# BestRand算法实现VLA评测基准低成本压缩与误差抑制

> 原文链接：https://robotics.hellobike.com/papers/2026-10-31-2611-00883-cd0f92ec
> 频道：papers | 更新日期：2026-10-31 | 标签：auto, arxiv, hf-daily-papers
> 核心摘要：针对具身智能模型评测成本高昂问题，提出BestRand基准压缩方法。在70+基准及VLA任务验证中，该方法将高误差风险降低50.2%，平均估计误差减少16.4%，构建成本仅为竞品51.3%，有效解决随机采样长尾不稳定难题。

# BestRand算法实现VLA评测基准低成本压缩与误差抑制

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[Random Is Almost Enough? Stable and Low-Cost Benchmark Compression via BestRand](https://arxiv.org/abs/2611.00883)（HF Daily Papers · arXiv 2611.00883）

## 核心内容

针对基础模型与VLA评测耗时痛点，研究揭示子集估计误差呈长尾分布，纯随机采样虽高效但存在罕见大误差风险。为此提出BestRand方法，在保留随机采样效率的同时显著抑制高误差尾部。实验覆盖70+文本、多模态及Agent基准并扩展至VLA评估，证实该方案兼具稳定性与低成本优势。

BestRand通过优化采样策略克服现代基础模型固有的评测不稳定性。对比现有竞争方法，其高误差风险降低50.2%，平均估计误差减少16.4%，且构建成本仅需51.3%。在分数范围、模型族及时间偏移等OOD场景下，平均估计误差进一步降低21.8%。相关代码与数据已在GitHub开源，为具身智能评测提供可靠压缩工具。

## 关键看点

- **VLA评测适配**：方法从通用基准扩展至VLA评估，解决具身模型全量评测成本高企问题，保障小样本下评分可靠性。
- **误差风险控制**：相比竞争方法，高误差风险降低50.2%，平均估计误差减少16.4%，有效消除随机采样的长尾失效缺陷。
- **构建成本优势**：在提升稳定性的同时，基准构建成本仅为同类方法的51.3%，大幅降低评测数据集筛选与验证开销。
- **OOD泛化能力**：在分数范围、模型家族及时间维度偏移场景下，平均OOD估计误差降低21.8%，适应快速迭代的模型生态。

## 工程与产业观察

随着VLA等多模态大模型成为具身智能核心，全量评测算力与时间成本呈指数级增长。BestRand以极低边际成本解决了基准压缩的长尾风险，为算法团队提供了高性价比的迭代验证工具。其开源特性使其能快速集成至现有MLOps流程，对加速具身智能模型选型与训练闭环具有直接工程价值。

## 英文摘要 (Abstract)

Benchmark compression reduces the often prohibitive time and cost of model evaluation by evaluating models on a small benchmark subset that closely matches full-benchmark results. To quantify estimation error, we derive an accurate analytical approximation to the error distribution when estimating full-benchmark scores from a subset, showing that most subsets yield small errors, while a few produce much larger errors, resulting in a long-tailed distribution. This means random sampling alone can achieve surprisingly strong performance with high probability, but remains unstable due to rare large errors. We further show that this instability is inherent to modern foundation models and persists even as more models are evaluated. Based on these findings, we introduce BestRand, a stable and low-cost benchmark compression method that preserves the efficiency of random sampling while substantially reducing its high-error tail. We validate BestRand on 70+ text, multimodal, and agent benchmarks, and further extend it to VLA evaluation. Compared with competitive methods, BestRand reduces high-error risk by 50.2% and estimation error by 16.4% on average, while requiring only 51.3% of the construction cost. It further reduces OOD estimation error by 21.8% on average across score-range, model-family, and temporal shifts. Together, our theory provides practical guidance for understanding and predicting the accuracy and reliability of benchmark compression, while BestRand enables stable and accurate compression at low cost. The code and data are publicly available onhttps://github.com/George-Ling3/BestRand.

Hugging Face Daily Papers 热度：38 upvotes
作者：George Ling, Shanshan Zhong, Lijin Yang, Hao Yang, Nanxiang Li, Zhongzhan Huang
:::

## 原始参考来源 (Sources)

- [HF Daily Papers · arXiv 2611.00883：Random Is Almost Enough? Stable and Low-Cost Benchmark Compression via BestRand](https://arxiv.org/abs/2611.00883)