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

Manfred Jaeger, Oliver Schulte

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

Abstrakt

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.
OriginalsprogEngelsk
TitelEighth International Workshop on Statistical Relational AI : Accepted Papers
Antal sider8
ForlagStatistical Relational AI (StarAI)
Publikationsdato2018
StatusUdgivet - 2018
BegivenhedInternational Workshop on Statistical Relational AI - Stockholm, Sverige
Varighed: 14 jul. 201814 jul. 2018
Konferencens nummer: 8
http://www.starai.org/2018/#

Konference

KonferenceInternational Workshop on Statistical Relational AI
Nummer8
LandSverige
ByStockholm
Periode14/07/201814/07/2018
Internetadresse

    Fingerprint

Citationsformater

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