Intervention-Based Movability Representations in Minimalist Embodied Predictive Agents
Year: 2026
Type: preprint
Abstract: This preprint studies whether minimalist embodied predictive agents can form movability-sensitive internal representations from intervention-driven experience. The work uses a 2D embodied world, vector causal perturbation diagnostics, causal-swap tests, blind generalization, temporal and noise perturbations, visual shortcut tests, distillation, and 3D stress tests. The central result is cautious: shortcut-deprived sensor agents form stable movability-like representations, while raw visual and 3D settings expose major unresolved limits.Version 2 updates the manuscript PDF and source archives so that the title page includes the Zenodo DOI. The scientific content is unchanged. Version 3 updates the title page to use the stable Zenodo concept DOI. The scientific content is unchanged.
Author: Yugen
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7165045046
Cites:
Cited by:
Related to: