Lifted Bayesian filtering in multi-entity systems

Stefan Lüdtke

doi:10.18453/rosdok_id00003022

Stefan Lüdtke

Lifted Bayesian filtering in multi-entity systems

Universität Rostock, 2020

https://doi.org/10.18453/rosdok_id00003022

Abstract: This thesis focuses on Bayesian filtering for systems that consist of multiple, interacting entites (e.g. agents or objects), which can naturally be described by Multiset Rewriting Systems (MRSs). The main insight is that the state space that is underling an MRS exhibits a certain symmetry, which can be exploited to increase inference efficiency. We provide an efficient, lifted filtering algorithm, which is able to achieve a factorial reduction in space and time complexity, compared to conventional, ground filtering.

doctoral thesis free access

title:

Lifted Bayesian filtering in multi-entity systems

contributing persons:

Stefan Lüdtke[VerfasserIn]
	0000-0002-1488-4236
	1121592104
Thomas Kirste[AkademischeR BetreuerIn]
	137244754
Heiner Stuckenschmidt[AkademischeR BetreuerIn]
	0000-0002-0209-3859
	124274242

contributing corporate bodies:

Universität Rostock[Grad-verleihende Institution]
	38329-6
Universität Rostock, Fakultät für Informatik und Elektrotechnik[Grad-verleihende Institution]
	10085032-7

abstract:

This thesis focuses on Bayesian filtering for systems that consist of multiple, interacting entites (e.g. agents or objects), which can naturally be described by Multiset Rewriting Systems (MRSs). The main insight is that the state space that is underling an MRS exhibits a certain symmetry, which can be exploited to increase inference efficiency. We provide an efficient, lifted filtering algorithm, which is able to achieve a factorial reduction in space and time complexity, compared to conventional, ground filtering. [English]

Diese Arbeit betrachtet Bayes'sche Filter in Systemen, die aus mehreren, interagierenden Entitäten (z.B. Agenten oder Objekten) bestehen. Die Systemdynamik solcher Systeme kann auf natürliche Art durch Multiset Rewriting Systems (MRS) spezifiziert werden. Die wesentliche Erkenntnis ist, dass der Zustandraum Symmetrien aufweist, die ausgenutzt werden können, um die Effizienz der Inferenz zu erhöhen. Wir führen einen effizienten, gelifteten Filter-Algorithmus ein, dessen Zeit- und Platzkomplexität gegenüber dem grundierten Algorithmus um einen faktoriellen Faktor reduziert ist. [German]

document type:

doctoral thesis

institution:

Faculty of Computer Science and Electrical Engineering

language:

English

subject class (DDC):

004 Data processing Computer sciences

publication /
production:

Rostock

Rostock: Universität Rostock

2020

statement of responsibility:

vorgelegt von Stefan Lüdtke

identifiers:

K10plus id:	1757652698
URN:	urn:nbn:de:gbv:28-rosdok_id00003022-4
DOI:	10.18453/rosdok_id00003022
PURL:	https://purl.uni-rostock.de/rosdok/id00003022

access condition:

open access

license/rights statement:

CC BY 4.0
This work is licensed under a
Creative Commons Attribution 4.0 International License.

RosDok id:

rosdok_disshab_0000002515

created / modified:

11.05.2021 / 08.08.2023

metadata license:

The metadata of this document was dedicated to the public domain
(CC0 1.0 Universal Public Domain Dedication).

fulltext	Luedtke_Dissertation_2021.pdf (3,23 MB; MD5: bba72e4f8cc015d841edbf1dee7021aa)

Lifted Bayesian filtering in multi-entity systems

Persistent citation

Portals

Rights

Export

Share

Contact

Service

MyCoRe