Meme War is a 3D simulation platform developed as a proof of concept on the incorporation of machine learning in games, particularly covering the aspect of meme gene co-evolution. While most research works have dedicated on exploring the use of machine learning in games, the success level has been rather limited due to several constraining factors. Meme War aspires to take a new approach of incorporating machine learning and computational intelligence that follows the Meme and Gene concept, and how the co-evolution of intelligent agents in game could lead to more human-like avatars in games. These shall be achieved through the integration of several inter-disciplinary artificial intelligence (AI) and game technologies to facilitate memetic and genetic co-evolution. Meme War shall serve as a starting reference and platform for research in memetic computing field as well as for the development of human-like believable games.
Imagine having non-player character as enemies in games that is able to change its attack patterns and style according to the players' strategy, it is envisioned that intelligent agents with learning capability will play a big role in the future of AI games. In the path towards the next generation AI game, future R&D efforts will focus on realizing this vision through the development of a meme gene co-evolutionary AI framework for games. The outcome of this project is in the form of a game AI middleware which can be easily integrated into the various game engines. This exciting research will benefit the whole gaming industries, ranging from the game developers to the gamers playing the game.
In Meme War, 3D agent simulation game that embeds various machine learning and computational intelligence algorithms, user is able to play to train the agent (e.g. Multiple Layer Perceptron) or configure the agent, and to perform a simulation to compete against other agents. In addition, the Meme War comes with a Map Editor to assist user designing a map with agents and buildings.