Skip to content

Event

Authors propose Causal World Action Model (CWAM)

1 reports · 1 sources

Juntao Cheng and other authors propose CWAM (Causal World Action Model): a sparse routing path is added to the action head of a pretrained world action model, separating the latent variables that truly drive actions from the nuisance variables on which scene prediction depends, while the world modeling objective remains unchanged. The authors claim to have proven that, under sparse constraints, the latent variables used to generate actions are merely an invertible function of the action-relevant latent variables, and that the true structure can be identified up to permutation of the latent variables, without requiring sufficient temporal dependence and without making assumptions about the evolution of unobserved states. Synthetic data and real-robot experiments validate the theory and show improvements in closed-loop success rate efficiency.

Report timeline

  1. CausalWAM: Learning the Causal Sources of Actions in World Action Models

    The authors propose CWAM (Causal WAM), which adds sparse routing paths to the action head of a world action model to separate the latent variables that truly drive actions from the nuisance variables that scene prediction depends on, while keeping the world modeling objective unchanged. The paper proves that under a sparsity constraint, the latent variables used to generate actions are an invertible function of the action-relevant latent variables alone, and the true structure is identifiable up to a permutation of the latent variables, without requiring sufficient temporal dependence or assumptions about how unobserved states evolve. Synthetic data and real-robot experiments validate the theory and show improved closed-loop success rate efficiency.

    arXiv cs.RO ·

Related topics

Robot models and learningData and simulationResearch papers

Interest trend

Current
3
Peak
4Nov 4 01:00
Past 24 hours
–

The trend compares participants observed throughout the same period, so its range may be smaller than the current interest count. Move or tap to inspect each hour; use the arrow keys on a keyboard.