Virtual Agent Behavior Generation Using Affective Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current virtual agent systems struggle to dynamically adjust their behavior to match the emotion, mood, and personality of users during conversations, often resulting in inappropriate or mismatched responses due to pre-assigned behaviors and limited context analysis.
Innovation Solution
A method for automatically generating facial expressions, body gestures, and vocal expressions for virtual agents based on emotion, mood, and personality vectors derived from user interactions, using a pleasure, arousal, and dominance model to create Behavior Markup Language (BML) that adapts in real-time to user affective context.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If pre-assigned behaviors are used for virtual agent utterances, then the implementation is simple and fast, but the behavior becomes inappropriate for different conversation contexts and user emotional states
Solution Approach 1:
The patent implements dynamic behavior generation by computing emotion vectors, mood vectors, and personality vectors in real-time during conversations. The virtual agent's behavior is continuously adapted based on the user's emotional state, mood, and personality traits rather than using static pre-assigned behaviors. This allows the system to respond appropriately to different conversation contexts while maintaining computational efficiency through vector-based representations.
2Adaptability or versatility
If behavior is dynamically generated based on user affect, then the behavior appropriateness improves, but the system complexity increases
Solution Approach 1:
The patent manages complexity by representing complex user states as compact parameter vectors: emotion vectors (capturing emotional dimensions), mood vectors (representing overall affective state), and personality vectors (encoding stable traits). These parameterized representations allow dynamic behavior generation through mathematical operations on vectors rather than complex rule-based systems, reducing computational overhead while maintaining adaptability.
Solution Approach 2:
The patent creates a universal framework where emotion vectors, mood vectors, and personality vectors can be computed from multiple input sources (text analysis, speech patterns, facial expressions) and applied across different conversation domains. This multi-functional approach allows the same vector computation mechanism to serve various affective computing needs without requiring domain-specific implementations.
3Productivity
If content analysis based solely on words is used, then the processing is simple and fast, but the understanding of user utterance meaning becomes flawed due to ignoring tone and context
Solution Approach 1:
The patent merges multiple analysis dimensions by combining word content analysis with tone analysis (from speech patterns), facial expression analysis, and conversation context analysis. These different modalities are integrated into unified emotion vectors that capture the complete affective meaning of user utterances, allowing the system to understand sarcasm, irony, and emotional nuance that would be missed by text-only analysis.
Solution Approach 2:
The patent introduces emotion vectors as intermediary representations that bridge raw input data (text, speech, facial expressions) and behavioral output. These vectors serve as a compact intermediary format that captures essential affective information from multiple sources, enabling efficient computation while preserving nuanced understanding of user intent beyond literal word meaning.
Data Source
AI summary
Systems and methods for automatically generating at least one of facial expressions, body gestures, vocal expressions, or verbal expressions for a virtual agent based on emotion, mood and/or personality of a user and/or the virtual agent are provided. Systems and method for determining a user's emotion, mood and/or personality are also provided.