Exploiting Innocuousness in Bayesian Networks

Alexander Motzek, Ralf Möller


Boolean combination functions in Bayesian networks, such as noisy-or, are often credited a property stating that inactive dependences (e.g., observed to false) do not ``cause any harm'' and an arc becomes vacuous and could have been left out. However, in classic Bayesian networks we are not able to express this property in local CPDs. By using novel ADBNs, we formalize the innocuousness property in CPDs and extend previous work on context-specific independencies. With an explicit representation of innocuousness in local CPDs, we provide a higher causal accuracy for CPD specifications and open new ways for more efficient and less-restricted reasoning in (A)DBNs.
Original languageEnglish
Title of host publicationAI 2015: Advances in Artificial Intelligence
EditorsBernhard Pfahringer, Jochen Renz
Number of pages13
Place of PublicationCham
PublisherSpringer International Publishing
Publication date2015
ISBN (Print)978-3-319-26349-6
ISBN (Electronic)978-3-319-26350-2
Publication statusPublished - 2015
Event28th Australasian Joint Conference on Artificial Intelligence - Canberra, Australia
Duration: 30.11.201504.12.2015
Conference number: 157849

DFG Research Classification Scheme

  • 409-01 Theoretical Computer Science


Dive into the research topics of 'Exploiting Innocuousness in Bayesian Networks'. Together they form a unique fingerprint.

Cite this