Evo-World: A Lightweight Vision-Language-Action Model with Future Motion Prediction via Optical-Flow Supervision

Key points

• Vision-language-action (VLA) models provide a unified framework for visual understanding, language grounding, and robot control, yet anticipating how the scene will evolve beyond the current observation can provide additional information for control.1
• We further design the training procedure to preserve these predictive representations during policy learning, regulating representation diversity while decoupling representation learning from action optimization.1
• On real-world AgileX robots, Evo-World outperforms the evaluated baselines across single-arm and bimanual manipulation tasks while using only 0.87B parameters and delivering lower GPU memory usage and higher inference efficiency.1
• These results demonstrate that future motion prediction through optical-flow supervision provides an effective and lightweight approach to enhancing the predictive capabilities of VLA models.1

Sources (1)

• [1]Evo-World: A Lightweight Vision-Language-Action Model with Future Motion Prediction via Optical-Flow Supervision
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 03:38 AM

  To address these limitations, we introduce Evo-World, a lightweight VLA model that augments policy learning with future motion prediction through optical-flow supervision, thereby providing a task-relevant predictive signal focused on motion rather than appearance.

  Vision-language-action (VLA) models provide a unified framework for visual understanding, language grounding, and robot control, yet anticipating how the scene will evolve beyond the current observation can provide additional information for control.

Extractive summary: sentences quoted from the sources.