Answers are grounded in this paper and cite their sources.
World models have shown substantial potential for learning environmental dynamics and predicting future states. However, as environments become more complex and interactive, intelligent systems must make reliable decisions under uncertainty, anticipate alternative outcomes, and adapt as conditions change. This calls for world models that connect prediction with reasoning, action, and continual adaptation. Existing efforts remain fragmented and lack a systematic synthesis. To address this gap, this paper proposes world model 2.0 as a general cognitive architecture for complex interactive environments, develops a taxonomy of six core capabilities, and examines their integration within this architecture. It reviews representative methods and applications, analyzes their relationships, and summarizes key challenges and future research directions.
No takes yet. Share an insight, caveat, or question.
Qiu et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: