Intelligent Storytelling

Yun-Gyung Cheong, Associate Professor, Department of AI, College of Computing and  Informatics, Sungkyunkwan University,South Korea

Course Type: Design and Technology-focused 

Keywords: AI, NLP, Computational Models, Narrative Analysis

University Department Level Credits Length Medium
Sungkyunkwan University Department of AI Graduate 3 15 In-person

Course Description

The goal of this course is to theoretically understand interactive stories and computationally model  them to implement technologies for automatic analysis and creation of narratives. The curriculum  includes story analysis based on narrative theory, various computational models applying artificial  intelligence algorithms, authoring tools for creating interactive stories, and case studies  implementing intelligent stories in VR and games. Students who take this course are expected to  be able to design and implement systems that enhance user experience by adding story elements  to the problems they want to solve.

Weekly Outline

Week 1. Course Introduction

Week 2. Narrative Analysis Theory 

Week 3. Characters 

Week 4. Conflict 

Week 5. Narrative Time 

Week 6. Genre 

Week 7. Point of View 

Week 8. AI Knowledge Representation and Inference 

Week 9. AI Story Planning Algorithms 

Week 10. Discourse and Presentation of Narrative 

Week 11. Story Comprehension, Cognitive Science Theories 

Week 12. Computational Models of Narrative, Story Datasets 

Week 13. Paper presentation – Interactive Storytelling Applications 

Week 14. Paper presentation – HCI 

Week 15. Project presentation 

Course Objectives ​

  • To understand the concepts and terms of narrative analysis theory.  
  • To analyze stories and interactive stories using interdisciplinary narrative analysis  theories.  
  • To build computational models to generate or analyze stories using technologies such as  data analysis, natural language processing, and machine learning.  
  • To apply AI and NLP techniques to practical problems in interactive storytelling  applications such as games and education.

Reading

  • The Storytelling Animal: How Stories Make Us Human, Jonathan Gottschall (Book,  2012) 
  • Story: Style, Structure, Substance, & the Principles of Screenwriting, Robert McKee  (Book, 2010) 
  • The Hero with a Thousand Faces, Joseph Campbell, (Book, 1949) 
  • The Writer’s Journey, Christopher Vogler (Book, 1992) 
  • Save The Cat, Blake Snyder (Book, 2005) 
  • 20 Master Plots, Tobias, (Book, 1993)
  • 7 Basic Plots, Christopher Booker (Book, 2004) 
  • Morphology of the Folktale, Vladimir Propp (Book, 1928) 
  • Introduction to the theory of narrative, 4th ed., Mieke Bal (Book, 2017)
  • Narrative Discourse: An Essay in Method, G. Genette (Book, 1983)
  • Story and Discourse: Narrative Structure in Fiction and Film, S. Chatman (Book,  1980) 
  • The Foundations of Screenwriting, Syd Field (Book, 2005) 
  • Save the Cat: The Last Book on Screenwriting You’ll Ever Need, Blake Snyder
  • (Book, 2005) 
  • The Plot Thickens: 8 Ways to Bring Fiction to Life, Noah Lukeman (Book, 2012)
  • A Man Without a Country, Kurt Vonnegut (Book, 2005) 
  • Now Write! Science Fiction, Fantasy, and Horror, Laurie Lamson (Book, 2014)
  • Save the Cat: Write a Novel, Jessica Brody (Book, 2018) 
  • Dictionary of Narratology, Gerald Prince (Book, 2003) 
  • Aspects of the Novel, E.M. Forster (Book, 1927)  
  • The Story Grid, Shawn Coyne (Book, 2015) 
  • Characters, Emotion & Viewpoint: Techniques and Exercises for Crafting Dynamic  Characters and Effective Viewpoints, Nancy Kress (Book, 2005) 
  • The Cambridge Introduction to Narrative. H. Porter Abbott (Book, 2008)
  • Story understanding. In Encyclopedia of Cognitive Science. London: Macmillan.  Mueller, Erik T. (Book Chapter, 2002).  
  • Scripts, Plans and Knowledge, Schank, Roger C. and Robert P. Abelson., International Joint Conference on Artificial Intelligence (Paper, 1975) 
  • TALE-SPIN, James Meehan. An Interactive Program that Writes Stories. Fifth  International Joint Conference on Artificial Intelligence (Paper, 1977). 
  • Facade: An Experiment in Building a Fully-Realized Interactive Drama, Mateas and  Stern, In Game Developer’s Conference: Game Design Track, San Jose, California,  (Paper, 2003).

