Learning Contact Dynamics using Physically Structured Neural Networks

Andreas Hochlehnert, Alexander Terenin, Steindor Saemundsson, Marc Deisenroth
Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, PMLR 130:2152-2160, 2021.

Abstract

Learning physically structured representations of dynamical systems that include contact between different objects is an important problem for learning-based approaches in robotics. Black-box neural networks can learn to approximately represent discontinuous dynamics, but they typically require large quantities of data and often suffer from pathological behaviour when forecasting for longer time horizons. In this work, we use connections between deep neural networks and differential equations to design a family of deep network architectures for representing contact dynamics between objects. We show that these networks can learn discontinuous contact events in a data-efficient manner from noisy observations in settings that are traditionally difficult for black-box approaches and recent physics inspired neural networks. Our results indicate that an idealised form of touch feedback—which is heavily relied upon by biological systems—is a key component of making this learning problem tractable. Together with the inductive biases introduced through the network architectures, our techniques enable accurate learning of contact dynamics from observations.

Cite this Paper


BibTeX
@InProceedings{pmlr-v130-hochlehnert21a, title = { Learning Contact Dynamics using Physically Structured Neural Networks }, author = {Hochlehnert, Andreas and Terenin, Alexander and Saemundsson, Steindor and Deisenroth, Marc}, booktitle = {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics}, pages = {2152--2160}, year = {2021}, editor = {Banerjee, Arindam and Fukumizu, Kenji}, volume = {130}, series = {Proceedings of Machine Learning Research}, month = {13--15 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v130/hochlehnert21a/hochlehnert21a.pdf}, url = {https://proceedings.mlr.press/v130/hochlehnert21a.html}, abstract = { Learning physically structured representations of dynamical systems that include contact between different objects is an important problem for learning-based approaches in robotics. Black-box neural networks can learn to approximately represent discontinuous dynamics, but they typically require large quantities of data and often suffer from pathological behaviour when forecasting for longer time horizons. In this work, we use connections between deep neural networks and differential equations to design a family of deep network architectures for representing contact dynamics between objects. We show that these networks can learn discontinuous contact events in a data-efficient manner from noisy observations in settings that are traditionally difficult for black-box approaches and recent physics inspired neural networks. Our results indicate that an idealised form of touch feedback—which is heavily relied upon by biological systems—is a key component of making this learning problem tractable. Together with the inductive biases introduced through the network architectures, our techniques enable accurate learning of contact dynamics from observations. } }
Endnote
%0 Conference Paper %T Learning Contact Dynamics using Physically Structured Neural Networks %A Andreas Hochlehnert %A Alexander Terenin %A Steindor Saemundsson %A Marc Deisenroth %B Proceedings of The 24th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2021 %E Arindam Banerjee %E Kenji Fukumizu %F pmlr-v130-hochlehnert21a %I PMLR %P 2152--2160 %U https://proceedings.mlr.press/v130/hochlehnert21a.html %V 130 %X Learning physically structured representations of dynamical systems that include contact between different objects is an important problem for learning-based approaches in robotics. Black-box neural networks can learn to approximately represent discontinuous dynamics, but they typically require large quantities of data and often suffer from pathological behaviour when forecasting for longer time horizons. In this work, we use connections between deep neural networks and differential equations to design a family of deep network architectures for representing contact dynamics between objects. We show that these networks can learn discontinuous contact events in a data-efficient manner from noisy observations in settings that are traditionally difficult for black-box approaches and recent physics inspired neural networks. Our results indicate that an idealised form of touch feedback—which is heavily relied upon by biological systems—is a key component of making this learning problem tractable. Together with the inductive biases introduced through the network architectures, our techniques enable accurate learning of contact dynamics from observations.
APA
Hochlehnert, A., Terenin, A., Saemundsson, S. & Deisenroth, M.. (2021). Learning Contact Dynamics using Physically Structured Neural Networks . Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 130:2152-2160 Available from https://proceedings.mlr.press/v130/hochlehnert21a.html.

Related Material