Search papers, labs, and topics across Lattice.
This paper introduces WorldMind, a novel decoupled framework designed to enhance state-aware behavior in non-player characters (NPCs) within game world models. By separating the interactive world modeling into four distinct layers鈥擴nderstanding, Decision, Control, and Generation鈥擶orldMind allows for a more responsive NPC behavior that is grounded in the evolving game state. Experiments on the newly created BOSS-140K dataset reveal that WorldMind outperforms existing models in approximately 70% of comparisons, showcasing its ability to produce more tactically appropriate and coherent NPC actions.
WorldMind achieves a breakthrough in NPC behavior by decoupling state understanding from action generation, leading to significantly more coherent and context-aware interactions in gameplay.
Game world models have recently demonstrated promising capabilities in generating visually coherent and action-controllable gameplay videos. However, non-player character (NPC) behavior in existing models is either implicitly entangled with video generation or explicitly prescribed through external control signals. Consequently, a game world model has to jointly understand the state, plan the NPC's response and render its visual outcome, limiting its ability to produce responsive and state-aware NPC behavior. The challenge lies in the lack of an explicit interface for state-grounded decision-making. To this end, we introduce WorldMind, to our knowledge the first decoupled framework for state-aware NPC behavior in game world models. WorldMind separates interactive world modeling into four layers: an Understanding Layer that constructs a compact state from generated frames; a Decision Layer that reasons over the compact state to plan the NPC's next action; a Control Layer that translates the actions into temporally aligned conditions; and a Generation Layer that synthesizes their visual outcomes. By reconnecting layers in a closed interaction loop, WorldMind grounds NPC behavior in the evolving game state. We further introduce BOSS-140K, a dataset of gameplay videos paired with rich internal game states, together with an agent that automates the collection at scale. Experiments on BOSS-140K demonstrate reliable compact state reconstruction and mechanics-grounded planning, with WorldMind preferred over the baselines in approximately 70% of pairwise comparisons for its more tactically appropriate and coherent NPC behavior. Project page: https://teawhite.cn/worldmind_projectpage/
