![Graph embeddings
↔
[-0.13391995429992676,
-0.03493423014879227,
0.005581242498010397,
-0.05787748470902443,]](https://image.slidesharecdn.com/treeofbabel-190706075531/85/The-Tree-of-Babel-10-320.jpg)
![Raw data to graph embeddings
Graph representation
{“nodes": [[1], [2], [18], … ],
"edge_index": [[1, 2], [2, 18], [2, 12], … ]}
Data formatting
{-0.13786359131336212,
-0.028311695903539658,
0.017919035628437996,
… ,
-0.05109648406505585}
Graph embeddings](https://image.slidesharecdn.com/treeofbabel-190706075531/85/The-Tree-of-Babel-12-320.jpg)
![Classification
Clustering Algorithms Self Classification
[Basic,
Broad,
Long,
Merged,
Clustered,
Cycle]
using graph properties](https://image.slidesharecdn.com/treeofbabel-190706075531/85/The-Tree-of-Babel-14-320.jpg)
The Tree of Babel
- 2. The Tree of Babel Machine Learning with narrative graphs Youjin Chung
- 5. Interactive Narrative Structure = design pattern = flow chart = graph
- 6. Why ML? …Checking 14M possibilities
- 7. 1. ML implementation to interactive narrative structure as graphs 2. Provide a web app for graph generation for writers/game makers Goal
- 8. 1. Graph Embeddings 2. Classification 3. Generation Flow
- 11. Transverse Reading Gallery ~740 graphs from gamebooks
- 12. Raw data to graph embeddings Graph representation {“nodes": [[1], [2], [18], … ], "edge_index": [[1, 2], [2, 18], [2, 12], … ]} Data formatting {-0.13786359131336212, -0.028311695903539658, 0.017919035628437996, … , -0.05109648406505585} Graph embeddings
- 14. Classification Clustering Algorithms Self Classification [Basic, Broad, Long, Merged, Clustered, Cycle] using graph properties
- 15. Basic basic directed graphs tree shapes clean + straightforward
- 16. Broad Larger widths Short playthrough Strong replayability
- 17. Linear The linearity of narrative structure Easy control of plot engaging storytelling through long playtime
- 18. Merged The graphs merge into certain points. important events in a story
- 19. modular structure detailed choices in a single situation organized by geography. Fragmentary, episodic. Clustered
- 20. Cyclic loop back unlock new status while looping time loops / geographic travel
- 21. 3. Generation
- 22. Autoencoder
- 23. Visualization
- 25. What’s next? Labeling nodes with sentences: similarity(node vector, sentence vector) Train Graph CNN with my own categories.
- 26. Takeaways ML implementation for narrative graphs Understanding of Graph Theory, Vector Embeddings, ML models and implementation Refining and Formatting Custom Data
- 27. Special Thanks: Allison Parrish, Gene Kogan, Dan Shiffman, Jeremy Douglass and Dongphil Yoo