Inference, Learning, and Population Size: Projectivity for SRL Models

Manfred Jaeger, Oliver Schulte

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

Abstract

A subtle difference between propositional and relational data is that in many relational models, marginal probabilities depend on the population or domain size. This paper connects the dependence on population size to the classic notion of projectivity from statistical theory: Projectivity implies that relational predictions are robust with respect to changes in domain size. We discuss projectivity for a number of common SRL systems, and identify syntactic fragments that are guaranteed to yield projective models. The syntactic conditions are restrictive, which suggests that projectivity is difficult to achieve in SRL, and care must be taken when working with different domain sizes.
Original languageEnglish
Title of host publicationEighth International Workshop on Statistical Relational AI : Accepted Papers
Number of pages8
PublisherStatistical Relational AI (StarAI)
Publication date2018
Publication statusPublished - 2018
EventInternational Workshop on Statistical Relational AI - Stockholm, Sweden
Duration: 14 Jul 201814 Jul 2018
Conference number: 8
http://www.starai.org/2018/#

Conference

ConferenceInternational Workshop on Statistical Relational AI
Number8
CountrySweden
CityStockholm
Period14/07/201814/07/2018
Internet address

Fingerprint

population size
learning
prediction

Cite this

Jaeger, M., & Schulte, O. (2018). Inference, Learning, and Population Size: Projectivity for SRL Models. In Eighth International Workshop on Statistical Relational AI: Accepted Papers Statistical Relational AI (StarAI).
Jaeger, Manfred ; Schulte, Oliver. / Inference, Learning, and Population Size : Projectivity for SRL Models. Eighth International Workshop on Statistical Relational AI: Accepted Papers. Statistical Relational AI (StarAI), 2018.
@inproceedings{95ec8cc7e16245e983184283817aceb4,
title = "Inference, Learning, and Population Size: Projectivity for SRL Models",
abstract = "A subtle difference between propositional and relational data is that in many relational models, marginal probabilities depend on the population or domain size. This paper connects the dependence on population size to the classic notion of projectivity from statistical theory: Projectivity implies that relational predictions are robust with respect to changes in domain size. We discuss projectivity for a number of common SRL systems, and identify syntactic fragments that are guaranteed to yield projective models. The syntactic conditions are restrictive, which suggests that projectivity is difficult to achieve in SRL, and care must be taken when working with different domain sizes.",
author = "Manfred Jaeger and Oliver Schulte",
year = "2018",
language = "English",
booktitle = "Eighth International Workshop on Statistical Relational AI",
publisher = "Statistical Relational AI (StarAI)",

}

Jaeger, M & Schulte, O 2018, Inference, Learning, and Population Size: Projectivity for SRL Models. in Eighth International Workshop on Statistical Relational AI: Accepted Papers. Statistical Relational AI (StarAI), International Workshop on Statistical Relational AI, Stockholm, Sweden, 14/07/2018.

Inference, Learning, and Population Size : Projectivity for SRL Models. / Jaeger, Manfred; Schulte, Oliver.

Eighth International Workshop on Statistical Relational AI: Accepted Papers. Statistical Relational AI (StarAI), 2018.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Inference, Learning, and Population Size

T2 - Projectivity for SRL Models

AU - Jaeger, Manfred

AU - Schulte, Oliver

PY - 2018

Y1 - 2018

N2 - A subtle difference between propositional and relational data is that in many relational models, marginal probabilities depend on the population or domain size. This paper connects the dependence on population size to the classic notion of projectivity from statistical theory: Projectivity implies that relational predictions are robust with respect to changes in domain size. We discuss projectivity for a number of common SRL systems, and identify syntactic fragments that are guaranteed to yield projective models. The syntactic conditions are restrictive, which suggests that projectivity is difficult to achieve in SRL, and care must be taken when working with different domain sizes.

AB - A subtle difference between propositional and relational data is that in many relational models, marginal probabilities depend on the population or domain size. This paper connects the dependence on population size to the classic notion of projectivity from statistical theory: Projectivity implies that relational predictions are robust with respect to changes in domain size. We discuss projectivity for a number of common SRL systems, and identify syntactic fragments that are guaranteed to yield projective models. The syntactic conditions are restrictive, which suggests that projectivity is difficult to achieve in SRL, and care must be taken when working with different domain sizes.

M3 - Article in proceeding

BT - Eighth International Workshop on Statistical Relational AI

PB - Statistical Relational AI (StarAI)

ER -

Jaeger M, Schulte O. Inference, Learning, and Population Size: Projectivity for SRL Models. In Eighth International Workshop on Statistical Relational AI: Accepted Papers. Statistical Relational AI (StarAI). 2018