Intactworld improves video world modeling with uncompressed features
IntactWorld: Joint World Modeling with Intact Features
Computer Vision and Pattern Recognition
Summary
Video generation models can create realistic images but often don’t fully understand how the real world works. The authors introduce IntactWorld, which keeps detailed features intact instead of compressing them, helping the model learn better world logic. They also developed a training method that reduces memory and speeds up processing by focusing on key feature tokens. This approach leads to better performance than previous methods on a standard benchmark.
What this means in practice
- •For video game developers: Improve the realism and logical consistency of in-game video scenes by using detailed, uncompressed feature modeling for dynamic environment generation.
- •For visual effects teams: Generate complex video scenes that better reflect real-world dynamics while reducing computational resources during rendering.
Authors
Boming Tan, Xiangdong Zhang, Yan Xia, Qi Zhu, Deyi Ji, Xue Yang, Shaofeng Zhang
Abstract
While recent video generation models synthesize highly realistic visuals, they lack a genuine understanding of intrinsic real-world logic. Existing methods attempt to understand the world by internalizing diverse world knowledge, yet constrained by computational overhead or dimensionality alignment, their learning processes inevitably compress features, causing a severe loss of structural information. To address this, we propose \textbf{IntactWorld}, a \textbf{Joint World Modeling Architecture} utilizing uncompressed \textbf{Intact Features}. Since data naturally reside on a low-dimensional manifold within a high-dimensional space, predicting the flow velocity $v$ within this uncompressed high-dimensional space induces a severe manifold gap. To successfully eliminate this optimization bottleneck, our framework instead predicts the clean feature $x_0$ at intermediate layers. Furthermore, to mitigate the computational overhead of incorporating complete world knowledge, we introduce a \textit{Full-to-Compact Training Paradigm}. By replacing raw full features with highly refined CLS tokens, this paradigm enables efficient single-branch guidance, reducing spatial memory consumption by 11.4\% and cutting inference latency by 43.8\%. Extensive evaluations demonstrate the effectiveness of IntactWorld, outperforming established baselines by 2.46 points on the VBench 2.0 benchmark.