Safety
Beyond Actions: Learning Future Operation Targets for Vision-Language-Action Models
arXiv:2610.13212v1 Announce Type: cross Abstract: Actions specify what a robot should execute, but do not describe what those commands will achieve. Vision-language-action (VLA) policies nevertheless
arXiv:2610.13212v1 Announce Type: cross Abstract: Actions specify what a robot should execute, but do not describe what those commands will achieve. Vision-language-action (VLA) policies nevertheless learn primarily by predicting action chunks from images, language, and state. We introduce Action--Future Coupled Effect Learning (AFCE), which makes a future-grounded physical Effect the policy's primary prediction target. An offline tokenizer constructs temporally ordered Effect tokens from demonstrated actions, realized robot responses, and sparse visual transitions. Reconstruction objectives encourage the representation to retain executable detail and observed physical change. A flow-matching VLA predicts Effects from current images, instruction, and robot state, while a state-conditioned decoder converts them into action chunks. Embodiment-specific interfaces allow single- and dual-arm tasks to share the Effect model. On DexJoCo, AFCE achieves 50.97% overall success, improving over pi_{0.5} and UVT by 11.82 and 15.21 percentage points, respectively. It leads on all five bimanual tasks, highlighting the value of Effect supervision for coordinated dexterous manipulation. The project page is available at https://afceproject.github.io/AFCE-Web/.
Related
- Where Predictive Supervision Goes Shapes What VLA Policies Learn
- PAVE: Predictive Alignment and Value-Guided Evolution for World-Action Policies
- VLAFlow: A Unified Training Framework for Vision-Language-Action Models via Co-training and Future Latent Alignment
Source: arXiv cs.AI | 2026-10-12