Viewings

Viewings

  • UP (Animation, 2009) 
  • Matrix (film, 1999) 
  • Parasite (film, 2019) 
  • Aningaaq (short film, 2013) 
  • Nobody (film, 2021)  
  • Sherlock (TV series, 2010-2017) 
  • As good as it gets (film, 1997) 
  • Star Wars (film, 1977) 
  • The Godfather (film, 1972) 
  • The Shawshank Redemption (film, 1994) 
  • Extraordinary Attorney Woo (TV series, 2022) 
  • Joker (film, 2019) 
  • Squid Game (Netflix series, 2021) 
  • Groundhog Day (film, 1993)
  • Rashomon (film, 1950) 
  • The Usual Suspects (film, 1995) 
  • American Beauty (film, 1999) 
  • The Remarried Empress Fantas (Web novel, Webtoon, 2018~2020) 
  • Breaking Bad (TV series, 2008-2013) 
  • Memories of Murder (film, 2003) 
  • Money Heist (TV series, 2017-2021) 
  • Knives Out (film, 2019) 
  • The Sixth Sense (film, 1999) 
  • Mission: Impossible (film, 1996) 
  • You Have to Burn the Rope (fame, 2008) 
  • Conversations with Friends (Sally Rooney) 
  • The Handmaiden (film, 2016) 
  • 500 Days of Summer (film, 2009) 
  • Memento (film, 2000) 
  • The Others (film, 2016) 
  • Fight Club (film, 2001) 
  • Where is the Friends’ House? (film, 1987) 
  • Please Look After Mom (Novel, 2008) 
  • Love virtually (Novel, 2011) 
  • Die Hard (film, 2988) 
  • Sunspring (short film, 2016) 
  • The last of Us (Digital Game, 2010) 
  • Heavy Rain (Digital Game, 2010) 
  • Indigo Prophecy (Digital Game, 2005) 
  • Detroit: Become Human (Digital Game, 2013) 
  • Bioshock: Infinite (Digital Game, 2013) 
  • Back to the Future (film, 1985) 
  • Black Mirror: Bandersnatch (Interactive film, 2018) 
  • Play novel (Digital Game, 2022) 
  • Choose Love (Interactive film, 2023)

IDN Artifacts

  • AI and NLP techniques 
  • Computational Models of Narrative

IDE and IDN Authoring Tools

  • IDE: Python, NLP toolkit, Google Colaboratory 
  • chatGPT, LLMs

Major Assignments (being assignments whose value is of 25% or more)

Paper review and presentation (30%) 

  • Purpose: Cultivating the ability to independently read, study, and analyze the  latest research papers on computational models, AI, and NLP that correspond to  narrative theories learned in the class. 
  • Requirements: The presenter will give a summary of the paper in a 10-15 minute  presentation, followed by the strengths and drawbacks of the research discussed by another student who plays the role of a reviewer in 5 minutes.  
  • Evaluation:  
    • Presentation  
      • Structure and Content: The presentation clearly explains the  content, making it easily understandable for the audience. Verify  that no critical details are missing. The presentation must cover the paper’s primary objectives, the problem it addresses, related research, methodologies employed, experiments and results, contributions to the field, and any limitations.
      • Presentation Style: The structure and aesthetics of the presentation materials should be clean and visually appealing to maintain audience engagement.
      • Presentation Delivery: Deliver the presentation at an appropriate pace to keep the audience engaged without rushing through or dragging the content. Avoid merely reading from the slides; instead, speak calmly, energetically, and engagingly to make the presentation lively and interesting. Strive for clarity and precision in explanations, ensuring they are insightful and appropriately paced to suit all audience members.
      • Comprehension and Reference: Thoroughly understand the paper  yourself before attempting to present it. This deep comprehension  will enhance the quality of your presentation. If certain aspects of  the paper are unclear from the content alone, consult additional  references or background theories to better grasp and convey the  material effectively. 
    • Reviewers evaluate the paper by discussing its strengths, including its  novelty and contributions to the research or application field. They also  identify any weaknesses or limitations. Additionally, they are encouraged  to share the questions raised during their review and highlight areas  needing improvement. Finally, they offer specific suggestions and outline  potential directions to further enhance the quality of the paper.
    • Paper List

Title            

venue

Link

The emotional arcs of stories are  dominated by six basic shapes 

EPJ Data Science

https://epjdatascience.springer open.com/articles/10.1140/epj ds/s13688-016-0093-1

Detecting Narrative Elements in  Informational Text 

NAACL 2022

https://aclanthology.org/2022. findings-naacl.133/

The Construction of Situation  Models in Narrative Comprehension:  An Event-Indexing Model

Psychological  

Science 1995

https://www.jstor.org/stable/4 0063035

Compute to Tell the Tale: Goal Driven Narrative Generation 

Multimedia 2022

https://dl.acm.org/doi/abs/10.1 145/3503161.3549202

“Let Your Characters Tell Their  Story”: A Dataset for Character Centric Narrative Understanding 

EMNLP 2021

https://aclanthology.org/2021. findings-emnlp.150.pdf

PeaCoK: Persona Commonsense  Knowledge for Consistent and  Engaging Narratives 

ACL 2023

https://aclanthology.org/2023. acl-long.362/

Conflicts, Villains, Resolutions: Towards models of Narrative Media  Framing 

ACL 2023

https://aclanthology.org/2023. acl-long.486.pdf

Generative Agents: Interactive  Simulacra of Human Behavior

 

https://arxiv.org/abs/2304.034 42

ATOMIC: An Atlas of Machine  Commonsense for If-then Reasoning AAAI 2019

 

https://ojs.aaai.org//index.php/ AAAI/article/view/4160

COMET: Commonsense  

Transformers for Automatic  

Knowledge Graph Construction 

ACL 2019

https://aclanthology.org/P19- 1470/

GLUCOSE: GeneraLized and  COntextualized Story Explanations 

EMNLP 2020

https://aclanthology.org/2020. emnlp-main.370/

Minding Language Models’ (Lack of) Theory of Mind:A Plug-and-Play Multi-Character

Belief Tracker ACL 2023 

 

https://aclanthology.org/2023. acl-long.359/

   
 

Automated storytelling via causal,  commonsense plot ordering 

AAAI 

https://ojs.aaai.org/index.php/ AAAI/article/view/16733/165 40

COINS: Dynamically Generating  COntextualized Inference Rules for  Narrative Story Completion 

ACL 2021

https://aclanthology.org/2021. acl-long.395/

A Corpus and Evaluation Framework  for Deeper Understanding of  Commonsense Stories

NAACL HLT  2016

https://cs.rochester.edu/nlp/ro cstories/

STORIUM: A Dataset and  

Evaluation Platform for Machine-in the-Loop Story Generation 

EMNLP 2020

https://aclanthology.org/2020. emnlp-main.525/

Plot-guided Adversarial Example  Construction for Evaluating Open domain Story Generation 

NAACL 2021

https://aclanthology.org/2021. naacl-main.343/

NARRASUM: A Large-Scale  Dataset for Abstractive 

Narrative Summarization 

EMNLP 2022

https://aclanthology.org/2022. findings-emnlp.14.pdf

Re3: Generating Longer Stories With  Recursive Reprompting and Revision EMNLP 2022

