# LARBRID算法解决扩散策略微调噪声混叠，机器人任务性能提升3倍

> 原文链接：https://robotics.hellobike.com/papers/2026-11-06-2611-07275-02b8c440
> 频道：papers | 更新日期：2026-11-06 | 标签：auto, arxiv
> 核心摘要：针对预训练扩散策略RL微调效率低及噪声混叠难题，提出LARBRID算法。通过LNL-MDP框架与Bridge Equation对齐噪声与动作空间表征，在机器人运动与操作任务中实现2-3倍性能提升，代码已开源。

# LARBRID算法解决扩散策略微调噪声混叠，机器人任务性能提升3倍

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[Efficient Diffusion Policy Fine Tuning with Latent Noise Representation Bridging](https://arxiv.org/abs/2611.07275)（arXiv 2611.07275）

## 核心内容

针对预训练扩散策略在强化学习微调中面临的计算昂贵、样本效率低及缺乏理论保证等挑战，研究团队提出Latent Representation Bridging for Diffusion（LARBRID）算法。该方法旨在解决现有初始噪声微调技术中因噪声空间与动作空间缺乏自然映射而导致的“噪声混叠”问题，为具身智能策略适配提供了新的理论支撑与工程解法。

技术上引入潜在噪声线性马尔可夫决策过程（LNL-MDP），推导出Bridge Equation以在共享表征下对齐噪声值与动作值，消除混叠并满足FineTuneRL理论的近最优保证。该框架通过优化初始分布引导预训练模型实现环境最优行为，在多样化的机器人运动控制与机械臂操作基准测试中，取得了相比基线方法2至3倍的性能增益，相关代码已在GitHub开源。

## 关键看点

- **理论创新**：提出LNL-MDP框架并推导Bridge Equation，从数学层面消除噪声混叠，确保微调过程具备FineTuneRL近最优理论保证
- **性能指标**：在机器人运动与操作任务基准测试中，LARBRID算法使预训练扩散策略性能提升2-3倍，显著优于传统RL微调方案
- **核心机制**：通过学习潜在表征对齐噪声与动作空间，优化初始噪声分布以引导预训练模型，规避了高成本的时间反向传播计算
- **开源资产**：完整算法实现代码已在github.com/Maincode-Research/LARBRID.git开源，支持社区复现与二次开发

## 工程与产业观察

该工作为具身大模型落地提供了关键微调范式。通过解决扩散策略RL适配中的噪声混叠痛点，大幅降低了将通用预训练模型迁移至特定机器人任务的算力与数据门槛。其理论完备性与显著的性能增益，使其成为当前具身智能算法工程师优化操作策略的重要参考基线。

## 英文摘要 (Abstract)

Pre-trained diffusion and flow policies have emerged as powerful backbones for robotics, yet adapting them to novel data via reinforcement learning (RL) remains computationally expensive, sample inefficient and lacks theoretical guarantees. Prior approaches that fine-tune the initial noise bypass costly backpropagation through time but struggle with the "noise aliasing" problem: there does not exista natural link between the noise space value and the action space value. This paper resolves these challenges by introducing the Latent Noise Linear Markov Decision Process (LNL-MDP). We derive the Bridge Equation for LNL-MDPs which eliminates the noise aliasing problem by aligning values with a shared representation between the noise and the action space. Furthermore, our LNL-MDP fine tuning framework falls under the FineTuneRL theory with near-optimal guarantee. Leveraging these insights, we propose Latent Representation Bridging for Diffusion (LARBRID), an algorithm that efficiently learns the representation to optimize the initial distribution that steers the pre-trained model to act optimally in the environment. It yields up to 2-3$\times$ performance increases on challenging control problems across a diverse suite of robotic locomotion and manipulation tasks. Code can be found at: github.com/Maincode-Research/LARBRID.git.

作者：Yuanchu Liang, Yue Yang, An-Chi He, Hanna Kurniawati
:::

## 原始参考来源 (Sources)

- [arXiv 2611.07275：Efficient Diffusion Policy Fine Tuning with Latent Noise Representation Bridging](https://arxiv.org/abs/2611.07275)