DexWorldModel: Causal Latent World Modeling towards Automated Learning of Embodied Tasks
For robotics manipulation, this work provides a more efficient and generalizable world model that reduces memory footprint and inference latency while enabling robust domain transfer.
The paper introduces the Causal Latent World Model (CLWM) to address bottlenecks in generative world-action models for manipulation, achieving state-of-the-art performance in dual-arm simulation and zero-shot sim-to-real transfer, outperforming baselines finetuned on real-world data.
Deploying generative World-Action Models for manipulation is severely bottlenecked by redundant pixel-level reconstruction, $\mathcal{O}(T)$ memory scaling, and sequential inference latency. We introduce the Causal Latent World Model (CLWM), which employs DINOv3 features as generative targets to disentangle interaction semantics from visual noise, yielding highly robust domain generalization. To overcome memory scaling, CLWM features a Dual-State Test-Time Training (TTT) Memory that guarantees a strict $\mathcal{O}(1)$ footprint for long-horizon tasks. To overcome deployment latency, we propose Speculative Asynchronous Inference (SAI) to mask partial diffusion denoising behind physical execution, cutting blocking latency by about $50\%$. To scale robust policies, we present EmbodiChain, an online framework that establishes the Efficiency Law by injecting an infinite flow of physics-grounded trajectories during training. Extensive experiments validate that CLWM achieves state-of-the-art performance in complex dual-arm simulation and unprecedented zero-shot sim-to-real transfer on physical robots, outperforming baselines explicitly finetuned on real-world data.