Scalable Identification of Partially Observed Systems with Certainty-Equivalent EM

Kunal Menda, Jean De Becdelievre, Jayesh Gupta, Ilan Kroo, Mykel Kochenderfer, Zachary Manchester
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:6830-6840, 2020.

Abstract

System identification is a key step for model-based control, estimator design, and output prediction. This work considers the offline identification of partially observed nonlinear systems. We empirically show that the certainty-equivalent approximation to expectation-maximization can be a reliable and scalable approach for high-dimensional deterministic systems, which are common in robotics. We formulate certainty-equivalent expectation-maximization as block coordinate-ascent, and provide an efficient implementation. The algorithm is tested on a simulated system of coupled Lorenz attractors, demonstrating its ability to identify high-dimensional systems that can be intractable for particle-based approaches. Our approach is also used to identify the dynamics of an aerobatic helicopter. By augmenting the state with unobserved fluid states, a model is learned that predicts the acceleration of the helicopter better than state-of-the-art approaches. The codebase for this work is available at https://github.com/sisl/CEEM.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-menda20a, title = {Scalable Identification of Partially Observed Systems with Certainty-Equivalent {EM}}, author = {Menda, Kunal and De Becdelievre, Jean and Gupta, Jayesh and Kroo, Ilan and Kochenderfer, Mykel and Manchester, Zachary}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {6830--6840}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/menda20a/menda20a.pdf}, url = {https://proceedings.mlr.press/v119/menda20a.html}, abstract = {System identification is a key step for model-based control, estimator design, and output prediction. This work considers the offline identification of partially observed nonlinear systems. We empirically show that the certainty-equivalent approximation to expectation-maximization can be a reliable and scalable approach for high-dimensional deterministic systems, which are common in robotics. We formulate certainty-equivalent expectation-maximization as block coordinate-ascent, and provide an efficient implementation. The algorithm is tested on a simulated system of coupled Lorenz attractors, demonstrating its ability to identify high-dimensional systems that can be intractable for particle-based approaches. Our approach is also used to identify the dynamics of an aerobatic helicopter. By augmenting the state with unobserved fluid states, a model is learned that predicts the acceleration of the helicopter better than state-of-the-art approaches. The codebase for this work is available at https://github.com/sisl/CEEM.} }
Endnote
%0 Conference Paper %T Scalable Identification of Partially Observed Systems with Certainty-Equivalent EM %A Kunal Menda %A Jean De Becdelievre %A Jayesh Gupta %A Ilan Kroo %A Mykel Kochenderfer %A Zachary Manchester %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-menda20a %I PMLR %P 6830--6840 %U https://proceedings.mlr.press/v119/menda20a.html %V 119 %X System identification is a key step for model-based control, estimator design, and output prediction. This work considers the offline identification of partially observed nonlinear systems. We empirically show that the certainty-equivalent approximation to expectation-maximization can be a reliable and scalable approach for high-dimensional deterministic systems, which are common in robotics. We formulate certainty-equivalent expectation-maximization as block coordinate-ascent, and provide an efficient implementation. The algorithm is tested on a simulated system of coupled Lorenz attractors, demonstrating its ability to identify high-dimensional systems that can be intractable for particle-based approaches. Our approach is also used to identify the dynamics of an aerobatic helicopter. By augmenting the state with unobserved fluid states, a model is learned that predicts the acceleration of the helicopter better than state-of-the-art approaches. The codebase for this work is available at https://github.com/sisl/CEEM.
APA
Menda, K., De Becdelievre, J., Gupta, J., Kroo, I., Kochenderfer, M. & Manchester, Z.. (2020). Scalable Identification of Partially Observed Systems with Certainty-Equivalent EM. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:6830-6840 Available from https://proceedings.mlr.press/v119/menda20a.html.

Related Material