CamJam: Using the Camelot Virtual Environment
Wednesday, October 9th, 2019 at the Georgia Institute of Technology
Camelot is a highly-customizable, easy-to-use virtual environment designed for creating interactive narrative games and game prototypes. If offers numerous characters, places, and items that can be customized and controlled via a series of simple text commands.
Commands are sent via standard input and output, so the environment can be controlled by any program which can read and write text to the command line. If you can use Python's print or Java's System.out.println, you can write an experience manager for Camelot!
This workshop will use the first public build release of Camelot, version 0.3. You can download it for Windows and Mac and read the documentation here. If you find bugs (which is likely) please let Alex and Stephen know.
About the Workshop
CamJam will be a half day workshop at AIIDE 2019 that will gather interested users to share their experiences using their own research tools to control Camelot and to learn more about the environment.
We will be inviting specific speakers to share their tools, but if you are interested in presenting at the workshop, please contact us. We are especially interested in experience managers using these technologies:
- Narrative Planning
- Reinforcement Learning
- Social Physics
- Grammar-Based Tools
- Answer Set Programming
- Behavior Trees
- Linear Logic
- Reactive Planning
The ideal demo would be around 30 seconds of content with at least one interaction that can happen in more than one way depending on the user's interaction. While full games are certainly welcome, the purpose of these demos is to discover how future development of Camelot can support a wide variety of interactive narrative tools in preparation for a competition/showcase.
Mr. Shirvani is a PhD student in the Narrative Intelligence Lab at the University of Kentucky. He is the primary architect of the Camelot virtual environment, and he will open the tutorial with an explanation of how Camelot is designed and a demonstration of how to use it's many features. Mr. Shirvani does research on plan-based models of emotion and affect for believable virtual characters in interactive virtual environments such as games, training simulations, and tutoring systems.
Dr. Ware is an Assistant Professor at the University of Kentucky where he directs the Narrative Intelligence Lab and teaches courses on artificial intelligence and game development. His Camelot demo (created jointly with Alireza Shirvani) will showcase intelligent agent behavior based on narrative planning. By reasoning about each agent's intentions and beliefs, and by considering the author's contrtains, a narrative planner can proactively decide what actions each NPC should take. Prof. Ware does research on narrative planning and story generation for virtual environments.
Dr. Harrison is an Assistant Professor at the University of Kentucky where he directs the CORGI Lab and teaches courses on machine learning and game development. In his Camelot demo, the behaviors of all agents have been learned from via many sessions of multi-agent reinforcement learning. Prof. Harrison does research on machine learning, especially reinforcement learning, for training intelligent agents in the real and virtual worlds.
Dr. Samuel is an Assistant Professor at the University of New Orleans where he teaches courses on artificial intelligence and game design. His Camelot demo will showcase agents who react intelligently to one another using the Ensemble Social Physics Engine, which he helped to develop. Prof. Samuel is active in both computer science and improvisational theater performance, and his research explores the intersection of these interests, especially for narrative generation.
Dr. Martens is an Assistant Professor at North Carolina State University where they direct the POEM Lab and teach courses on artificial intelligence and game design. Their Camelot demo will showcase agents whose behaviors are defined using Villanelle, an agent authoring language which combines scripting and generativity, using a logic-based foundation that unifies both. Dr. Martens does research on how formal reasoning tools can interface with human cognition, and their work has significant applications in virtual environments.
Dr. Eger is an Invited Professor at the University of Costa Rica where he directs the SLOTH Lab. His Camelot demo will be an automatically generated murder mystery, using a tool developed with Henry Mohr and backed by Ostari, an implementation of Dynamic Epistemic Logic developed jointly with Prof. Chris Martens. Prof. Eger does research on modeling the minds of intelligent agents to help them interface more naturally with people.
All times are for Atlanta, Georgia, USA, which is in Eastern Daylight Time.
|Alireza Shirvani: How to Use Camelot||1:30 PM||-||2:15 PM|
|Stephen Ware: Story Graph Pruning||2:15 PM||-||2:30 PM|
|Break||2:30 PM||-||2:45 PM|
|Brent Harrison: Multi-Agent Reinforcement Learning||2:45 PM||-||3:00 PM|
|Ben Samuel: Ensemble Social Physics Engine||3:00 PM||-||3:15 PM|
|Chris Martens: Villanelle Behavior Trees||3:15 PM||-||3:30 PM|
|Markus Eger: A Generated Murder Most Foul||3:30 PM||-||3:45 PM|
|Break||3:45 PM||-||4:00 PM|
|Jam Time: Build something with Camelot||4:00 PM||-||4:30 PM|
|Community Discussion||4:30 PM||-||5:00 PM|
About the Interactive Narrative Competition / Showcase
Camelot was developed in the hopes of providing a common context in which many different experience management techniques can be compared. In the next year, we hope to organize a competition or showcase for interactive narrative technology using Camelot, so this workshop will help us to get feedback from the community on what is needed for such an event to be valuable to other researchers. We'd love to see you at the workshop to get your feedback!