# LASR指标量化世界模型潜变量动作支撑度

> 原文链接：https://robotics.hellobike.com/papers/2026-11-16-2611-17775-3139d6ea
> 频道：papers | 更新日期：2026-11-16 | 标签：auto, arxiv
> 核心摘要：针对无标注视频训练的世界模型难以区分动作编码与状态捷径问题，提出LASR无量纲评估指标。该指标在机器人操作数据上验证了动作信息检测能力，为具身智能离线评估提供标准化协议。

# LASR指标量化世界模型潜变量动作支撑度

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[Beyond Prediction Gain: LASR as a Metric for Action-Grounded Latents in World Models](https://arxiv.org/abs/2611.17775)（arXiv 2611.17775）

## 核心内容

针对世界模型从无标注视频学习时潜变量易混淆动作与状态预测捷径的难题，研究提出Latent Action Support Ratio（LASR）指标。该指标通过离线评估衡量潜变量对动作相关目标的预测支撑度是否超越当前状态及非动作替代项，解决了传统重建损失或下游任务性能无法隔离动作编码有效性的痛点，并在机器人操作数据集上完成了正负对照验证。

LASR包含GeoLASR（典型支撑度）与WorstLASR（最弱情况）两个互补维度，采用episode-held-out数据划分策略确保评估泛化性。实验表明该指标能准确识别真实动作信息并拒绝非动作捷径，支持对不完美潜变量进行分级比较。研究进一步利用LASR训练并审计轻量级残差适配器，通过配对比较评估不确定性，确立了世界模型动作表征的标准化离线评测协议。

## 关键看点

- **LASR指标定义**：无量纲度量潜变量在离线数据上对动作目标的预测支撑度，剔除状态与非动作捷径干扰
- **双维评估体系**：GeoLASR量化典型动作支撑水平，WorstLASR定位最差表现案例，实现潜变量质量分级
- **实验验证机制**：基于机器人操作数据构建正负对照组，证实指标可检出真实动作信息并排斥非动作预测捷径
- **模型优化应用**：利用LASR指导轻量级残差适配器训练与审计，通过配对比较量化点估计改进的不确定性

## 工程与产业观察

LASR填补了世界模型从纯视觉数据学习动作表征时的离线评估空白。对于依赖大规模无标注视频预训练的具身算法团队，该指标提供了比下游任务更细粒度的表征诊断工具，有助于在部署前排查潜变量是否真正编码了可控动作而非环境统计规律，提升Sim-to-Real迁移的可解释性与安全性。

## 英文摘要 (Abstract)

World models that support planning, control, and interpretable reasoning require latent representations that genuinely encode action. Yet the large-scale video data from which such models are learned overwhelmingly lack action annotations. Latent-action models infer action-like variables from observation sequences, enabling action-aware learning from unlabeled videos or trajectories. A fundamental concern, however, remains: a latent that improves future prediction may simply encode the current state, episode progress, task identity, or other predictive shortcuts. Existing evaluation practices---reconstruction loss, rollout fidelity, action recovery, or downstream performance---cannot isolate this distinction. We introduce the Latent Action Support Ratio (LASR), a dimensionless metric that measures, under offline evaluation on held-out data, whether a learned latent provides predictive support for action-related targets beyond the current state and non-action alternatives. LASR provides two complementary summaries: GeoLASR captures typical support, while WorstLASR identifies the weakest case. We instantiate the LASR benchmark with recorded actions and corresponding physical state changes as targets, episode-held-out splits, one predictor type for all comparisons, and positive and negative controls. Experiments on recorded robot manipulation data and a matched multi-task panel demonstrate that LASR detects known action information, rejects representative non-action shortcuts, and enables graded comparisons among imperfect latent representations. We further use LASR to train and audit a lightweight residual adapter, obtaining descriptively better point estimates whose uncertainty is assessed with paired comparisons. LASR thus provides a principled offline protocol for measuring action-related predictive support in world-model latents.

作者：Yu Zhou, Changsong Lei, Yixiang Wei, Kairui Wen, Huankun Sheng, Ming Li, Yeying Fan, Yu-Hui Wen, Yong-Jin Liu
:::

## 原始参考来源 (Sources)

- [arXiv 2611.17775：Beyond Prediction Gain: LASR as a Metric for Action-Grounded Latents in World Models](https://arxiv.org/abs/2611.17775)