Ontology-guided Semantic Composition for Zero-Shot Learning

Jiaoyan Chen; Freddy Lecue; Yuxia Geng; Jeff Z. Pan; Huajun Chen

Ontology-guided Semantic Composition for Zero-Shot Learning

Jiaoyan Chen, Freddy Lecue, Yuxia Geng, Jeff Z. Pan, Huajun Chen

Computing Science

Research output: Working paper

Abstract

Zero-shot learning (ZSL) is a popular research problem that aims at predicting for those classes that have never appeared in the training stage by utilizing the inter-class relationship with some side information. In this study, we propose to model the compositional and expressive semantics of class labels by an OWL (Web Ontology Language) ontology, and further develop a new ZSL framework with ontology embedding. The effectiveness has been verified by some primary experiments on ani- mal image classification and visual question answering.

Original language	English
Publisher	ArXiv
Number of pages	5
Publication status	Submitted - 30 Jun 2020

Bibliographical note

Accepted by KR 2020 - 17th International Conference on Principles of Knowledge Representation and Reasoning.

Acknowledgments
We want to thank Ian Horrocks from University of Ox-ford for helpful discussions. The work is supported by the AIDA project (Alan Turing Institute), the SIRIUS Centre for Scalable Data Access (Research Council of Norway), Samsung Research UK, Siemens AG, and the EPSRC projects AnaLOG (EP/P025943/1), OASIS (EP/S032347/1) and UKFIRES (EP/S019111/1)

Access to Document

https://arxiv.org/pdf/2006.16917.pdfLicence: Unspecified

Cite this

@techreport{17baf4c2143c44beb5bfd21b082e07da,

title = "Ontology-guided Semantic Composition for Zero-Shot Learning",

abstract = "Zero-shot learning (ZSL) is a popular research problem that aims at predicting for those classes that have never appeared in the training stage by utilizing the inter-class relationship with some side information. In this study, we propose to model the compositional and expressive semantics of class labels by an OWL (Web Ontology Language) ontology, and further develop a new ZSL framework with ontology embedding. The effectiveness has been verified by some primary experiments on ani- mal image classification and visual question answering.",

author = "Jiaoyan Chen and Freddy Lecue and Yuxia Geng and Pan, {Jeff Z.} and Huajun Chen",

note = "Accepted by KR 2020 - 17th International Conference on Principles of Knowledge Representation and Reasoning. Acknowledgments We want to thank Ian Horrocks from University of Ox-ford for helpful discussions. The work is supported by the AIDA project (Alan Turing Institute), the SIRIUS Centre for Scalable Data Access (Research Council of Norway), Samsung Research UK, Siemens AG, and the EPSRC projects AnaLOG (EP/P025943/1), OASIS (EP/S032347/1) and UKFIRES (EP/S019111/1)",

year = "2020",

month = jun,

day = "30",

language = "English",

publisher = "ArXiv",

type = "WorkingPaper",

institution = "ArXiv",

}

TY - UNPB

T1 - Ontology-guided Semantic Composition for Zero-Shot Learning

AU - Chen, Jiaoyan

AU - Lecue, Freddy

AU - Geng, Yuxia

AU - Pan, Jeff Z.

AU - Chen, Huajun

N1 - Accepted by KR 2020 - 17th International Conference on Principles of Knowledge Representation and Reasoning. Acknowledgments We want to thank Ian Horrocks from University of Ox-ford for helpful discussions. The work is supported by the AIDA project (Alan Turing Institute), the SIRIUS Centre for Scalable Data Access (Research Council of Norway), Samsung Research UK, Siemens AG, and the EPSRC projects AnaLOG (EP/P025943/1), OASIS (EP/S032347/1) and UKFIRES (EP/S019111/1)

PY - 2020/6/30

Y1 - 2020/6/30

N2 - Zero-shot learning (ZSL) is a popular research problem that aims at predicting for those classes that have never appeared in the training stage by utilizing the inter-class relationship with some side information. In this study, we propose to model the compositional and expressive semantics of class labels by an OWL (Web Ontology Language) ontology, and further develop a new ZSL framework with ontology embedding. The effectiveness has been verified by some primary experiments on ani- mal image classification and visual question answering.

AB - Zero-shot learning (ZSL) is a popular research problem that aims at predicting for those classes that have never appeared in the training stage by utilizing the inter-class relationship with some side information. In this study, we propose to model the compositional and expressive semantics of class labels by an OWL (Web Ontology Language) ontology, and further develop a new ZSL framework with ontology embedding. The effectiveness has been verified by some primary experiments on ani- mal image classification and visual question answering.

UR - https://arxiv.org/abs/2006.16917

M3 - Working paper

BT - Ontology-guided Semantic Composition for Zero-Shot Learning

PB - ArXiv

ER -

Ontology-guided Semantic Composition for Zero-Shot Learning

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this