CS 7634 / 4803 SVW / LCC 8803 MR : A.I. Storytelling in Virtual Worlds

Home Schedule Readings Assignments Flyer

Team Project

The course project will be a semester-long effor to research and develop a Alternate Reality Game (ARG). An ARG is a type of interactive story that plays out in the real world. Mobile ARGs use mobile location-aware devices such as smart-phones and GPS receivers to augment the players' experience such that the story unfolds as the player moves about the physical world. Moblie ARGs have been proposed as a new and emerging form of entertainment, and also theorized to have positive health attributes if they require walking, biking, or running. However, ARGs (mobile and otherwise) currently require a substantial form of human support. Unlike computer games, that can be run on consoles without human intervention, ARGs require human game masters to monitor the game situation, direct the action, and make changes to the story structure on the fly. ARGs also often utilize human confederate actors that play the roles of fictional characters to enhance the sense of realism; the use of confederate actors is an expensive and time-consuming means of entertaining people. These limitations have prevented ARGs from becoming a mainstream form of entertainment (or education or training or health activity). However, recent developments in the field of artificial intelligenct -- especially the branch of artificial intelligence pertaining to games and interactive entertainment -- present a way forward. AI Game Masters and virtual autonomous character agents have been the focus of recent AI research.

(An Alternate Reality Game is not Augmented Reality, which is the super-imposition of graphics on the real world through head-sets or other mobile devices. Alternate Reality games attempt to make people feel as if they are in an alternate reality through storytelling.)

More information about ARGs on Wikipedia. Examples of ARGs:

Teams of 2-4 people will work together to design an develop and playable ARG system that uses AI as an integral part of the execution and user experience of the game. By "ARG system" I mean a piece of software that delivers and manages the execution of the game in the real world. That is, while there may be work to define and scope a particular story (or, more correctly, a space of interactive stories), the main consideration is not the storytelling abilities of the team. Instead, the emphasis is on the technology through which story experiences are delivered to human players. I do not expect the teams to be great storytellers or for the games themselves to be "fun." I do expect the teams to be exceptional technically. By "AI as an integral part," I mean that no software solution will be accepted without a non-trivial AI implementation. Further, this AI system should be integral to the ARG system in two ways: (1) the game design is simply unplayable if the AI system isn't there, and (2) the AI cannot be replaced by the trivial application of human effort. As noted in the introductory paragraph, ARGs have significant technical and human-effort limitations. The ARG system should be designed and developed in such a way that AI is used to overcome some technical and/or human-effort limitation. This means you have the opportunity to design a system for delivering entertainment (or educational or training) experiences never tried before.


Preferred (most ARGs are multiplayer cooperative games), but not strictly necessary.

Mobile platforms

Your ARG system will need to manage players who are at times away from traditional computing platforms (e.g. desktops). There are many ways to deal with this:

Your ARG system must be playable, however, your grade will primarily reflect your innovative use of AI. Therefore, I encourage you to find light-weight and simple solutions to delivering ARG story content to the player while he or she is mobile. Javascript in the browser and SMS may be the most light-weight engineering solutions. Solving this problem of dealing with mobile platforms is a necessary but not sufficient condition of a succesful project.

If you have smart devices, you are encouraged to use them. If you do not, I have 2 iPhone 3GS (with data and voice plans) and 1 iPhone 3G (no data or voice plan, wireless only) in my lab. There will be means by which they can be checked out. Note that iPhone development requires a developer license ($100) and license provisioning is a pain in the butt.

Artificial Intelligence

Any non-trivial implementation of AI that is integral to game play and non-trivial will be accepted. Almost all AI techniques (search, case-based reasoning, machine learning, decision processes, reactive planners, optimization, Bayesian inference, etc.) are applicable to game AI and thus applicable to this project. It is incument on your team to identify the problem that must be solved with AI and also the best technique to implement. You are encouraged to talk through your ideas with the instructor and to read ahead with the class readings to get a good idea of what is possible, and what has been done before. It is acceptable to reimplement techniques that you find in the course reading materials.

