The knowledge graph as the default data model for learning on heterogeneous knowledge

Research output: Contribution to journalArticle

Abstract

In modern machine learning, raw data is the pre-ferred input for our models. Where a decade ago data scien-tists were still engineering features, manually picking out the details they thought salient, they now prefer the data in their raw form. As long as we can assume that all relevant and ir-relevant information is present in the input data, we can de-sign deep models that build up intermediate representations to sift out relevant features. However, these models are often domain specific and tailored to the task at hand, and therefore unsuited for learning on heterogeneous knowledge: informa-tion of different types and from different domains. If we can develop methods that operate on this form of knowledge, we can dispense with a great deal of ad-hoc feature engineering and train deep models end-to-end in many more domains. To accomplish this, we first need a data model capable of ex-pressing heterogeneous knowledge naturally in various do-mains, in as usable a form as possible, and satisfying as many use cases as possible. In this position paper, we argue that the knowledge graph is a suitable candidate for this data model. This paper describes current research and discusses some of the promises and challenges of this approach.
LanguageEnglish
Pages39-57
Number of pages19
JournalData Science
Volume1
Issue number1-2
DOIs
StatePublished - 8 Dec 2017

Cite this

@article{d589859477984c67919e8ae0d5586e4c,
title = "The knowledge graph as the default data model for learning on heterogeneous knowledge",
abstract = "In modern machine learning, raw data is the pre-ferred input for our models. Where a decade ago data scien-tists were still engineering features, manually picking out the details they thought salient, they now prefer the data in their raw form. As long as we can assume that all relevant and ir-relevant information is present in the input data, we can de-sign deep models that build up intermediate representations to sift out relevant features. However, these models are often domain specific and tailored to the task at hand, and therefore unsuited for learning on heterogeneous knowledge: informa-tion of different types and from different domains. If we can develop methods that operate on this form of knowledge, we can dispense with a great deal of ad-hoc feature engineering and train deep models end-to-end in many more domains. To accomplish this, we first need a data model capable of ex-pressing heterogeneous knowledge naturally in various do-mains, in as usable a form as possible, and satisfying as many use cases as possible. In this position paper, we argue that the knowledge graph is a suitable candidate for this data model. This paper describes current research and discusses some of the promises and challenges of this approach.",
keywords = "End-to-End Learning, Knowledge Graphs, Machine Learning, Position paper, Semantic Web",
author = "Xander Wilcke and Peter Bloem and {De Boer}, Victor",
year = "2017",
month = "12",
day = "8",
doi = "10.3233/DS-170007",
language = "English",
volume = "1",
pages = "39--57",
journal = "Data Science",
issn = "2451-8484",
publisher = "IOS Press",
number = "1-2",

}

The knowledge graph as the default data model for learning on heterogeneous knowledge. / Wilcke, Xander; Bloem, Peter; De Boer, Victor.

In: Data Science, Vol. 1, No. 1-2, 08.12.2017, p. 39-57.

Research output: Contribution to journalArticle

TY - JOUR

T1 - The knowledge graph as the default data model for learning on heterogeneous knowledge

AU - Wilcke,Xander

AU - Bloem,Peter

AU - De Boer,Victor

PY - 2017/12/8

Y1 - 2017/12/8

N2 - In modern machine learning, raw data is the pre-ferred input for our models. Where a decade ago data scien-tists were still engineering features, manually picking out the details they thought salient, they now prefer the data in their raw form. As long as we can assume that all relevant and ir-relevant information is present in the input data, we can de-sign deep models that build up intermediate representations to sift out relevant features. However, these models are often domain specific and tailored to the task at hand, and therefore unsuited for learning on heterogeneous knowledge: informa-tion of different types and from different domains. If we can develop methods that operate on this form of knowledge, we can dispense with a great deal of ad-hoc feature engineering and train deep models end-to-end in many more domains. To accomplish this, we first need a data model capable of ex-pressing heterogeneous knowledge naturally in various do-mains, in as usable a form as possible, and satisfying as many use cases as possible. In this position paper, we argue that the knowledge graph is a suitable candidate for this data model. This paper describes current research and discusses some of the promises and challenges of this approach.

AB - In modern machine learning, raw data is the pre-ferred input for our models. Where a decade ago data scien-tists were still engineering features, manually picking out the details they thought salient, they now prefer the data in their raw form. As long as we can assume that all relevant and ir-relevant information is present in the input data, we can de-sign deep models that build up intermediate representations to sift out relevant features. However, these models are often domain specific and tailored to the task at hand, and therefore unsuited for learning on heterogeneous knowledge: informa-tion of different types and from different domains. If we can develop methods that operate on this form of knowledge, we can dispense with a great deal of ad-hoc feature engineering and train deep models end-to-end in many more domains. To accomplish this, we first need a data model capable of ex-pressing heterogeneous knowledge naturally in various do-mains, in as usable a form as possible, and satisfying as many use cases as possible. In this position paper, we argue that the knowledge graph is a suitable candidate for this data model. This paper describes current research and discusses some of the promises and challenges of this approach.

KW - End-to-End Learning

KW - Knowledge Graphs

KW - Machine Learning

KW - Position paper

KW - Semantic Web

U2 - 10.3233/DS-170007

DO - 10.3233/DS-170007

M3 - Article

VL - 1

SP - 39

EP - 57

JO - Data Science

T2 - Data Science

JF - Data Science

SN - 2451-8484

IS - 1-2

ER -