GameFactorly：利用生成式交互视频创造新游戏
TLDR
GameFactory通过多阶段训练与领域适配器，生成开放域动作可控的游戏视频。
评分理由
The paper presents a novel framework for action-controlled game video generation with scene generalizability, leveraging pre-trained video diffusion models and a dedicated dataset. Its strengths include a clear methodology and experimental results, but it lacks explicit connection to established world model concepts and real-world deployment beyond game video generation.
Read-first 评分解释
综合优先阅读分 41.1，由主题、引用、图谱、方法、可复现性和近期性等信号加权得到。 原始总分保留为 20。
研究版图角色
前沿论文
排序敏感性
稳定性：volatile；排名波动范围：298。