DynaVieW: Schema-Guided World Model for Video Scene Dynamics PredictionDynaVieW：基于模式引导的世界模型实现视觉动态预测与仿真
S 1.7T11 sources1 个来源R7-research
Researchers propose DynaVieW, a schema-guided world model optimized for predicting and simulating temporal visual scene evolution in videos and multi-image sequences.
The model learns interleaved state-transition sequences where states represent visual scenes from keyframes and transitions capture hierarchical dynamic constituents through a mixture-of-experts architecture.
DynaVieW jointly optimizes transition prediction and state simulation using cross-expert selective attention and schema token re-weighted loss to enable robust visual dynamics understanding.