The three faces of faithfulness

Jiji ZHANG, Peter SPIRTES

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

4 Citations (Scopus)

Abstract

In the causal inference framework of Spirtes, Glymour, and Scheines (SGS), inferences about causal relationships are made from samples from probability distributions and a number of assumptions relating causal relations to probability distributions. The most controversial of these assumptions is the Causal Faithfulness Assumption, which roughly states that if a conditional independence statement is true of a probability distribution generated by a causal structure, it is entailed by the causal structure and not just for particular parameter values. In this paper we show that the addition of the Causal Faithfulness Assumption plays three quite different roles in the SGS framework: (i) it reduces the degree of underdetermination of causal structure by probability distribution; (ii) computationally, it justifies reliable (constraint-based) causal inference algorithms that would otherwise have to be slower in order to be reliable; and (iii) statistically, it implies that those algorithms reliably obtain the correct answer at smaller sample sizes than would otherwise be the case. We also consider a number of variations on the Causal Faithfulness Assumption, and show how they affect each of these three roles.
Original languageEnglish
Pages (from-to)1011-1027
Number of pages17
JournalSynthese
Volume193
Issue number4
Early online date11 Feb 2015
DOIs
Publication statusPublished - Apr 2016

Fingerprint

Causal
Faithfulness
Causal Inference
Inference
Underdetermination
Small Sample
Causal Relation
Sample Size

Keywords

  • Bayes nets
  • Causal inference
  • Faithfulness
  • Graphical models

Cite this

ZHANG, Jiji ; SPIRTES, Peter. / The three faces of faithfulness. In: Synthese. 2016 ; Vol. 193, No. 4. pp. 1011-1027.
@article{a58210284f254140bef9e3bad475b7c7,
title = "The three faces of faithfulness",
abstract = "In the causal inference framework of Spirtes, Glymour, and Scheines (SGS), inferences about causal relationships are made from samples from probability distributions and a number of assumptions relating causal relations to probability distributions. The most controversial of these assumptions is the Causal Faithfulness Assumption, which roughly states that if a conditional independence statement is true of a probability distribution generated by a causal structure, it is entailed by the causal structure and not just for particular parameter values. In this paper we show that the addition of the Causal Faithfulness Assumption plays three quite different roles in the SGS framework: (i) it reduces the degree of underdetermination of causal structure by probability distribution; (ii) computationally, it justifies reliable (constraint-based) causal inference algorithms that would otherwise have to be slower in order to be reliable; and (iii) statistically, it implies that those algorithms reliably obtain the correct answer at smaller sample sizes than would otherwise be the case. We also consider a number of variations on the Causal Faithfulness Assumption, and show how they affect each of these three roles.",
keywords = "Bayes nets, Causal inference, Faithfulness, Graphical models",
author = "Jiji ZHANG and Peter SPIRTES",
year = "2016",
month = "4",
doi = "10.1007/s11229-015-0673-9",
language = "English",
volume = "193",
pages = "1011--1027",
journal = "Synthese",
issn = "0039-7857",
publisher = "Springer Netherlands",
number = "4",

}

The three faces of faithfulness. / ZHANG, Jiji; SPIRTES, Peter.

In: Synthese, Vol. 193, No. 4, 04.2016, p. 1011-1027.

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

TY - JOUR

T1 - The three faces of faithfulness

AU - ZHANG, Jiji

AU - SPIRTES, Peter

PY - 2016/4

Y1 - 2016/4

N2 - In the causal inference framework of Spirtes, Glymour, and Scheines (SGS), inferences about causal relationships are made from samples from probability distributions and a number of assumptions relating causal relations to probability distributions. The most controversial of these assumptions is the Causal Faithfulness Assumption, which roughly states that if a conditional independence statement is true of a probability distribution generated by a causal structure, it is entailed by the causal structure and not just for particular parameter values. In this paper we show that the addition of the Causal Faithfulness Assumption plays three quite different roles in the SGS framework: (i) it reduces the degree of underdetermination of causal structure by probability distribution; (ii) computationally, it justifies reliable (constraint-based) causal inference algorithms that would otherwise have to be slower in order to be reliable; and (iii) statistically, it implies that those algorithms reliably obtain the correct answer at smaller sample sizes than would otherwise be the case. We also consider a number of variations on the Causal Faithfulness Assumption, and show how they affect each of these three roles.

AB - In the causal inference framework of Spirtes, Glymour, and Scheines (SGS), inferences about causal relationships are made from samples from probability distributions and a number of assumptions relating causal relations to probability distributions. The most controversial of these assumptions is the Causal Faithfulness Assumption, which roughly states that if a conditional independence statement is true of a probability distribution generated by a causal structure, it is entailed by the causal structure and not just for particular parameter values. In this paper we show that the addition of the Causal Faithfulness Assumption plays three quite different roles in the SGS framework: (i) it reduces the degree of underdetermination of causal structure by probability distribution; (ii) computationally, it justifies reliable (constraint-based) causal inference algorithms that would otherwise have to be slower in order to be reliable; and (iii) statistically, it implies that those algorithms reliably obtain the correct answer at smaller sample sizes than would otherwise be the case. We also consider a number of variations on the Causal Faithfulness Assumption, and show how they affect each of these three roles.

KW - Bayes nets

KW - Causal inference

KW - Faithfulness

KW - Graphical models

UR - http://commons.ln.edu.hk/sw_master/4379

UR - https://www.scopus.com/inward/record.uri?eid=2-s2.0-84922615612&doi=10.1007%2fs11229-015-0673-9&partnerID=40&md5=7081160aab486990d549dfb2b7b9870b

U2 - 10.1007/s11229-015-0673-9

DO - 10.1007/s11229-015-0673-9

M3 - Journal Article (refereed)

VL - 193

SP - 1011

EP - 1027

JO - Synthese

JF - Synthese

SN - 0039-7857

IS - 4

ER -