A network data structure representing narrative characters as vertices and their interactions or relationships as edges, used to model social dynamics, drive procedural story generation, and analyze narrative structure through graph theory and social network analysis.
Semantic Classification
Content
Technical Details
- Network Properties:
- Vertices represent narrative characters
- Edges represent interactions or relationships
- Relationship types: positive, negative, neutral
- Dynamic changes drive narrative emergence
- Research Findings (2024):
- LLM-generated stories show higher density and clustering among amiable characters
- Strong bias toward positive relationships in AI stories
- LLMs struggle with complex relationships in longer narratives
- Human stories have more conflict-driven social dynamics
- Notable Systems:
- NetworkING: Uses character relationships for interactive narrative generation
- STORYVERSE: Co-authoring dynamic plot with LLM-based character simulation
- Applications Across Fields: Literary analysis, AI narrative generation, game design, complex systems modeling
Applications
- Procedural story generation
- Interactive narrative game design
- Literary criticism and analysis
- Social simulation in virtual worlds
- NPC relationship systems