Learning from pairwise marginal independencies

Johannes Textor; Alexander Idelberger; Maciej Lis̈kiewicz

Learning from pairwise marginal independencies

Johannes Textor, Alexander Idelberger, Maciej Lis̈kiewicz

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

We consider graphs that represent pairwise marginal independencies amongst a set of variables (for instance, the zero entries of a covari-ance matrix for normal data). We characterize the directed acyclic graphs (DAGs) that faithfully explain a given set of independencies, and derive algorithms to efficiently enumerate such structures. Our results map out the space of faithful causal models for a given set of pairwise marginal independence relations. This allows us to show the extent to which causal inference is possible without using conditional independence tests.

Original language	English
Title of host publication	Uncertainty in Artificial Intelligence
Subtitle of host publication	Proceedings of the Thirty-First Conference (2015), July 12-16, 2015, Amsterdam, Netherlands
Editors	Marina Meila, Tom Heskes
Publisher	AUAI Press
Pages	882-891
Number of pages	10
ISBN (Print)	978-0-9966431-0-8
Publication status	Published - 2015
Event	31st Conference on Uncertainty in Artificial Intelligence, UAI 2015 - Amsterdam, Netherlands Duration: 12 Jul 2015 → 16 Jul 2015

Conference

Conference	31st Conference on Uncertainty in Artificial Intelligence, UAI 2015
Country/Territory	Netherlands
City	Amsterdam
Period	12/07/15 → 16/07/15

Access to Document

proceedingsFinal published version, 505 KB

http://auai.org/uai2015/proceedings.shtml

Cite this

@inproceedings{076650a470bc430c85fb6df576c6121f,

title = "Learning from pairwise marginal independencies",

abstract = "We consider graphs that represent pairwise marginal independencies amongst a set of variables (for instance, the zero entries of a covari-ance matrix for normal data). We characterize the directed acyclic graphs (DAGs) that faithfully explain a given set of independencies, and derive algorithms to efficiently enumerate such structures. Our results map out the space of faithful causal models for a given set of pairwise marginal independence relations. This allows us to show the extent to which causal inference is possible without using conditional independence tests.",

author = "Johannes Textor and Alexander Idelberger and Maciej Li{\"s}kiewicz",

year = "2015",

language = "English",

isbn = "978-0-9966431-0-8",

pages = "882--891",

editor = "Meila, {Marina } and Heskes, {Tom }",

booktitle = "Uncertainty in Artificial Intelligence",

publisher = "AUAI Press",

note = "31st Conference on Uncertainty in Artificial Intelligence, UAI 2015 ; Conference date: 12-07-2015 Through 16-07-2015",

}

Textor, J, Idelberger, A & Lis̈kiewicz, M 2015, Learning from pairwise marginal independencies. in M Meila & T Heskes (eds), Uncertainty in Artificial Intelligence : Proceedings of the Thirty-First Conference (2015), July 12-16, 2015, Amsterdam, Netherlands. AUAI Press, pp. 882-891, 31st Conference on Uncertainty in Artificial Intelligence, UAI 2015, Amsterdam, Netherlands, 12/07/15. <http://auai.org/uai2015/proceedings.shtml>

Learning from pairwise marginal independencies. / Textor, Johannes; Idelberger, Alexander; Lis̈kiewicz, Maciej.
Uncertainty in Artificial Intelligence : Proceedings of the Thirty-First Conference (2015), July 12-16, 2015, Amsterdam, Netherlands. ed. / Marina Meila; Tom Heskes. AUAI Press, 2015. p. 882-891.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Learning from pairwise marginal independencies

AU - Textor, Johannes

AU - Idelberger, Alexander

AU - Lis̈kiewicz, Maciej

PY - 2015

Y1 - 2015

N2 - We consider graphs that represent pairwise marginal independencies amongst a set of variables (for instance, the zero entries of a covari-ance matrix for normal data). We characterize the directed acyclic graphs (DAGs) that faithfully explain a given set of independencies, and derive algorithms to efficiently enumerate such structures. Our results map out the space of faithful causal models for a given set of pairwise marginal independence relations. This allows us to show the extent to which causal inference is possible without using conditional independence tests.

AB - We consider graphs that represent pairwise marginal independencies amongst a set of variables (for instance, the zero entries of a covari-ance matrix for normal data). We characterize the directed acyclic graphs (DAGs) that faithfully explain a given set of independencies, and derive algorithms to efficiently enumerate such structures. Our results map out the space of faithful causal models for a given set of pairwise marginal independence relations. This allows us to show the extent to which causal inference is possible without using conditional independence tests.

UR - http://www.scopus.com/inward/record.url?scp=84943563772&partnerID=8YFLogxK

M3 - Conference contribution

SN - 978-0-9966431-0-8

SP - 882

EP - 891

BT - Uncertainty in Artificial Intelligence

A2 - Meila, Marina

A2 - Heskes, Tom

PB - AUAI Press

T2 - 31st Conference on Uncertainty in Artificial Intelligence, UAI 2015

Y2 - 12 July 2015 through 16 July 2015

ER -

Learning from pairwise marginal independencies

Abstract

Conference

Access to Document

Fingerprint

Cite this