Prediction and Control with Temporal Segment Models

Nikhil Mishra, Pieter Abbeel, Igor Mordatch
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:2459-2468, 2017.

Abstract

We introduce a method for learning the dynamics of complex nonlinear systems based on deep generative models over temporal segments of states and actions. Unlike dynamics models that operate over individual discrete timesteps, we learn the distribution over future state trajectories conditioned on past state, past action, and planned future action trajectories, as well as a latent prior over action trajectories. Our approach is based on convolutional autoregressive models and variational autoencoders. It makes stable and accurate predictions over long horizons for complex, stochastic systems, effectively expressing uncertainty and modeling the effects of collisions, sensory noise, and action delays. The learned dynamics model and action prior can be used for end-to-end, fully differentiable trajectory optimization and model-based policy optimization, which we use to evaluate the performance and sample-efficiency of our method.

Cite this Paper

BibTeX
@InProceedings{pmlr-v70-mishra17a, title = {Prediction and Control with Temporal Segment Models}, author = {Nikhil Mishra and Pieter Abbeel and Igor Mordatch}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {2459--2468}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/mishra17a/mishra17a.pdf}, url = {https://proceedings.mlr.press/v70/mishra17a.html}, abstract = {We introduce a method for learning the dynamics of complex nonlinear systems based on deep generative models over temporal segments of states and actions. Unlike dynamics models that operate over individual discrete timesteps, we learn the distribution over future state trajectories conditioned on past state, past action, and planned future action trajectories, as well as a latent prior over action trajectories. Our approach is based on convolutional autoregressive models and variational autoencoders. It makes stable and accurate predictions over long horizons for complex, stochastic systems, effectively expressing uncertainty and modeling the effects of collisions, sensory noise, and action delays. The learned dynamics model and action prior can be used for end-to-end, fully differentiable trajectory optimization and model-based policy optimization, which we use to evaluate the performance and sample-efficiency of our method.} }
Endnote
%0 Conference Paper %T Prediction and Control with Temporal Segment Models %A Nikhil Mishra %A Pieter Abbeel %A Igor Mordatch %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-mishra17a %I PMLR %P 2459--2468 %U https://proceedings.mlr.press/v70/mishra17a.html %V 70 %X We introduce a method for learning the dynamics of complex nonlinear systems based on deep generative models over temporal segments of states and actions. Unlike dynamics models that operate over individual discrete timesteps, we learn the distribution over future state trajectories conditioned on past state, past action, and planned future action trajectories, as well as a latent prior over action trajectories. Our approach is based on convolutional autoregressive models and variational autoencoders. It makes stable and accurate predictions over long horizons for complex, stochastic systems, effectively expressing uncertainty and modeling the effects of collisions, sensory noise, and action delays. The learned dynamics model and action prior can be used for end-to-end, fully differentiable trajectory optimization and model-based policy optimization, which we use to evaluate the performance and sample-efficiency of our method.
APA
Mishra, N., Abbeel, P. & Mordatch, I.. (2017). Prediction and Control with Temporal Segment Models. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:2459-2468 Available from https://proceedings.mlr.press/v70/mishra17a.html.

Related Material