Knowledge Graph Valence Modeling for Dynamic Digital Characters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing representations of characters, whether in written works, interactive digital characters, or chatbots, lack the depth and dynamism of human interactions due to static personality depictions and limited emotional complexity, leading to inadequate portrayal of nuanced emotional states and behaviors.
Innovation Solution
A system and method utilizing knowledge graphs with dynamic valence scores to represent characters, where valence curves capture emotional associations over time, enabling sophisticated and interactive character portrayals through valence-enabled knowledge graphs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If static personality depictions are used in character representations, then the system is simple and easy to implement, but the emotional depth and dynamism of character interactions are insufficient
Solution Approach 1:
The patent transforms static character personality representations into dynamic ones by introducing temporal dimensions. Character traits are no longer fixed values but evolve over time through experiences, allowing the system to capture emotional depth and behavioral changes while maintaining manageable complexity through structured temporal modeling.
Solution Approach 2:
The patent adds a temporal dimension to character representations by introducing timelines and historical contexts. This transforms one-dimensional static trait descriptions into multi-dimensional dynamic models that incorporate when and how traits change, enabling sophisticated emotional portrayals without requiring exponential increases in system complexity.
2Ease of operation
If logic-driven programming with fixed storylines is used for interactive digital characters, then the system is easy to control, but the characters cannot reflect the depth and dynamism of actual human interactions
Solution Approach 1:
The patent enables characters to autonomously generate their own narratives and responses based on their internal state models and historical experiences. Rather than relying on pre-programmed storylines, characters self-determine their actions and reactions, creating authentic interactions that reflect human-like decision-making while maintaining systematic control through defined behavioral frameworks.
Solution Approach 2:
The patent transforms fixed storyline parameters into dynamic parameters that change based on character states, experiences, and contextual factors. This allows the system to maintain ease of control through structured parameter management while achieving authenticity through continuous parameter adaptation based on simulated psychological processes.
3Adaptability or versatility
If large language models with narrative devices are used to approximate emotional states, then the system can generate varied responses, but it lacks a deep and realistic personality that informs simulated emotions
Solution Approach 1:
The patent introduces an intermediary layer between the language model and the generated responses. This intermediary consists of structured character models that include personality traits, historical experiences, and emotional states. These intermediaries process and filter language model outputs to ensure they are consistent with the character's defined personality, preventing information loss while maintaining response variety.
Solution Approach 2:
The patent performs preliminary actions by pre-defining comprehensive character profiles including personality structures, historical timelines, and emotional frameworks before generating interactions. This preliminary structuring ensures that subsequent language model responses are grounded in deep personality information, preventing loss of character depth while enabling varied emotional expressions.
Data Source
AI summary
The present disclosure provides a system and method for creating digital characters using dynamic valence scores applied to knowledge graphs. The system is comprised of at least a knowledge graph interface to manage nodes and edges, as well as a valence score generator and a valence curve modeller. At any selected point in a character's timeline, the system outputs textual or numeric representations of the positivity and intensity of associations held within the graph. Meanwhile, the method to create a digital character comprises creating the nodes and edges in the graph network interface, applying the curve modeller to generate dynamic valence scores over time, and utilizing a network interrogator to convert the valence scores into textual or numeric representations of the digital character's affective state at a relative point in time.