Here are a few examples of the usage of AI in Alternate Reality Games and mobile platforms:

(See T-Square for papers not available on the Internet)

The example movie from August 23rd class. This is a mock-up of what I felt would be a good ARG using mobile devices.

Grading and Evaluation

The ARG system must be playable, meaning the instructor can play through a game. The game must be played on campus and must not take more than 1 hour. That being said, your project will be primarily graded on the way you use AI to solve a problem and drive the game play experience. The problem and solution are of your choosing, but it must be a compelling problem and a reasonable solution. If you have any questions about whether your problem and solution are appropriate, you should talk with the instructor. The game play experience itself does not need to be fun or entertainming. I don't expect you to be a great storyteller or a great game designer. I expect you to be great at proposing and solving technical challenges.


Pitching your project [NEW 09.23.11]

When presenting your project to the class, treat it as a sales pitch. You need to convince the others in the class that you have a game design and an AI solution that will enable you to deliver an unique ARG gaming experience that scales up, meaning that it can be played with lower overhead than ARGs historically have been. The presentation should last 5-7 minutes. This is a short amount of time, so be efficient, pithy, and concise!

Here are notes taken from computer game industry veterans on how to "sell" a game concept to upper management. Some of this will be relevant and some will not be. You do not need to do all of this (or any of this) in your presentation. This is merely a set of suggestions as a starting point. Italics are notes specific to the class project only


Project Critiques

Teams will be asked to report on the designs and progress of other teams as if one were preparing a report for an investor interested in commercializing ARGs. The purpose of this task is not to critique presenters, but to help teams learn from each others' experiences.

[NEW 09.20.11] Critique 1 evaluations will be approximately 5 pages in length. Each report will critique on the 4 other project teams. You should look at the other projects from the perspective of someone deciding whether to invest money into a new game technology. In other words, will each project solve a real problem when it comes to creating and deploying ARGs, and will the solution work? For each other team project, you should spend 1+ pages critiquing each of the other projects. A project critique will include (a) a brief summary of the proposed problem and approach to solving the problem, (b) a list of concerns about whether the proposed solution will actually address the proposed problem and feasibility, and (c) recommendations for improvements to the project. Recommendations can be to modify the approach, to simplify aspects of the solution, to adjust the problem so as to be more compelling. Finally, the report will close with an overall assessment of all projects and summary of the best approach, or best combination of approaches in the case that you see things in various projects that, combined, will make for a compelling system.

[NEW 11.21.11] Critique 2 evaluations will happen a bit differently. Final presentations will briefly overview the flavor of the experiece that players have (or are expected to have). A bulk of the final presentations will describe the underlying AI technologies used in sufficient detail for all to understand how they work. Each team will write a critique of other projects other than their own. Each critique of each project will be approximately 3+ pages in length and have the following elements:

  1. Describe the AI approach(es) from the literature we reviewed that most closely matches the project. Describe how the AI in these approaches might be used to implement the design proposed for the project. You may have to describe how the AI approach from the literature must be adapted (significantly) or combined with other AI approaches from other literature.
  2. Summarize the actual AI approach used in the project. Provide an argument for why the project team made the right choice of AI technology (as opposed to the alternatives you posed in part 1). In other words, defend the other project team and justify their design decisions.

In total, critique reports will be 12 pages in length. Use 1-inch margins and a 12-point font (or less).

In-Class Exercises

Throughout the semester, there will be several in-class participatory exercises that augment course material. The exercises will be announced in advance and may require out-of-class preparation in advance.

Presentations and Discussions

Students will be asked to assist with the presentation of course material from the conference and journal papers on the reading list. Students may also be asked to help facilitate a discussion about the course material that is presented. Presentations will be short ~ 5-15 minutes. The purpose is of the presentation and discussion is not to cover all the details of the paper, but to lay out the relevant points at a high level of detail, going into low level detail only when necessary to make a point.

Consider the following questions for your presentation and discussions:

It is good have an opinion and to share it.

I recommend preparing questions for the class -- that will look good for your grade.