Learning Centre

Infinite-Horizon Gaussian Processes

 |  Login

Show simple item record

dc.contributor Aalto-yliopisto fi
dc.contributor Aalto University en
dc.contributor.author Solin, Arno
dc.contributor.author Hensman, James
dc.contributor.author Turner, Richard E.
dc.date.accessioned 2019-01-14T09:24:22Z
dc.date.available 2019-01-14T09:24:22Z
dc.date.issued 2018
dc.identifier.citation Solin , A , Hensman , J & Turner , R E 2018 , Infinite-Horizon Gaussian Processes . in 32nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montréal, Canada. . Advances in Neural Information Processing Systems , vol. 31 , Curran Associates, Inc. , pp. 3490-3499 , Conference on Neural Information Processing Systems , Montréal , Canada , 02/12/2018 . < http://papers.nips.cc/paper/7608-infinite-horizon-gaussian-processes.pdf > en
dc.identifier.issn 1049-5258
dc.identifier.other PURE UUID: cc4edcc6-ad67-493f-b735-d0fa1cb4d3ed
dc.identifier.other PURE ITEMURL: https://research.aalto.fi/en/publications/cc4edcc6-ad67-493f-b735-d0fa1cb4d3ed
dc.identifier.other PURE LINK: http://papers.nips.cc/paper/7608-infinite-horizon-gaussian-processes.pdf
dc.identifier.other PURE LINK: https://papers.nips.cc/paper/7608-infinite-horizon-gaussian-processes
dc.identifier.other PURE FILEURL: https://research.aalto.fi/files/30995961/SCI_Solin_Infinite_Horizon_Gaussian_Processes.article.pdf
dc.identifier.uri https://aaltodoc.aalto.fi/handle/123456789/36021
dc.description.abstract Gaussian processes provide a flexible framework for forecasting, removing noise, and interpreting long temporal datasets. State space modelling (Kalman filtering) enables these non-parametric models to be deployed on long datasets by reducing the complexity to linear in the number of data points. The complexity is still cubic in the state dimension m which is an impediment to practical application. In certain special cases (Gaussian likelihood, regular spacing) the GP posterior will reach a steady posterior state when the data are very long. We leverage this and formulate an inference scheme for GPs with general likelihoods, where inference is based on single-sweep EP (assumed density filtering). The infinite-horizon model tackles the cubic cost in the state dimensionality and reduces the cost in the state dimension m to O(m^2) per data point. The model is extended to online-learning of hyperparameters. We show examples for large finite-length modelling problems, and present how the method runs in real-time on a smartphone on a continuous data stream updated at 100 Hz. en
dc.format.extent 10
dc.format.extent 3490-3499
dc.format.mimetype application/pdf
dc.language.iso en en
dc.publisher IEEE
dc.relation.ispartof Conference on Neural Information Processing Systems en
dc.relation.ispartofseries 32nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montréal, Canada. en
dc.relation.ispartofseries Advances in Neural Information Processing Systems en
dc.relation.ispartofseries Volume 31 en
dc.rights openAccess en
dc.title Infinite-Horizon Gaussian Processes en
dc.type A4 Artikkeli konferenssijulkaisussa fi
dc.description.version Peer reviewed en
dc.contributor.department Professorship Solin A.
dc.contributor.department PROWLER.io Limited
dc.contributor.department University of Cambridge
dc.contributor.department Department of Computer Science en
dc.identifier.urn URN:NBN:fi:aalto-201901141204
dc.type.version acceptedVersion

Files in this item

Files Size Format View

There are no open access files associated with this item.

This item appears in the following Collection(s)

Show simple item record

Search archive

Advanced Search

article-iconSubmit a publication