Register-Routed Delayed Fusion: Rewiring Shortcut-Prone Observation Fusion in Visuomotor Imitation

Published October 5, 2026 · Category: Robotics

Overview

arXiv:2610.02813v1 Announce Type: new Abstract: Visuomotor imitation policies combine high-dimensional visual observations with compact signals such as proprioception, and their fusion topology determines when and through which tokens these streams interact. In dense token fusion, visual tokens may attend directly to compact tokens from the first encoder layer, allowing action-predictive compact cues to influence spatial visual representations early in their formation. We ask whether controlling this route improves visual responsiveness and policy behavior. We introduce Register-Routed Delayed Fusion (RRDF), which masks direct compact-visual attention and stages cross-modal interaction through a learned register workspace. Its isolate-collect-route schedule protects an early stream-separated prefix and later permits only register-mediated exchange, while compact conditioning remains available to the native action generator. Across five simulation tasks and three real-robot tasks, RRDF matches or improves dense ACT under nominal conditions. Appearance-shift evaluations on four simulation tasks and held-out-position evaluations on three real-robot tasks also favor RRDF. Phase-matched input probes show lower measured state-to-image sensitivity, while ablations indicate that adding registers alone does not reproduce the full performance gain. These results support controlling cross-modal propagation while retaining compact action conditioning.

Source

Originally published at arxiv.org.

Related Articles

• Around the World: Unified Learned Locomotion on a 270 g Continuous-Rotation Quadruped
• RoboChemGym: A Protocol-Driven Generative Simulation Framework for Long-Horizon Chemical Manipulation
• GeoScaffold: Learning Compact Geometric Latents via Reconstruction for Efficient Vision-Language Navigation