• Instructor: David Thue
• Contact: Office in Venus floor 2, telephone 599-6412, e-mail davidthue[ ]ru.is
• Class Times: Tuesdays 8:30-10:05 (M114) & Thursdays 10:20-11:55 (M114)
• Online Forum: Piazza Course Page
• Project Tracking Software: Trello

The ability to create and revise stories is fundamental to human interaction, but computers are still in their infancy of doing either very well. In this course, we will explore computational storytelling from the perspective of Artificial Intelligence. We will read and analyze key papers from the literature, discuss how such technologies might be applied in different domains (e.g., computer games or training simulations), and obtain hands-on experience by building prototypes that extend previous research.

There is no course textbook, but you have access to a wide range of relevant research papers via ICIDS.org's Springer LNCS Conference Proceedings and AAAI's Library of Conference Proceedings (AIIDE), Symposia Papers, and Workshop Papers. All of these sites should allow you to download their papers free of charge.

The course reading list is available here: Course Presentation Schedule

Upon completion of the course, students should be able to:

• Discuss current challenges in computational storytelling
• Describe a variety of algorithms and techniques for addressing those challenges
• Present and critique related research, both orally and in writing
• Pursue original research that extends the concepts discussed in class
• Write a conference-level research report
• Identify computational storytelling projects that could be pursued as thesis research

Please try to use the course discussion system (Piazza) for posting questions regarding lectures or your projects, rather than sending e-mail. That way we can build a shared repository of useful questions/answers.

Piazza Course Page

Please note that attendance during both classes each week (Tuesday and Thursday mornings) is required. Please inform the instructor if this is hard for you for some reason, such as scheduling conflicts or sick leave.

This course will combine presentations, discussions, and brainstorming in class with a hands-on term project.

Most classes will be dedicated to presenting and discussing topics in computational storytelling, with each topic being grounded by an assigned reading and/or gameplay session. Before coming to each class, everyone will read the assigned paper or play the assigned game. During class, one student will present the assigned paper/game using slides and/or live demonstrations (Paper Presentation), and everyone else will hand in a Written Review of the paper/game before the presentation starts (max 2 pages, single spaced). Guidelines for presentations and reviews will be provided.

Each written review will receive one of three grades: 0 (missing or severely lacking), 0.5 (incomplete), or 1 (complete). Each review will be worth up to 1% of the final grade, for a total of 17% of the final grade. Of the 19 written reviews that each student will complete, the scores of two reviews that scored the lowest will be ignored. For example, for a student who scored 0 for one review, 0.5 for 3 reviews, 1 for 15 reviews, their total score out of 17% would be 15*1% + 2*0.5% = 16% (the 0 and one of the 0.5s would be ignored).

Each student will present 3* of the assigned papers and/or games, using slides (and possibly demonstrations) to: 1) describe necessary background information, 2) explain and critique the technique(s) and evaluation(s) (if any) discussed in the paper, and 3) highlight any follow-up work that has been done since the assigned work was published. Each presentation is expected to last roughly 45 minutes, followed by an in-class discussion of the assigned game/paper after a 10 minute break.

Each presentation will be graded based on its content, its organization, and the clarity with which it is delivered. If the score of any student's third presentation is higher than one of their two earlier presentations, the lowest earlier score will be dropped and the third score will count double. For example, if Student A obtained presentation scores 8, 7, and 10 on their first, second, and third presentations, then the score from the second presentation (7) would be dropped and the 10 would be counted double (8+2*10 = 28). Meanwhile, if Student B obtained scores of 8, 10, and 7, then no scores would be dropped because the third score was the lowest (8+10+7 = 25).

NB: Each student must complete 3* Paper Presentations to qualify for submitting the Final Report.

* This number may decrease if more students enrol in the course.

The presentation schedule is available here: Course Presentation Schedule

The participation grade will be based on the instructor's subjective evaluation of the student's participation throughout the semester. This evaluation will focus mostly on their activity during class discussions, but may also consider their use of the online forums and project tracking tools.

Beginning in Week 6, students will work on a research project in teams of 2 to 3. The topic of the project must be related to computational storytelling, and the project itself must focus on using AI techniques to address a particular challenge. We will form teams on September 16 (Week 5) following an in-class Idea Jam, during which everyone will propose and discuss several ideas for potential term projects. Each team will propose a project via written hand-in, and, upon its approval by the instructor, work to complete their project for the remainder of the term. Each team's project will be evaluated in four parts: a Project Proposal (due on September 19), Project Updates from Week 6 onward, a Final Presentation (date TBA), and a Final Report in the format of a standard conference paper (due date TBA, but likely during final exams; this course has no final exam).

Project Updates are short presentations (5 minutes of slides/demos + 5 minutes for discussion) given by one team at the end of each class from Week 6 onward. Whenever a student presents a paper at the beginning of a class, their team will give a project update at the end of that class.

* This number may decrease if more students enrol in the course.