Computer Science > Machine Learning
[Submitted on 10 Dec 2023 (v1), revised 8 Jan 2024 (this version, v3), latest version 22 Jul 2024 (v5)]
Title:Ensemble Kalman Filtering Meets Gaussian Process SSM for Non-Mean-Field and Online Inference
View PDF HTML (experimental)Abstract:Gaussian process state-space models (GPSSMs) are a versatile and principled family of nonlinear dynamical system models. However, existing variational learning and inference methods for GPSSMs often necessitate optimizing a substantial number of variational parameters, leading to inadequate performance and efficiency. To overcome this issue, we propose incorporating the ensemble Kalman filter (EnKF), a well-established model-based filtering technique, into the variational inference framework to approximate the posterior distribution of latent states. This utilization of EnKF can effectively exploit the dependencies between latent states and GP dynamics, while eliminating the need for parameterizing the variational distribution, thereby significantly reducing the number of variational parameters. Moreover, we show that our proposed algorithm allows straightforward evaluation of an approximated evidence lower bound (ELBO) in variational inference via simply summating multiple terms with readily available closed-form solutions. Leveraging automatic differentiation tools, we hence can maximize the ELBO and train the GPSSM efficiently. We also extend the proposed algorithm to accommodate an online setting and provide detailed algorithmic analyses and insights. Extensive evaluation on diverse real and synthetic datasets demonstrates the superiority of our EnKF-aided variational inference algorithms in terms of learning and inference performance compared to existing methods.
Submission history
From: Zhidi Lin [view email][v1] Sun, 10 Dec 2023 15:22:30 UTC (1,360 KB)
[v2] Sat, 23 Dec 2023 06:57:38 UTC (1,306 KB)
[v3] Mon, 8 Jan 2024 10:51:55 UTC (1,307 KB)
[v4] Sun, 14 Jan 2024 08:37:23 UTC (1,305 KB)
[v5] Mon, 22 Jul 2024 07:45:40 UTC (1,405 KB)
Current browse context:
cs.LG
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.