Structured Inference Networks for Nonlinear State Space Models. Krishnan, R., G., Shalit, U., & Sontag, D.
Paper abstract bibtex Gaussian state space models have been used for decades as generative models of sequential data. They admit an intuitive probabilistic interpretation, have a simple functional form, and enjoy widespread adoption. We introduce a unified algorithm to efficiently learn a broad class of linear and non-linear state space models, including variants where the emission and tran-sition distributions are modeled by deep neural networks. Our learning algorithm simultaneously learns a compiled inference network and the generative model, leveraging a structured variational approximation parameterized by recurrent neural networks to mimic the posterior distribution. We apply the learning algorithm to both synthetic and real-world datasets, demonstrating its scalability and versatility. We find that using the structured approximation to the posterior results in models with significantly higher held-out likelihood.
@article{
title = {Structured Inference Networks for Nonlinear State Space Models},
type = {article},
id = {10b7ed7d-a5c8-35c5-9125-ea85d6c71a9f},
created = {2018-02-14T12:53:37.892Z},
file_attached = {true},
profile_id = {ea20c45e-0866-3373-a162-5a7398e4ac0b},
group_id = {2d721da3-627a-3ca3-862a-cd2d130d920f},
last_modified = {2018-02-14T12:53:39.759Z},
read = {false},
starred = {false},
authored = {false},
confirmed = {false},
hidden = {false},
folder_uuids = {d1061552-d7f6-480b-a466-74128dc5a51f},
private_publication = {false},
abstract = {Gaussian state space models have been used for decades as generative models of sequential data. They admit an intuitive probabilistic interpretation, have a simple functional form, and enjoy widespread adoption. We introduce a unified algorithm to efficiently learn a broad class of linear and non-linear state space models, including variants where the emission and tran-sition distributions are modeled by deep neural networks. Our learning algorithm simultaneously learns a compiled inference network and the generative model, leveraging a structured variational approximation parameterized by recurrent neural networks to mimic the posterior distribution. We apply the learning algorithm to both synthetic and real-world datasets, demonstrating its scalability and versatility. We find that using the structured approximation to the posterior results in models with significantly higher held-out likelihood.},
bibtype = {article},
author = {Krishnan, Rahul G and Shalit, Uri and Sontag, David}
}