Papers

Dune:面向 VLA 测试时适配的零阶优化器,用历史误差回归降低梯度估计方差

Dune: A Zeroth-Order Optimizer with Time-Conditioned Historical Error Regression for VLA Test-Time Adaptation

Why we picked it

机器人端部署 VLA 模型时,反向传播太吃内存,纯零阶方法又噪声大。这篇论文用历史误差预测融合当前估计来压方差,不用反向传播也不用额外前向,还在 LIBERO 和真机上验证了。做机器人部署和端侧适配的工程师值得看方法细节。

论文提出 Dune,一个用于 Vision-Language-Action(VLA)模型测试时适配(TTA)的零阶(ZO)优化器。核心问题是:基于反向传播的适配在资源受限机器人上内存开销大,而零阶方法在扰动预算有限时估计方差高,长期更新容易停滞或退化。

Body

Dune 的机制是从过往优化动态中学习一个轻量的时间条件化误差预测器,将预测结果与即时 ZO 扰动观测融合,从而在不做反向传播、不增加前向传播的前提下实现方差缩减。论文还从理论上刻画了高维小样本场景下估计方差对收敛的制约。

实验在 LIBERO 基准和真实机器人上验证,Dune 在低内存占用下取得更高的成功率,消融实验确认了闭环稳定性和受扰后恢复能力的提升。

Original · arXiv cs.AI

Dune: A Zeroth-Order Optimizer with Time-Conditioned Historical Error Regression for VLA Test-Time Adaptation

Test-Time Adaptation (TTA) enables Vision-Language-Action (VLA) models to handle distribution shifts during deployment without source-domain retraining. However, backpropagation-based adaptation incurs heavy memory overhead on resource-constrained robots, whereas zeroth-order (ZO) methods suffer from high estimation variance under limited perturbation budgets. We observe that while ZO estimation enables low-memory adaptation with rapid initial gains, it often stalls or degrades over long-horizon updates due to noisy gradient estimates. To address these limitations, we propose Dune, a zeroth-order optimizer that stabilizes long-horizon adaptation by coupling online historical error prediction with current perturbation observations. Specifically, Dune learns a lightweight, time-conditioned error predictor from past optimization dynamics and fuses its predictions with instantaneous ZO estimates, achieving effective variance reduction without backpropagation or extra forward passes. Additionally, we theoretically characterize how estimation variance restricts convergence in high-dimensional, small-population regimes, motivating the use of historical statistics. Extensive experiments on LIBERO and real-world robots demonstrate that Dune achieves superior success rates under a low memory footprint, with ablations confirming improved closed-loop stability and post-disturbance recovery.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include VLA, zeroth-order optimization, Test-Time Adaptation, LIBERO.

Source attribution: curated from arXiv cs.AI (@Chengyang Dong, Ze Chen, Kai Ye, Chutian Wang, Heng Xu, Wei Mo, Maomao Li, Jiayan Xu, Jiayuan Wu, Yihao Xiao, Linhua Ye, Zhonghua Peng, Ruize Han, Jialong Sun).