Lifted Filtering via Exchangeable Decomposition

Lüdtke, Stefan; Schröder, Max; Bader, Sebastian; Kersting, Kristian; Kirste, Thomas

doi:10.24963/ijcai.2018/703

Cited by 7 publications

(5 citation statements)

References 1 publication

(1 reference statement)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Graph rewriting systems on multi-hypergraphs are a generalization of multiset rewriting systems, as used in Lifted Marginal Filtering (Lüdtke et al 2018). Specifically, from Figure 4: Graphical model representation of the Bayesian filtering model described by our approach.…”

Section: Bayesian Filtering In Multi-hypergraphsmentioning

confidence: 99%

“…Each LMHG represents a distribution over (ground) MHG. Conceptually, LMHGs are an extension of lifted multiset states, as used in Lifted Marginal Filtering (Lüdtke et al 2018). The illustrated example represents the case where 4 eccentrics are attached to the shelf-top and shelf-bottom boards.…”

Section: Bayesian Filtering In Multi-hypergraphsmentioning

confidence: 99%

“…In contrast to Lüdtke et al (2018), we assume that similar parts never need to be distinguished explicitly, thus we do not require a splitting operator that would handle identification.…”

Section: Bayesian Filtering In Multi-hypergraphsmentioning

confidence: 99%

See 2 more Smart Citations

Activity Recognition in Assembly Tasks by Bayesian Filtering in Multi-Hypergraphs

Felske¹,

Lüdtke²,

Bader³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

We study sensor-based human activity recognition in manual work processes like assembly tasks. In such processes, the system states often have a rich structure, involving object properties and relations. Thus, estimating the hidden system state from sensor observations by recursive Bayesian filtering can be very challenging, due to the combinatorial explosion in the number of system states.To alleviate this problem, we propose an efficient Bayesian filtering model for such processes. In our approach, system states are represented by multi-hypergraphs, and the system dynamics is modeled by graph rewriting rules. We show a preliminary concept that allows to represent distributions over multi-hypergraphs more compactly than by full enumeration, and present an inference algorithm that works directly on this compact representation. We demonstrate the applicability of the algorithm on a real dataset.

show abstract

Section: Bayesian Filtering In Multi-hypergraphsmentioning

confidence: 99%

Section: Bayesian Filtering In Multi-hypergraphsmentioning

confidence: 99%

See 1 more Smart Citation

Activity Recognition in Assembly Tasks by Bayesian Filtering in Multi-Hypergraphs

Felske¹,

Lüdtke²,

Bader³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Furthermore, symbolic and hybrid HAR models [ 34 , 35 , 36 ] can also integrate context data. They model the causal structure of activities, e.g., via precondition-effect rules, and do not only estimate the currently performed activity, but maintain a distribution over system states , which can also include factors such as the locations or states of objects or subjects.…”

Section: Related Workmentioning

confidence: 99%

Context-Aware Human Activity Recognition in Industrial Processes

Niemann

Lüdtke

Bartelt

et al. 2021

Sensors

Self Cite

View full text Add to dashboard Cite

The automatic, sensor-based assessment of human activities is highly relevant for production and logistics, to optimise the economics and ergonomics of these processes. One challenge for accurate activity recognition in these domains is the context-dependence of activities: Similar movements can correspond to different activities, depending on, e.g., the object handled or the location of the subject. In this paper, we propose to explicitly make use of such context information in an activity recognition model. Our first contribution is a publicly available, semantically annotated motion capturing dataset of subjects performing order picking and packaging activities, where context information is recorded explicitly. The second contribution is an activity recognition model that integrates movement data and context information. We empirically show that by using context information, activity recognition performance increases substantially. Additionally, we analyse which of the pieces of context information is most relevant for activity recognition. The insights provided by this paper can help others to design appropriate sensor set-ups in real warehouses for time management.

show abstract

“…Such exchangeability arises whenever a system consists of multiple similar parts. Other examples include link prediction in social networks (Singla and Domingos, 2008) or multi-agent activity recognition (Lüdtke et al, 2018). Unfortunately, treestructured SPNs that are generated by SPN structure learning algorithms (as well as graphical models based on the notion of conditional independence) do not efficiently encode distributions over exchangeable RVs.…”

Section: Introductionmentioning

confidence: 99%

Exchangeability-Aware Sum-Product Networks

Lüdtke¹,

Bartelt²,

Stuckenschmidt³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Sum-Product Networks (SPNs) are expressive probabilistic models that provide exact, tractable inference. They achieve this efficiency by making used of local independence.On the other hand, mixtures of exchangeable variable models (MEVMs) are a class of tractable probabilistic models that make use of exchangeability of random variables to render inference tractable. Exchangeability, which arises naturally in systems consisting of multiple, interrelated entities, has not been considered for efficient representation and inference in SPNs yet.The contribution of this paper is a novel probabilistic model which we call Exchangeability-Aware Sum-Product Networks (XSPNs). It contains both SPNs and MEVMs as special cases, and combines the ability of SPNs to efficiently learn deep probabilistic models with the ability of MEVMs to efficiently handle exchangeable random variables.We also introduce a structure learning algorithm for XSPNs and empirically show that they can be more accurate and efficient than conventional SPNs when the data contains repeated, interchangeable parts.

show abstract

Lifted Filtering via Exchangeable Decomposition

Cited by 7 publications

References 1 publication

Activity Recognition in Assembly Tasks by Bayesian Filtering in Multi-Hypergraphs

Activity Recognition in Assembly Tasks by Bayesian Filtering in Multi-Hypergraphs

Context-Aware Human Activity Recognition in Industrial Processes

Exchangeability-Aware Sum-Product Networks

Contact Info

Product

Resources

About