ParaGnosis: A Tool for Parallel Knowledge Compilation

Dal, Giso H.; Laarman, Alfons; Lucas, Peter

doi:10.1007/978-3-031-32157-3_2

Cited by 1 publication

(5 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have also tested large versions of our MBD models, in order to investigate the limits of weighted model counting. We are able to perform weighted model counting in networks that are significantly larger than the well-known Munin network using ParaGnosis [39,Ch.7].…”

Section: Discussionmentioning

confidence: 99%

“…As a comparison, the well known Munin network is considered to be large, and has 1041 variables with 98423 probabilities [40,Ch.5]. This demonstrates the inference capabilities of weighted model counting using ParaGnosis [39,Ch.7]. Figure 8.7 show a nearly linear increase in compilation time with respect to the number of probabilities in the networks.…”

Section: Larger Networkmentioning

confidence: 91%

“…The research question that is explored in the remainder of this chapter is whether partitioned probabilistic weighted model counting, using the ParaGnosis toolkit [39,Ch.7 ], can effectively deal with large Bayesian MBD models, in such a way that acceptable diagnostic results are obtained.…”

Section: Methodology 139mentioning

confidence: 99%

“…We have created larger systems using the description in Section 8.4.1, and perform weighted model counting using ParaGnosis [39,Ch.7]. It is clear that the created systems are particularly strenuous on the inference side, due to the joining node in the network.…”

Section: Larger Networkmentioning

confidence: 99%

“…This tool can be found on GitHub 1 and Zenodo 2 to ensure long-term availability and has a DOI 3 . ParaGnosis was awarded the Functional, Reusable and Available badges by TACAS 2023: the 29th International Conference on Tools and Algorithms for the Construction and Analysis of Systems and is published [39,Ch.7].…”

Section: Implementation and Evaluationmentioning

confidence: 99%

See 4 more Smart Citations

Probabilistic inference using partitioned bayesian networks : introducing a compositional framework

Dal

View full text Add to dashboard Cite

Previous chapters have introduced many theoretical concepts that help improve the performance of probabilistic inference. These concepts were implemented in a tool called ParaGnosis, an open-source tool that supports inference queries on discrete Bayesian networks through WMC. This tool can be found on GitHub 1 and Zenodo 2 to ensure long-term availability and has a DOI 3 . ParaGnosis was awarded the Functional, Reusable and Available badges by TACAS 2023: the 29th International Conference on Tools and Algorithms for the Construction and Analysis of Systems. This chapter is based on [39]. My contributions are the writing of the paper and I am the creator of ParaGnosis, where concepts are implemented from [40-43]. Chapter 8: Model-based Probabilistic Diagnosis in Large ModelsModel-based diagnostics or diagnosis is concerned with diagnosing faults or malfunction of physical or cyber-physical systems using a model of the structure and behavior of a real-world system. As cyber-physical systems can be extremely large and complex the associated models can be equally large and complex, imposing a challenge on the computational feasibility of reasoning with such models. In this chapter, we create such a system and evaluate its performance and diagnostic behavior.This chapter is based on a submitted article. My contributions are the writing of the paper and performing the empirical evaluation using ParaGnosis. Chapter 9: Discussion And ConclusionsResearch presented in this thesis is briefly summarized, and brought into context of what was achieved and other research. We also look toward future work and hypothesize on promising research directions.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Larger Networkmentioning

confidence: 91%

Section: Methodology 139mentioning

confidence: 99%

Section: Larger Networkmentioning

confidence: 99%

Section: Implementation and Evaluationmentioning

confidence: 99%

See 3 more Smart Citations