots. These experiences often involve interactions between the virtual characters and users in a conversational manner. Traditionally, such interactions have been driven by scripted dialogue trees and/or dialogue flows written by designers, with key phrases inserted into messages from the virtual characters to personalize the messages to the users. However, virtual characters that utilize dialogue trees and/or dialogue flows are limited to a predetermined set of conversational topics and unable to improvise, which can lead to interactions that feel robotic. (3) More recently, large language models (LLMs) and/or other types of generative models have been incorporated into interactions with virtual characters. These generative models are capable of generating text that is contextually relevant to user input, thereby allowing for more natural, dynamic, and engaging conversations. However, generative models are trained on large diverse datasets and lack grounding in a consistent character background and/or persona, which can cause the generative models to “hallucinate”

Claims

1. A computer-implemented method for synthesizing an interaction with a virtual character, the method comprising: performing, via one or more processors, the steps of: matching a first message from a user to a first set of memories associated with the virtual character; determining at least a portion of the first set of memories based on a plurality of factors associated with the first set of memories; inputting a first prompt that includes (i) one or more instructions associated with the virtual character, (ii) the at least a portion of the first set of memories, and (iii) the first message into a machine learning model; generating, via execution of the machine learning model based on the first prompt, a first response by the virtual character to the first message; and causing the first response to be outputted to the user. || 11. One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of: matching a first message from a user to a first set of memories associated with a virtual character; generating a first ranking of the first set of memories based on one or more factors associated with the first set of memories; determining at least a portion of the first set of memories based on the first ranking; inputting a first prompt that includes (i) one or more instructions associated with the virtual character, (ii) the at least the portion of the first set of memories, and (iii) the first message into a machine learning model; generating, via execution of the machine learning model based on the first prompt, a first response by the virtual character to the first message; and causing the first response to be outputted to the user. || 20. A system, comprising: one or more memories that store instructions, and one or more processors that are coupled to the one or more memories and, when executing the instructions, are configured to perform the steps of: matching a first message from a user to a set of memories associated with a virtual character; determining at least a portion of the set of memories based on a ranking of the set of memories by a set of scores associated with the set of memories; inputting a first prompt that includes (i) one or more instructions associated with the virtual character, (ii) the at least a portion of the set of memories, and (iii) the first message into a machine learning model; generating, via execution of the machine learning model based on the first prompt, a first response by the virtual character to the first message; and causing the first response to be outputted to the user.