Abstract

Conditional independence (CI) tests play a central role in statistical inference, machine learning, and causal discovery. Most existing CI tests assume that the samples are independently and identically distributed (i.i.d.). However, this assumption often does not hold in the case of relational data. We define Relational Conditional Independence (RCI), a generalization of CI to the relational setting. We show how, under a set of structural assumptions, we can test for RCI by reducing the task of testing for RCI on non-i.i.d. data to the problem of testing for CI on several data sets each of which consists of i.i.d. samples. We develop Kernel Relational CI test (KRCIT), a nonparametric test as a practical approach to testing for RCI by relaxing the structural assumptions used in our analysis of RCI. We describe results of experiments with synthetic relational data that show the benefits of KRCIT relative to traditional CI tests that don't account for the non-i.i.d. nature of relational data.

Original languageEnglish (US)
StatePublished - Jan 1 2017
Event33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017 - Sydney, Australia
Duration: Aug 11 2017Aug 15 2017

Other

Other33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017
CountryAustralia
CitySydney
Period8/11/178/15/17

Fingerprint

Testing
Learning systems
Experiments

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence

Cite this

Lee, S., & Honavar, V. G. (2017). A kernel conditional independence test for relational data. Paper presented at 33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017, Sydney, Australia.
Lee, Sanghack ; Honavar, Vasant Gajanan. / A kernel conditional independence test for relational data. Paper presented at 33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017, Sydney, Australia.
@conference{95ae8ee084c64faf95ee525eb5c069c4,
title = "A kernel conditional independence test for relational data",
abstract = "Conditional independence (CI) tests play a central role in statistical inference, machine learning, and causal discovery. Most existing CI tests assume that the samples are independently and identically distributed (i.i.d.). However, this assumption often does not hold in the case of relational data. We define Relational Conditional Independence (RCI), a generalization of CI to the relational setting. We show how, under a set of structural assumptions, we can test for RCI by reducing the task of testing for RCI on non-i.i.d. data to the problem of testing for CI on several data sets each of which consists of i.i.d. samples. We develop Kernel Relational CI test (KRCIT), a nonparametric test as a practical approach to testing for RCI by relaxing the structural assumptions used in our analysis of RCI. We describe results of experiments with synthetic relational data that show the benefits of KRCIT relative to traditional CI tests that don't account for the non-i.i.d. nature of relational data.",
author = "Sanghack Lee and Honavar, {Vasant Gajanan}",
year = "2017",
month = "1",
day = "1",
language = "English (US)",
note = "33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017 ; Conference date: 11-08-2017 Through 15-08-2017",

}

Lee, S & Honavar, VG 2017, 'A kernel conditional independence test for relational data' Paper presented at 33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017, Sydney, Australia, 8/11/17 - 8/15/17, .

A kernel conditional independence test for relational data. / Lee, Sanghack; Honavar, Vasant Gajanan.

2017. Paper presented at 33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017, Sydney, Australia.

Research output: Contribution to conferencePaper

TY - CONF

T1 - A kernel conditional independence test for relational data

AU - Lee, Sanghack

AU - Honavar, Vasant Gajanan

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Conditional independence (CI) tests play a central role in statistical inference, machine learning, and causal discovery. Most existing CI tests assume that the samples are independently and identically distributed (i.i.d.). However, this assumption often does not hold in the case of relational data. We define Relational Conditional Independence (RCI), a generalization of CI to the relational setting. We show how, under a set of structural assumptions, we can test for RCI by reducing the task of testing for RCI on non-i.i.d. data to the problem of testing for CI on several data sets each of which consists of i.i.d. samples. We develop Kernel Relational CI test (KRCIT), a nonparametric test as a practical approach to testing for RCI by relaxing the structural assumptions used in our analysis of RCI. We describe results of experiments with synthetic relational data that show the benefits of KRCIT relative to traditional CI tests that don't account for the non-i.i.d. nature of relational data.

AB - Conditional independence (CI) tests play a central role in statistical inference, machine learning, and causal discovery. Most existing CI tests assume that the samples are independently and identically distributed (i.i.d.). However, this assumption often does not hold in the case of relational data. We define Relational Conditional Independence (RCI), a generalization of CI to the relational setting. We show how, under a set of structural assumptions, we can test for RCI by reducing the task of testing for RCI on non-i.i.d. data to the problem of testing for CI on several data sets each of which consists of i.i.d. samples. We develop Kernel Relational CI test (KRCIT), a nonparametric test as a practical approach to testing for RCI by relaxing the structural assumptions used in our analysis of RCI. We describe results of experiments with synthetic relational data that show the benefits of KRCIT relative to traditional CI tests that don't account for the non-i.i.d. nature of relational data.

UR - http://www.scopus.com/inward/record.url?scp=85031095069&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85031095069&partnerID=8YFLogxK

M3 - Paper

ER -

Lee S, Honavar VG. A kernel conditional independence test for relational data. 2017. Paper presented at 33rd Conference on Uncertainty in Artificial Intelligence, UAI 2017, Sydney, Australia.