arXiv cs.AI· Erik Bauer, Jakob Thumm, Christopher Agia, Rohan Sinha, Chen Xu, Haruki Nishimura, Jeannette Bohg, Marco Pavone, Masha Itkina·· 3 小时前AI 评分42
行为图：用于解读机器人部署数据的无监督图结构
Behavior Graphs: Unsupervised Graph Structures for Interpreting Robot Deployment Data
AI 导读
研究者提出行为图（behavior graphs），将机器人策略 rollout 压缩为单一可解释的有向图，帮助分析部署数据并指导数据采集。用户研究中，18 人里有 16 人更偏好行为图而非 rollout 视频，感知时间需求降低 40.3%、努力程度降低 37.6%。在 43 个数据集上，行为图还将进度预测的秩相关性提升 22.9%、平均绝对误差降低 53.5%，优于最强基线。
正文 · 原文
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.
| Comments: | 8 pages, 6 figures, 3 tables. Project website: this https URL |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2611.00405 [cs.RO] |
| (or arXiv:2611.00405v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2611.00405 arXiv-issued DOI via DataCite |
Submission history
From: Erik Bauer [view email]
[v1]
Fri, 30 Oct 2026 21:51:01 UTC (2,030 KB)
来源：arXiv cs.AI · arxiv.org