Computer games are already a part of our lives. To have players engaged in gaming worlds, the demand for human-like Bot, new, novel and even player-customized and/or adaptive content keeps increasing, especially when matching players is required, and manual content production is expensive, unscalable, and not time-to-market.
Procedural Content Generation (PCG) for Games is something that generates game content through algorithmic means to effectively and efficiently increase gameplay experience with creative, variant and adaptive content, and reduce production cost and effort. In recent three years, we, GAME Lab, have proposed methods used for procedurally generating game content, such as game level and game object, with an emphasis on commercial applications. On game level generation, our methods control the gameplay experience with adaptive game levels for Match-3 Game, Slot Game, and Car Racing Game. Based on summarized rules from Chinese feature building properties, we develop a parameterized Chinese feature building generation tool and Chinese feature village tool for creating and evolving game content with Chinese style. Also, we devise a sword editing tool with learning capability to generate a series of swords in consistent with attributes of a game player.
Regarding to our research on Game Bot for Eight-ball and Mahjong, the proposed AI methods based on heuristic rules, MCTS, and machine learning with scalable game bot strength will share with you about the current research status and possible future research directions.
戴文凯博士于台湾交通大学资讯工程学系取得硕、博士学位，目前是台湾科技大学资讯工程学系教授。自获得计算机图形学(Computer Graphics)学位之后，持续二十多年来，主要致力于游戏多媒体产业之产学合作与相关实用技术之科技研发，以及主持游戏实验室GAME Lab培育游戏创新科研人才。目前所从事的产学合作项目包含游戏引擎、串流传输技术、实时绘制与动画、自动化建模与游戏内容生成、智能计算机视觉、多媒体系统开发等范畴，并实际应用于竞速类、动作类、休闲类、棋牌类、桌游等游戏。同时，累计发表了有百余篇国内外核心期刊与高水平国际会议论文，获取七项发明专利。另外，基于关注游戏产业的社会责任，戴教授亦积极协助编订台湾游戏产业白皮书，筹建产政学研游戏论坛、引导游戏业参与海内外游戏峰会，提升游戏关键技术研发能量。