Bringing Back Semantics to Knowledge Graph Embeddings: An Interpretability Approach

Antoine Domingues; Nitisha Jain; Albert Merono Penuela; Elena Simperl

doi:10.1007/978-3-031-71170-1_17

Bringing Back Semantics to Knowledge Graph Embeddings: An Interpretability Approach

Antoine Domingues^*, Nitisha Jain^*, Albert Merono Penuela, Elena Simperl

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

256 Downloads (Pure)

Abstract

Knowledge Graph Embeddings Models project entities and relations from Knowledge Graphs into a vector space. Despite their widespread application, concerns persist about the ability of these models to capture entity similarity effectively. To address this, we introduce InterpretE, a novel neuro-symbolic approach to derive interpretable vector spaces with human-understandable dimensions in terms of the features of the entities. We demonstrate the efficacy of InterpretE in encapsulating desired semantic features, presenting evaluations both in the vector space as well as in terms of semantic similarity measurements.

Original language	English
Title of host publication	Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings
Editors	Tarek R. Besold, Artur d’Avila Garcez, Ernesto Jimenez-Ruiz, Pranava Madhyastha, Benedikt Wagner, Roberto Confalonieri
Publisher	Springer, Cham
Pages	192-203
Number of pages	12
Volume	14979
ISBN (Electronic)	978-3-031-71167-1
ISBN (Print)	9783031711695
DOIs	https://doi.org/10.1007/978-3-031-71170-1_17 https://doi.org/10.1007/978-3-031-71170-1_17
Publication status	Published - 10 Sept 2024
Event	18th International Conference on Neural-Symbolic Learning and Reasoning, NeSy 2024 - Barcelona, Spain Duration: 9 Sept 2024 → 12 Sept 2024

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14980 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	18th International Conference on Neural-Symbolic Learning and Reasoning, NeSy 2024
Country/Territory	Spain
City	Barcelona
Period	9/09/2024 → 12/09/2024

Keywords

interpretable vectors
knowledge graph embeddings
semantic similarity

Access to Document

NeSy_submissionAccepted author manuscript, 1.83 MB
637244_1_En_17_Chapter_OnlinePDFProof, 1.44 MB

https://link.springer.com/10.1007/978-3-031-71170-1_17

Cite this

Domingues, A., Jain, N., Merono Penuela, A., & Simperl, E. (2024). Bringing Back Semantics to Knowledge Graph Embeddings: An Interpretability Approach. In T. R. Besold, A. d’Avila Garcez, E. Jimenez-Ruiz, P. Madhyastha, B. Wagner, & R. Confalonieri (Eds.), Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings (Vol. 14979, pp. 192-203). Article Chapter 17 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14980 LNAI). Springer, Cham. https://doi.org/10.1007/978-3-031-71170-1_17, https://doi.org/10.1007/978-3-031-71170-1_17

Domingues, Antoine ; Jain, Nitisha ; Merono Penuela, Albert et al. / Bringing Back Semantics to Knowledge Graph Embeddings : An Interpretability Approach. Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings. editor / Tarek R. Besold ; Artur d’Avila Garcez ; Ernesto Jimenez-Ruiz ; Pranava Madhyastha ; Benedikt Wagner ; Roberto Confalonieri. Vol. 14979 Springer, Cham, 2024. pp. 192-203 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inbook{755db6c66d7543b6b986e4dcb2c875d1,

title = "Bringing Back Semantics to Knowledge Graph Embeddings: An Interpretability Approach",

abstract = "Knowledge Graph Embeddings Models project entities and relations from Knowledge Graphs into a vector space. Despite their widespread application, concerns persist about the ability of these models to capture entity similarity effectively. To address this, we introduce InterpretE, a novel neuro-symbolic approach to derive interpretable vector spaces with human-understandable dimensions in terms of the features of the entities. We demonstrate the efficacy of InterpretE in encapsulating desired semantic features, presenting evaluations both in the vector space as well as in terms of semantic similarity measurements.",

keywords = "interpretable vectors, knowledge graph embeddings, semantic similarity",

author = "Antoine Domingues and Nitisha Jain and {Merono Penuela}, Albert and Elena Simperl",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.; 18th International Conference on Neural-Symbolic Learning and Reasoning, NeSy 2024 ; Conference date: 09-09-2024 Through 12-09-2024",

