TuneNet: One-Shot Residual Tuning for System Identification and Sim-to-Real Robot Task Transfer

Adam Allevato, Elaine Schaertl Short, Mitch Pryor, Andrea Thomaz
Proceedings of the Conference on Robot Learning, PMLR 100:445-455, 2020.

Abstract

As researchers teach robots to perform more and more complex tasks, the need for realistic simulation environments is growing. Existing techniques for closing the reality gap by approximating real-world physics often require extensive real world data and/or thousands of simulation samples. This paper presents TuneNet, a new machine learning-based method to directly tune the parameters of one model to match another using an iterative residual tuning technique. TuneNet estimates the parameter difference between two models using a single observation from the target and minimal simulation, allowing rapid, accurate and sample-efficient parameter estimation. The system can be trained via supervised learning over an auto-generated simulated dataset. We show that TuneNet can perform system identification even when the true parameter values lie well outside the distribution seen during training, and demonstrate that simulators tuned with TuneNet outperform existing techniques for predicting rigid body motion. Finally, we show that our method can estimate real-world parameter values, allowing a robot to perform sim-to-real task transfer on a dynamic manipulation task unseen during training. Code and videos are available online at http://bit.ly/2lf1bAw.

Cite this Paper

BibTeX
@InProceedings{pmlr-v100-allevato20a, title = {TuneNet: One-Shot Residual Tuning for System Identification and Sim-to-Real Robot Task Transfer}, author = {Allevato, Adam and Short, Elaine Schaertl and Pryor, Mitch and Thomaz, Andrea}, booktitle = {Proceedings of the Conference on Robot Learning}, pages = {445--455}, year = {2020}, editor = {Kaelbling, Leslie Pack and Kragic, Danica and Sugiura, Komei}, volume = {100}, series = {Proceedings of Machine Learning Research}, month = {30 Oct--01 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v100/allevato20a/allevato20a.pdf}, url = {https://proceedings.mlr.press/v100/allevato20a.html}, abstract = {As researchers teach robots to perform more and more complex tasks, the need for realistic simulation environments is growing. Existing techniques for closing the reality gap by approximating real-world physics often require extensive real world data and/or thousands of simulation samples. This paper presents TuneNet, a new machine learning-based method to directly tune the parameters of one model to match another using an iterative residual tuning technique. TuneNet estimates the parameter difference between two models using a single observation from the target and minimal simulation, allowing rapid, accurate and sample-efficient parameter estimation. The system can be trained via supervised learning over an auto-generated simulated dataset. We show that TuneNet can perform system identification even when the true parameter values lie well outside the distribution seen during training, and demonstrate that simulators tuned with TuneNet outperform existing techniques for predicting rigid body motion. Finally, we show that our method can estimate real-world parameter values, allowing a robot to perform sim-to-real task transfer on a dynamic manipulation task unseen during training. Code and videos are available online at http://bit.ly/2lf1bAw.} }
Endnote
%0 Conference Paper %T TuneNet: One-Shot Residual Tuning for System Identification and Sim-to-Real Robot Task Transfer %A Adam Allevato %A Elaine Schaertl Short %A Mitch Pryor %A Andrea Thomaz %B Proceedings of the Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2020 %E Leslie Pack Kaelbling %E Danica Kragic %E Komei Sugiura %F pmlr-v100-allevato20a %I PMLR %P 445--455 %U https://proceedings.mlr.press/v100/allevato20a.html %V 100 %X As researchers teach robots to perform more and more complex tasks, the need for realistic simulation environments is growing. Existing techniques for closing the reality gap by approximating real-world physics often require extensive real world data and/or thousands of simulation samples. This paper presents TuneNet, a new machine learning-based method to directly tune the parameters of one model to match another using an iterative residual tuning technique. TuneNet estimates the parameter difference between two models using a single observation from the target and minimal simulation, allowing rapid, accurate and sample-efficient parameter estimation. The system can be trained via supervised learning over an auto-generated simulated dataset. We show that TuneNet can perform system identification even when the true parameter values lie well outside the distribution seen during training, and demonstrate that simulators tuned with TuneNet outperform existing techniques for predicting rigid body motion. Finally, we show that our method can estimate real-world parameter values, allowing a robot to perform sim-to-real task transfer on a dynamic manipulation task unseen during training. Code and videos are available online at http://bit.ly/2lf1bAw.
APA
Allevato, A., Short, E.S., Pryor, M. & Thomaz, A.. (2020). TuneNet: One-Shot Residual Tuning for System Identification and Sim-to-Real Robot Task Transfer. Proceedings of the Conference on Robot Learning, in Proceedings of Machine Learning Research 100:445-455 Available from https://proceedings.mlr.press/v100/allevato20a.html.

Related Material