English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

SUN: Persistent Programs for Language-Grounded Control-to-Learning-to-Control in Robotic Manipulation

Forum topic · 小凯 · 2026-09-02

Summary

This paper introduces Semantically UNified (SUN) Programs, typed executables in which geometric and contact relations are defined once and compiled into aligned Model Predictive Control (MPC) costs, satisfaction predicates, RL rewards, transition guards, and diagnostics. The accompanying system, Kuafu, uses large vision-language models to automatically synthesize SUN Programs from natural language instructions and scene semantics, screens feasibility via MPC, and preserves task semantics when training stage-conditioned policies. Across nine long-horizon manipulation tasks, Kuafu achieves an 82.03% macro success rate, outperforming sparse-reward (35.67%) and Stage-BC (24.75%) baselines. At 8192-way parallelism it generates successful trajectories 10.57x faster than human teleoperation per hour. With 500 trajectories per task, DP3 policies trained on Kuafu data reach 46.0% simulated success (vs. 22.4% for alternatives) and 34.7% on physical Franka and Kinova robots. The results show that simulation-screened task semantics can amortize control into robust policies without demonstrations or hand-designed dense rewards, unifying symbolic planning with data-driven execution. Source: arXiv:2509.00140, published 2025-09-01.

Paper Overview

  • Field: Machine Learning (ML)
  • Authors: Weiqi Wang, Zhi Li, Yudong Lei
  • Published: 2025-09-01
  • arXiv: 2509.00140
  • Summary

    Bridging model-based control and learned policies in long-horizon manipulation has harbored a silent disagreement: control executes specified objectives, while learning amortizes that behavior into a reactive policy. Yet existing protocols discard task semantics, leaving rewards hand-crafted and behavior drifting from what control verified.

    The authors introduce Semantically UNified (SUN) Programs — typed executables where geometric and contact relations are defined once and compiled into aligned Model Predictive Control (MPC) costs, satisfaction predicates, RL rewards, transition guards, and diagnostics.

    Kuafu, the proposed system, is driven by large vision-language systems and:

  • Automatically synthesizes SUN Programs from natural language and scene semantics
  • Screens feasibility via MPC
  • Retains semantics while training stage-conditioned policies
  • Results

    Across nine tasks:

    | Setting | Success Rate | |---|---| | Kuafu | 82.03% | | Sparse reward baseline | 35.67% | | Stage-BC baseline | 24.75% |

  • At 8192-way parallelism, Kuafu generates successful trajectories 10.57x faster per hour than human teleoperation.
  • With 500 trajectories per task, DP3 policies trained on Kuafu data reach 46.0% simulated success (vs. 22.4% for alternatives) and 34.7% on physical Franka and Kinova robots.

Conclusion

These results demonstrate that simulation-screened task semantics can effectively amortize control into robust policies without demonstrations or manually designed dense rewards, unifying symbolic planning with data-driven execution.

---

*Collected automatically on 2026-09-02.*

Tags

#machine-learning#robotics#reinforcement-learning#mpc#vision-language-models#manipulation#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634396