Adapting Neural Models with Sequential Monte Carlo Dropout

Pamela Carreno, Dana Kulic, Michael Burke
Proceedings of The 6th Conference on Robot Learning, PMLR 205:1542-1552, 2023.

Abstract

The ability to adapt to changing environments and settings is essential for robots acting in dynamic and unstructured environments or working alongside humans with varied abilities or preferences. This work introduces an extremely simple and effective approach to adapting neural models in response to changing settings, without requiring any specialised meta-learning strategies. We first train a standard network using dropout, which is analogous to learning an ensemble of predictive models or distribution over predictions. At run-time, we use a particle filter to maintain a distribution over dropout masks to adapt the neural model to changing settings in an online manner. Experimental results show improved performance in control problems requiring both online and look-ahead prediction, and showcase the interpretability of the inferred masks in a human behaviour modelling task for drone tele-operation.

Cite this Paper

BibTeX
@InProceedings{pmlr-v205-carreno23a, title = {Adapting Neural Models with Sequential Monte Carlo Dropout}, author = {Carreno, Pamela and Kulic, Dana and Burke, Michael}, booktitle = {Proceedings of The 6th Conference on Robot Learning}, pages = {1542--1552}, year = {2023}, editor = {Liu, Karen and Kulic, Dana and Ichnowski, Jeff}, volume = {205}, series = {Proceedings of Machine Learning Research}, month = {14--18 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v205/carreno23a/carreno23a.pdf}, url = {https://proceedings.mlr.press/v205/carreno23a.html}, abstract = {The ability to adapt to changing environments and settings is essential for robots acting in dynamic and unstructured environments or working alongside humans with varied abilities or preferences. This work introduces an extremely simple and effective approach to adapting neural models in response to changing settings, without requiring any specialised meta-learning strategies. We first train a standard network using dropout, which is analogous to learning an ensemble of predictive models or distribution over predictions. At run-time, we use a particle filter to maintain a distribution over dropout masks to adapt the neural model to changing settings in an online manner. Experimental results show improved performance in control problems requiring both online and look-ahead prediction, and showcase the interpretability of the inferred masks in a human behaviour modelling task for drone tele-operation. } }
Endnote
%0 Conference Paper %T Adapting Neural Models with Sequential Monte Carlo Dropout %A Pamela Carreno %A Dana Kulic %A Michael Burke %B Proceedings of The 6th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2023 %E Karen Liu %E Dana Kulic %E Jeff Ichnowski %F pmlr-v205-carreno23a %I PMLR %P 1542--1552 %U https://proceedings.mlr.press/v205/carreno23a.html %V 205 %X The ability to adapt to changing environments and settings is essential for robots acting in dynamic and unstructured environments or working alongside humans with varied abilities or preferences. This work introduces an extremely simple and effective approach to adapting neural models in response to changing settings, without requiring any specialised meta-learning strategies. We first train a standard network using dropout, which is analogous to learning an ensemble of predictive models or distribution over predictions. At run-time, we use a particle filter to maintain a distribution over dropout masks to adapt the neural model to changing settings in an online manner. Experimental results show improved performance in control problems requiring both online and look-ahead prediction, and showcase the interpretability of the inferred masks in a human behaviour modelling task for drone tele-operation.
APA
Carreno, P., Kulic, D. & Burke, M.. (2023). Adapting Neural Models with Sequential Monte Carlo Dropout. Proceedings of The 6th Conference on Robot Learning, in Proceedings of Machine Learning Research 205:1542-1552 Available from https://proceedings.mlr.press/v205/carreno23a.html.

Related Material