year = "2024",

month = sep,

day = "10",

doi = "10.1007/978-3-031-71170-1_17",

language = "English",

isbn = "9783031711695",

volume = "14979",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer, Cham",

pages = "192--203",

editor = "Besold, {Tarek R.} and {d{\textquoteright}Avila Garcez}, Artur and Ernesto Jimenez-Ruiz and Pranava Madhyastha and Benedikt Wagner and Roberto Confalonieri",

booktitle = "Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings",

}

Domingues, A, Jain, N , Merono Penuela, A & Simperl, E 2024, Bringing Back Semantics to Knowledge Graph Embeddings: An Interpretability Approach. in TR Besold, A d’Avila Garcez, E Jimenez-Ruiz, P Madhyastha, B Wagner & R Confalonieri (eds), Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings. vol. 14979, Chapter 17, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14980 LNAI, Springer, Cham, pp. 192-203, 18th International Conference on Neural-Symbolic Learning and Reasoning, NeSy 2024, Barcelona, Spain, 9/09/2024. https://doi.org/10.1007/978-3-031-71170-1_17, https://doi.org/10.1007/978-3-031-71170-1_17

Bringing Back Semantics to Knowledge Graph Embeddings: An Interpretability Approach. / Domingues, Antoine; Jain, Nitisha ; Merono Penuela, Albert et al.
Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings. ed. / Tarek R. Besold; Artur d’Avila Garcez; Ernesto Jimenez-Ruiz; Pranava Madhyastha; Benedikt Wagner; Roberto Confalonieri. Vol. 14979 Springer, Cham, 2024. p. 192-203 Chapter 17 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14980 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

TY - CHAP

T1 - Bringing Back Semantics to Knowledge Graph Embeddings

T2 - 18th International Conference on Neural-Symbolic Learning and Reasoning, NeSy 2024

AU - Domingues, Antoine

AU - Jain, Nitisha

AU - Merono Penuela, Albert

AU - Simperl, Elena

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

PY - 2024/9/10

Y1 - 2024/9/10

N2 - Knowledge Graph Embeddings Models project entities and relations from Knowledge Graphs into a vector space. Despite their widespread application, concerns persist about the ability of these models to capture entity similarity effectively. To address this, we introduce InterpretE, a novel neuro-symbolic approach to derive interpretable vector spaces with human-understandable dimensions in terms of the features of the entities. We demonstrate the efficacy of InterpretE in encapsulating desired semantic features, presenting evaluations both in the vector space as well as in terms of semantic similarity measurements.

AB - Knowledge Graph Embeddings Models project entities and relations from Knowledge Graphs into a vector space. Despite their widespread application, concerns persist about the ability of these models to capture entity similarity effectively. To address this, we introduce InterpretE, a novel neuro-symbolic approach to derive interpretable vector spaces with human-understandable dimensions in terms of the features of the entities. We demonstrate the efficacy of InterpretE in encapsulating desired semantic features, presenting evaluations both in the vector space as well as in terms of semantic similarity measurements.

KW - interpretable vectors

KW - knowledge graph embeddings

KW - semantic similarity

UR - http://www.scopus.com/inward/record.url?scp=85204904471&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-71170-1_17

DO - 10.1007/978-3-031-71170-1_17

M3 - Conference paper

SN - 9783031711695

VL - 14979

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 192

EP - 203

BT - Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings

A2 - Besold, Tarek R.

A2 - d’Avila Garcez, Artur

A2 - Jimenez-Ruiz, Ernesto

A2 - Madhyastha, Pranava

A2 - Wagner, Benedikt

A2 - Confalonieri, Roberto

PB - Springer, Cham

Y2 - 9 September 2024 through 12 September 2024

ER -

Domingues A, Jain N , Merono Penuela A , Simperl E. Bringing Back Semantics to Knowledge Graph Embeddings: An Interpretability Approach. In Besold TR, d’Avila Garcez A, Jimenez-Ruiz E, Madhyastha P, Wagner B, Confalonieri R, editors, Neural-Symbolic Learning and Reasoning - 18th International Conference, NeSy 2024, Proceedings. Vol. 14979. Springer, Cham. 2024. p. 192-203. Chapter 17. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-71170-1_17, 10.1007/978-3-031-71170-1_17