cs.ROOct 30, 2026

Behavior Graphs: Unsupervised Graph Structures for Interpreting Robot Deployment Data

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

Organizations: Stanford University · Toyota Research Institute · NVIDIA Research

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.

Figures & tables

Explore similar work

CardsList
  1. GGraph: A GPU-Accelerated Library for Graph-Based Robot Deployment and Learning

    Oct 20, 2026Yash Rampuria, Rohan Garg, Daniele Gammelli +1Robot LearningRobotic RL

  2. Deployment-Time Reliability of Learned Robot Policies

    Mar 12, 2026Christopher AgiaRoboticsRobotic Policy Evaluation

  3. SIR: Structured Image Representations for Explainable Robot Learning

    Jun 29, 2026Paul Mattes, Jan Schwab, Jens Bosch +5Robot Policy LearningGraph Representation Learning