Sim2Val: Leveraging Correlation Across Test Platforms for Variance-Reduced Metric Estimation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Luo, Rachel, Yang, Heng, Watson, Michael, Sharma, Apoorva, Veer, Sushant, Schmerling, Edward, Pavone, Marco
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916931600121856
author Luo, Rachel
Yang, Heng
Watson, Michael
Sharma, Apoorva
Veer, Sushant
Schmerling, Edward
Pavone, Marco
author_facet Luo, Rachel
Yang, Heng
Watson, Michael
Sharma, Apoorva
Veer, Sushant
Schmerling, Edward
Pavone, Marco
contents Learning-based robotic systems demand rigorous validation to assure reliable performance, but extensive real-world testing is often prohibitively expensive, and if conducted may still yield insufficient data for high-confidence guarantees. In this work we introduce Sim2Val, a general estimation framework that leverages paired data across test platforms, e.g., paired simulation and real-world observations, to achieve better estimates of real-world metrics via the method of control variates. By incorporating cheap and abundant auxiliary measurements (for example, simulator outputs) as control variates for costly real-world samples, our method provably reduces the variance of Monte Carlo estimates and thus requires significantly fewer real-world samples to attain a specified confidence bound on the mean performance. We provide theoretical analysis characterizing the variance and sample-efficiency improvement, and demonstrate empirically in autonomous driving and quadruped robotics settings that our approach achieves high-probability bounds with markedly improved sample efficiency. Our technique can lower the real-world testing burden for validating the performance of the stack, thereby enabling more efficient and cost-effective experimental evaluation of robotic systems.
format Preprint
id arxiv_https___arxiv_org_abs_2506_20553
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sim2Val: Leveraging Correlation Across Test Platforms for Variance-Reduced Metric Estimation
Luo, Rachel
Yang, Heng
Watson, Michael
Sharma, Apoorva
Veer, Sushant
Schmerling, Edward
Pavone, Marco
Robotics
Learning-based robotic systems demand rigorous validation to assure reliable performance, but extensive real-world testing is often prohibitively expensive, and if conducted may still yield insufficient data for high-confidence guarantees. In this work we introduce Sim2Val, a general estimation framework that leverages paired data across test platforms, e.g., paired simulation and real-world observations, to achieve better estimates of real-world metrics via the method of control variates. By incorporating cheap and abundant auxiliary measurements (for example, simulator outputs) as control variates for costly real-world samples, our method provably reduces the variance of Monte Carlo estimates and thus requires significantly fewer real-world samples to attain a specified confidence bound on the mean performance. We provide theoretical analysis characterizing the variance and sample-efficiency improvement, and demonstrate empirically in autonomous driving and quadruped robotics settings that our approach achieves high-probability bounds with markedly improved sample efficiency. Our technique can lower the real-world testing burden for validating the performance of the stack, thereby enabling more efficient and cost-effective experimental evaluation of robotic systems.
title Sim2Val: Leveraging Correlation Across Test Platforms for Variance-Reduced Metric Estimation
topic Robotics
url https://arxiv.org/abs/2506.20553