PAC-Bayesian Error Bound, via Rényi Divergence, for a Class of Linear Time-Invariant State-Space Models
Abstract
In this paper we derive a PAC-Bayesian error bound for a class of stochastic dynamical systems with inputs, namely, for linear time-invariant stochastic state-space models (stochastic LTI systems for short). This class of systems is widely used in control engineering and econometrics, in particular, they represent a special case of recurrent neural networks. In this paper we 1) formalize the learning problem for stochastic LTI systems with inputs, 2) derive a PAC-Bayesian error bound for such systems, and 3) discuss various consequences of this error bound.
Cite
Text
Eringis et al. "PAC-Bayesian Error Bound, via Rényi Divergence, for a Class of Linear Time-Invariant State-Space Models." International Conference on Machine Learning, 2024.Markdown
[Eringis et al. "PAC-Bayesian Error Bound, via Rényi Divergence, for a Class of Linear Time-Invariant State-Space Models." International Conference on Machine Learning, 2024.](https://mlanthology.org/icml/2024/eringis2024icml-pacbayesian/)BibTeX
@inproceedings{eringis2024icml-pacbayesian,
title = {{PAC-Bayesian Error Bound, via Rényi Divergence, for a Class of Linear Time-Invariant State-Space Models}},
author = {Eringis, Deividas and Leth, John and Tan, Zheng-Hua and Wisniewski, Rafal and Petreczky, Mihaly},
booktitle = {International Conference on Machine Learning},
year = {2024},
pages = {12560-12587},
volume = {235},
url = {https://mlanthology.org/icml/2024/eringis2024icml-pacbayesian/}
}