Overview
Welcome to the Second Workshop on Game AI Algorithms and Multi-Agent Learning! This is a workshop co-located with IJCAI 2025 Conference in Montreal, Canada.
Games serve as sources of inspiration and open test-fields for human intelligence. Various games, ranging from traditional card games to video games and virtual reality games, have shared a history of evolution for more than 5000 years with civilization and technology of the human society. Games contain incarnations of diverse models in the game theory, and also provide rich simulations of real-world scenarios. In the present age, games continue to catalyze the development of artificial intelligence, proposing novel research problems and providing benchmarks for learning algorithms. To date, many games including Chess and Go have witnessed superhuman performance of AI algorithms, while human-level AI for harder games like Mahjong remains an open problem.
The Second Workshop on Game AI Algorithms and Multi-Agent Learning aims to investigate specific solutions to different forms of games as well as general learning theories, frameworks, and methods for game playing and decision making.
Topics
We solicit contributions from AI researchers who are interested in games and decision-making algorithms. Our topics of interest include but are not limited to:
- AI algorithms for game-playing
- Game theories and multi-agent systems
- General formulations of game solutions
- Game mechanics and environment design
- Feature engineering in game AI
- Evaluation of agents' rankings and game balance
Schedule
| Time | Agenda |
|---|---|
| 09:00 | |
| 09:15 | |
| 09:45 | |
| 10:30 | |
| 11:00 | |
| 11:00 | |
| 11:20 | |
| 11:50 | |
| 12:10 | |
| 12:30 | |
| 14:00 | |
| 16:00 |
Submission Requirements
Submissions can contain recent work on all relevant topics. All submissions will go through a light review process to ensure they fit with the scope of the workshop and to verify that they meet an acceptable quality threshold.
- Format guidelines:The LaTeX styles and Word template is available at https://www.mdpi.com/authors/latex
- Page limitation: Up to 8 pages + unlimited references/appendices
- Please submit your paper to https://cmt3.research.microsoft.com/GAAMAL2025.
Special Call For Papers
As a continuation and expansion of the IJCAI Mahjong AI Competition in previous years, the GAAMAL workshop also welcome papers from teams of previous Mahjong AI competitions to present their algorithms and models for the competitions.
Important Dates
All deadlines are "anywhere on earth" (UTC-12)- Submission Deadline: May 9, 2025
- Paper Notification: June 4, 2025
- Workshop time: August 18, 2025
Accepted Papers
- Guangchong Zhou, Feng Hong, Zeren Zhang and Guoliang Fan, Diversified Experience Replay for Multi-Agent Reinforcement Learning [Paper]
- Andrew Hamara, Greg Hamerly, Pablo Rivas, and Andrew C. Freeman, Learning to Plan via Supervised Contrastive Learning and Strategic Interpolation: A Chess Case Study [Paper]
- Lingfeng Li , Yunlong Lu, Yongyi Wang, Qifan Zheng, and Wenxin Li, Mxplainer: Explain and Learn Insights by Imitating Mahjong Agents [Paper]
- Patrick Benjamin, Alessandro Abate, Networked Communication in Mean-Field Games with Function Approximation and Empirical Mean-Field Estimation [Paper]
Program Chair
Organization Committee
Contact
Please send your inquiries to gaamal.workshop@gmail.com, or Xingbo Wang jacksimbol@stu.pku.edu.cn
-
Attention! According to IJCAI Regulations:
At least one author of each accepted paper must travel to the IJCAI venue in person.
Multiple submissions of the same paper to more IJCAI workshops are forbidden.
2025-03-17