Cosmos-Surg-dVRK: World Foundation Model-based Automated Online Evaluation of Surgical Robot Policy Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zbinden, Lukas, Nelson, Nigel, Chen, Juo-Tung, Chen, Xinhao, Kim, Ji Woong, Azizian, Mahdi, Krieger, Axel, Huver, Sean
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911247305277440
author Zbinden, Lukas
Nelson, Nigel
Chen, Juo-Tung
Chen, Xinhao
Kim, Ji Woong
Azizian, Mahdi
Krieger, Axel
Huver, Sean
author_facet Zbinden, Lukas
Nelson, Nigel
Chen, Juo-Tung
Chen, Xinhao
Kim, Ji Woong
Azizian, Mahdi
Krieger, Axel
Huver, Sean
contents The rise of surgical robots and vision-language-action models has accelerated the development of autonomous surgical policies and efficient assessment strategies. However, evaluating these policies directly on physical robotic platforms such as the da Vinci Research Kit (dVRK) remains hindered by high costs, time demands, reproducibility challenges, and variability in execution. World foundation models (WFM) for physical AI offer a transformative approach to simulate complex real-world surgical tasks, such as soft tissue deformation, with high fidelity. This work introduces Cosmos-Surg-dVRK, a surgical finetune of the Cosmos WFM, which, together with a trained video classifier, enables fully automated online evaluation and benchmarking of surgical policies. We evaluate Cosmos-Surg-dVRK using two distinct surgical datasets. On tabletop suture pad tasks, the automated pipeline achieves strong correlation between online rollouts in Cosmos-Surg-dVRK and policy outcomes on the real dVRK Si platform, as well as good agreement between human labelers and the V-JEPA 2-derived video classifier. Additionally, preliminary experiments with ex-vivo porcine cholecystectomy tasks in Cosmos-Surg-dVRK demonstrate promising alignment with real-world evaluations, highlighting the platform's potential for more complex surgical procedures.
format Preprint
id arxiv_https___arxiv_org_abs_2510_16240
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Cosmos-Surg-dVRK: World Foundation Model-based Automated Online Evaluation of Surgical Robot Policy Learning
Zbinden, Lukas
Nelson, Nigel
Chen, Juo-Tung
Chen, Xinhao
Kim, Ji Woong
Azizian, Mahdi
Krieger, Axel
Huver, Sean
Robotics
The rise of surgical robots and vision-language-action models has accelerated the development of autonomous surgical policies and efficient assessment strategies. However, evaluating these policies directly on physical robotic platforms such as the da Vinci Research Kit (dVRK) remains hindered by high costs, time demands, reproducibility challenges, and variability in execution. World foundation models (WFM) for physical AI offer a transformative approach to simulate complex real-world surgical tasks, such as soft tissue deformation, with high fidelity. This work introduces Cosmos-Surg-dVRK, a surgical finetune of the Cosmos WFM, which, together with a trained video classifier, enables fully automated online evaluation and benchmarking of surgical policies. We evaluate Cosmos-Surg-dVRK using two distinct surgical datasets. On tabletop suture pad tasks, the automated pipeline achieves strong correlation between online rollouts in Cosmos-Surg-dVRK and policy outcomes on the real dVRK Si platform, as well as good agreement between human labelers and the V-JEPA 2-derived video classifier. Additionally, preliminary experiments with ex-vivo porcine cholecystectomy tasks in Cosmos-Surg-dVRK demonstrate promising alignment with real-world evaluations, highlighting the platform's potential for more complex surgical procedures.
title Cosmos-Surg-dVRK: World Foundation Model-based Automated Online Evaluation of Surgical Robot Policy Learning
topic Robotics
url https://arxiv.org/abs/2510.16240