Flexible and efficient inference with particles for the variational Gaussian approximation
| dc.contributor.author | Galy-Fajou, Théo | |
| dc.contributor.author | Perrone, Valerio | |
| dc.contributor.author | Opper, Manfred | |
| dc.date.accessioned | 2022-01-17T13:55:53Z | |
| dc.date.available | 2022-01-17T13:55:53Z | |
| dc.date.issued | 2021-07-30 | |
| dc.description.abstract | Variational inference is a powerful framework, used to approximate intractable posteriors through variational distributions. The de facto standard is to rely on Gaussian variational families, which come with numerous advantages: they are easy to sample from, simple to parametrize, and many expectations are known in closed-form or readily computed by quadrature. In this paper, we view the Gaussian variational approximation problem through the lens of gradient flows. We introduce a flexible and efficient algorithm based on a linear flow leading to a particle-based approximation. We prove that, with a sufficient number of particles, our algorithm converges linearly to the exact solution for Gaussian targets, and a low-rank approximation otherwise. In addition to the theoretical analysis, we show, on a set of synthetic and real-world high-dimensional problems, that our algorithm outperforms existing methods with Gaussian targets while performing on a par with non-Gaussian targets. | en |
| dc.description.sponsorship | DFG, 414044773, Open Access Publizieren 2021 - 2022 / Technische Universität Berlin | en |
| dc.identifier.eissn | 1099-4300 | |
| dc.identifier.uri | https://depositonce.tu-berlin.de/handle/11303/16139 | |
| dc.identifier.uri | http://dx.doi.org/10.14279/depositonce-14913 | |
| dc.language.iso | en | en |
| dc.relation.ispartof | 10.14279/depositonce-17605 | |
| dc.rights.uri | https://creativecommons.org/licenses/by/4.0/ | en |
| dc.subject.ddc | 510 Mathematik | de |
| dc.subject.other | variational inference | en |
| dc.subject.other | Gaussian | en |
| dc.subject.other | particle flow | en |
| dc.subject.other | variable flow | en |
| dc.title | Flexible and efficient inference with particles for the variational Gaussian approximation | en |
| dc.type | Article | en |
| dc.type.version | publishedVersion | en |
| dcterms.bibliographicCitation.articlenumber | 990 | en |
| dcterms.bibliographicCitation.doi | 10.3390/e23080990 | en |
| dcterms.bibliographicCitation.issue | 8 | en |
| dcterms.bibliographicCitation.journaltitle | Entropy | en |
| dcterms.bibliographicCitation.originalpublishername | MDPI | en |
| dcterms.bibliographicCitation.originalpublisherplace | Basel | en |
| dcterms.bibliographicCitation.volume | 23 | en |
| tub.accessrights.dnb | free | en |
| tub.affiliation | Fak. 4 Elektrotechnik und Informatik::Inst. Softwaretechnik und Theoretische Informatik::FG Methoden der Künstlichen Intelligenz | de |
| tub.affiliation.faculty | Fak. 4 Elektrotechnik und Informatik | de |
| tub.affiliation.group | FG Methoden der Künstlichen Intelligenz | de |
| tub.affiliation.institute | Inst. Softwaretechnik und Theoretische Informatik | de |
| tub.publisher.universityorinstitution | Technische Universität Berlin | en |