EXAG 2023

AIIDE Workshop on Experimental AI in Games
October 8th, 2023

Call for Papers

EXAG 2023 will be accepting papers to three different tracks:

  • Full papers -- Regular papers submitted for oral presentation (up to nine (9) pages). These papers will be incorporated into the proceedings and will be presented as 10-minute talks with 5 minutes of Q&A.

  • Short papers -- Short papers (up to four (4) pages) describing a position, project, or proposal related to any topic of interest to the workshop. These papers will be incorporated into the proceedings and will be presented as 5-minute talks with 3 minutes of Q&A.

  • Lightning talks -- (One-page extended abstract) Presenters are given 5 minutes to present on a relevant topic of interest. Examples of talk formats are calls to action, explanations of design spaces, deep dives, and opinion talks.

Papers will be presented in groups based on their topics. The ordering of topics may depend on speaker availability, especially when accommodating remote presenters. Lightning talks will happen in the afternoon session and are designed to spark conversation prior to attendees dismissing for evening activities.

Submission Instructions

Research Papers

Full and short papers should be submitted on EasyChair. Please select the Experimental AI in Games track.

What to Submit

Papers describe AI research results that establish new entertainment AI challenges, make advances on existing problems, enable new forms of interactive digital entertainment, and/or use AI to improve the game design and development process. Papers are held to the highest standards of academic rigor. In general:

  • Results should be validated in a prototype or test-bed system (e.g., game, robot, generative algorithm), but need not be tested in a commercial environment.
  • The contribution of the paper should be clearly articulated, usually in the introduction.
  • The title and claims made in the paper should match the evaluation carried out and the results obtained. Overly broad titles are discouraged.
  • The paper should demonstrate knowledge of related systems and other approaches to solving similar problems, usually in a Related Work section.

Format

Papers should be formatted in CEUR two-column, camera-ready style and should be anonymized for double-blind review. There are two-column templates available for various platforms. Overleaf users can use the available template. If you prefer to use Microsoft Word, ODT, or an different LaTeX variant, please use these templates from CEUR.

Length

Authors are allotted 9 pages of content for full papers and 4 pages of content for short papers, with no limit on the number of pages for references. Thus, authors are encouraged to submit a paper of length proportional to its contribution. The length of typical submissions is expected to be approximately 6-7 pages of primary content (including figures and tables but excluding references), with 1-2 pages of appendix content. Note, reviewers may, but are not required to, read the appendices, and therefore the paper’s central thesis should be understandable without them.

Submissions longer than the alloted page counts will be considered for desk rejection. Papers whose lengths are incommensurate with their contributions will be rejected.

Evaluation Criteria

Submissions will be peer reviewed. Abstracts and other submitted materials will be judged on technical merit, accessibility to developers and researchers, originality, presentation, impact, and significance. Submissions do not need to score well in all of these categories.

Lightning Talks

Lightning talk proposals should instead be emailed to the organizers at: exag20xx at gmail.com. Please include a one-page PDF with your talk's extended abstract. Please include your name and the title of your talk.

Important Dates

Deadline for paper submissions: (UPDATED!) July 27th, 2023

Notification for accepted papers: August 24th, 2023

Publication-ready submission due: September 11th, 2023

Workshop date: October 8th, 2023

Topics

We are pleased to be a broad workshop with a large range of topics covered, and we aim to continue and expand our scope this year. This is a non-comprehensive list of topics of interest to EXAG:

  • The development of new game systems made possible by AI, from roguelike Unexplored’s procedurally-generated dungeons and puzzles to stealth game Third Eye Crime’s visualization of AI logic.

  • The use of AI in support of mixed-initiative co-creative play experiences, including collaborative storytelling games like Bad News.

  • Better living through AI—improving game development and design through new and interesting applications of AI, from intelligent design tools to automated QA.

  • Discussion of experimental AI technologies used in games and play environments that may extend beyond the digital space into physical, embodied environments. Examples include (but are not limited to) AI in games that involve tangible user interfaces, interactive play spaces like museums, or embodied gameplay in virtual or augmented reality environments.

  • Cross-pollination from AI subfields not typically used in games, like computational linguistics, machine vision, and procedural music.

  • Traditional AI techniques applied in new ways that break genre conventions, like Left 4 Dead’s drama management or Black And White’s learning creatures.

  • Human-centered evaluation of intelligent systems for digital entertainment.

  • Discussion of interesting but relatively unknown historical examples of experimental AI in games and related areas. Here, we are especially interested in examples from the game industry, such as Captain Blood’s (1988) modular icon-based interface for procedural communication with NPCs, Intellivision World Series Baseball’s (1983) telecast-influenced procedural camera system, or Skool Daze’s (1984) real-time simulation of NPC agendas. In some cases, the investigation may reveal that these historical examples actually represent the state of the art in game AI today.

  • Reports on failed experiments related to any topic in our purview, with insight into what went wrong and how others can learn from the failure.

  • Industry case studies in which non-academic developers present their applications of experimental game AI in commercial game contexts.

Previous work presented at EXAG includes dynamic NPC knowledge modeling, blending or inventing game designs, intelligent design tools, hybrid generative algorithms, formal logics for games, generation of procedural game entities, procedural game narratives, and more.

Not sure if your topic is a fit? Drop us a line!