 

https://aclanthology.org/2022. emnlp-main.296/

DOC: Improving Long Story  Coherence With Detailed Outline  Control 

ACL 2023

https://aclanthology.org/2023. acl-long.190/

Synthesizing Coherent Story with  Auto-Regressive Latent Diffusion  Models 

SOTA model

https://arxiv.org/abs/2211.109 50

Pun Generation with Surprise 

ACL 2019

https://aclanthology.org/N19- 1172/

Genre-Controllable Story Generation  via Supervised 

Contrastive Learning 

WEB 2021

https://dl.acm.org/doi/10.1145 /3485447.3512004

Go Back in Time: Generating  Flashbacks in Stories with Event  Temporal Prompts 

NAACL 2022

https://aclanthology.org/2022. naacl-main.104.pdf

Are Fairy Tales Fair? Analyzing  Gender Bias in Temporal Narrative  Event Chains of Children’s Fairy  Tales 

ACL 2023

https://aclanthology.org/2023. acl-long.359/

TaleBrush: Visual Sketching of Story  Generation with Pretrained Language  Models 

CHI 2022

https://dl.acm.org/doi/abs/10.1 145/3491102.3501819

Final Project (60%) 

  • Purpose: Applying the narrative theories and the computational models of  narrative learned in the course to create and understand stories automatically, or to develop interactive story applications 
  • Requirements: The project shall contain Idea, the direct background of the idea,  related work, method (model, framework, algorithm), example of operation  (scenario). A team can consist of 1-2 members, but if a team is composed of 3  members, they must conduct data analysis, experiments, or system implementation. 
  • Evaluation:  
    • Contribution of the proposed Methods or Systems 
    • Novelty of the method 
    • Related research survey section should contain related narrative analysis  theory, and previous research works similar to the proposed idea  
    • Technical soundness of methods (logically plausible) & evaluation 
    • Feasibility of the method 
    • Completeness of the idea 
    • Usability of the idea 
    • Team member role distribution  
    • Presentation (5-7 minutes presentation + 3 minutes Q&A)

Course Best Practices

  • Narrative theory and computational theory are arranged in parallel, examining how the  content learned theoretically is used to automatically create stories, and structured to  expose students to the latest AI and NLP technologies.  
  • Two sessions per week: 75-minute video lecture session, 75-minute hands-on practice &  paper presentation session (applying narrative theory to actual stories for analysis,  presenting research on related computational models). 
  • <<use of additional lecture resources for offline studying>> YouTube videos that  effectively demonstrate the application of narrative analysis theories learned in class to  actual movie scenes.
    • Andrew Stanton: The clues to a great story (https://youtu.be/KxDwieKpawg) ○ The Secret to Great Characters — Characterization Explained (https://youtu.be/43Vrnaz8fYU)
    • The Soul of Good Character Design (https://youtu.be/SM3IQFgP-d8)
    • The Purpose of Conflict (https://youtu.be/6_sri6K_IoM)
    • How to Create Story Conflict (https://youtu.be/z-CIZvS9NEg)
    • Types of Foreshadowing in Films — What is Indirect vs. Direct Foreshadowing? (https://youtu.be/JOas2BLjPR4)
    • What is a Red Herring — 5 Techniques to Mislead & Distract an Audience (https://youtu.be/47ntBElzaWk)
    • Movie Genres Explained — Types of Films & the Art of Subverting Film Genres (https://youtu.be/rDVVE8ZHJ3o)
    • The POV Shot — The Art of the Subjective Camera and “Point of View Shot”  (https://youtu.be/BLCQAmTleP0)
    • What is the Fourth Wall? The Best Examples of Breaking the Fourth Wall (https://youtu.be/PZL13w9TqbA)
    • Your Script Is Missing This: Setups and Payoffs (https://youtu.be/qGDdpXLc1CQ)
    • What is Theme — 5 Ways to Layer Theme into a Screenplay (https://youtu.be/9ELleu9J05g
  • <<relationship between course material and assignments>> Student participation and two  small assignments constitute 10% of the overall grade. After taking lessons on narrative  and character analysis theories, students are expected to apply these concepts to real-world  stories. For each assignment, they analyze their favorite stories and produce concise reports  demonstrating their understanding and application of the theories. 
  • <<dissemination options for students’ work>> The class, comprising 25 students, is  organized into three groups for the final project presentation. Each group presented their  work followed by peer reviews and Q&A sessions. This structure facilitates collaborative  learning and enables students to receive constructive feedback from their classmates on  their projects.