论文arXiv 2611.00405今天

Behavior Graphs：无监督图结构压缩机器人部署数据提升分析效率

研究提出Behavior Graphs将高维策略轨迹压缩为可解释有向图。用户研究显示其较视频分析降低40.3%时间需求与37.6%认知负荷；在43个数据集上进度预测MAE降低53.5%，有效支撑具身智能数据飞轮闭环。

AI 整理

仿真与训练
数据与采集

本文由 AI 根据公开内容整理，可能有误，请以原文为准：Behavior Graphs: Unsupervised Graph Structures for Interpreting Robot Deployment Data（arXiv 2611.00405）

核心内容

针对机器人部署产生海量高维Rollout数据难以分析、阻碍数据飞轮闭环的痛点，arXiv论文提出Behavior Graphs方法。该方法通过无监督学习识别相似行为模式并构建时序有向图，将复杂策略执行过程压缩为单一可解释结构。在三项真实任务用户研究中，18名参与者里有16人偏好该图谱而非原始视频，且得出相同结论的同时感知时间需求降低40.3%、努力程度降低37.6%。

技术层面，Behavior Graphs通过学习跨Rollout的行为模式相似性、建立时序连接及关联结果来构建图谱。除提升可解释性外，该结构还具备进度预测能力：在43个数据集测试中，相比最强基线模型，其预测进度与真实时序的秩相关性提升22.9%，平均绝对误差（MAE）降低53.5%。这为大规模机器人集群部署数据的结构化理解与采集决策提供了量化工具。

关键看点

分析效率：用户研究显示，使用Behavior Graphs分析部署数据比观看视频降低40.3%感知时间需求和37.6%认知负荷，且结论一致性相当
预测精度：在43个数据集上，该图谱使进度预测的平均绝对误差（MAE）较最强基线降低53.5%，秩相关性提升22.9%
构建机制：采用无监督学习识别跨Rollout相似行为模式，建立时序连接并关联任务结果，将高维轨迹压缩为单一可解释有向图
数据闭环：通过将非结构化部署数据结构化，帮助开发者精准定位失败模式与成功路径，指导后续数据采集以修补模型缺陷

工程与产业观察

该工作解决了具身智能数据飞轮中“数据多但洞察少”的工程瓶颈。Behavior Graphs将非结构化Rollout转化为结构化拓扑，不仅大幅降低人工审核成本，更以高精度进度预测量化训练成效。这对依赖大规模真机数据迭代的团队具有直接参考价值，是打通部署数据到算法优化闭环的关键基础设施。

英文摘要 (Abstract)

Data flywheels promise continuous improvement via insights from deployment data. Understanding how the deployed model succeeds and fails can steer subsequent data collection towards patching failures and reinforcing successes. However, robot deployments create large volumes of high-dimensional rollout data, which is challenging to understand and confidently ground data collection decisions in. To make deployment data easier to analyze and act on, we propose behavior graphs, which compress policy rollouts into a single interpretable, directed graph. We construct the graphs by learning to identify similar behavior modes across rollouts, connecting the behaviors temporally, and finally linking them to outcomes. In a user study on three real-world tasks, where participants analyzed rollouts to choose a data collection strategy, 16 of 18 preferred behavior graphs to rollout videos, reaching comparable conclusions with 40.3% lower perceived temporal demand and 37.6% lower effort. Beyond interpretability for deployment data, behavior graphs are capable progress predictors: across 43 datasets, they improve the rank correlation between predicted progress and true temporal order by 22.9% and reduce mean absolute error by 53.5% relative to the strongest baseline. As the number of deployed robots grows, behavior graphs lend structure to high-dimensional policy rollouts and make deployment data easier to interpret and reason about.

作者：Erik Bauer, Jakob Thumm, Christopher Agia, Rohan Sinha, Chen Xu, Haruki Nishimura, Jeannette Bohg, Marco Pavone, Masha Itkina