A General-Purpose Protocol for Multi-agent Based Explanations

Giovanni Ciatto; Matteo Magnini; Berk Buzcu; Reyhan Aydoğan; Andrea Omicini

doi:10.1007/978-3-031-40878-6_3

A General-Purpose Protocol for Multi-agent Based Explanations

Giovanni Ciatto^*, Matteo Magnini, Berk Buzcu, Reyhan Aydoğan, Andrea Omicini

^*Corresponding author for this work

Interactive Intelligence

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

7 Downloads (Pure)

Abstract

Building on prior works on explanation negotiation protocols, this paper proposes a general-purpose protocol for multi-agent systems where recommender agents may need to provide explanations for their recommendations. The protocol specifies the roles and responsibilities of the explainee and the explainer agent and the types of information that should be exchanged between them to ensure a clear and effective explanation. However, it does not prescribe any particular sort of recommendation or explanation, hence remaining agnostic w.r.t. such notions. Novelty lays in the extended support for both ordinary and contrastive explanations, as well as for the situation where no explanation is needed as none is requested by the explainee. Accordingly, we formally present and analyse the protocol, motivating its design and discussing its generality. We also discuss the reification of the protocol into a re-usable software library, namely PyXMas, which is meant to support developers willing to build explainable MAS leveraging our protocol. Finally, we discuss how custom notions of recommendation and explanation can be easily plugged into PyXMas.

Original language	English
Title of host publication	Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers
Editors	Davide Calvaresi, Amro Najjar, Andrea Omicini, Rachele Carli, Giovanni Ciatto, Reyhan Aydogan, Yazan Mualla, Kary Främling
Publisher	Springer
Pages	38-58
Number of pages	21
ISBN (Print)	9783031408779
DOIs	https://doi.org/10.1007/978-3-031-40878-6_3
Publication status	Published - 2023
Event	Proceedings of the 5th International Workshop on EXTRAAMAS 2023 - London, United Kingdom Duration: 29 May 2023 → 29 May 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14127 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	Proceedings of the 5th International Workshop on EXTRAAMAS 2023
Country/Territory	United Kingdom
City	London
Period	29/05/23 → 29/05/23

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

explanation protocols
multi-agent systems
PyXMas
recommender systems
Spade
XAI

Access to Document

10.1007/978-3-031-40878-6_3

978-3-031-40878-6_3Final published version, 1.3 MB

Cite this

Ciatto, G., Magnini, M., Buzcu, B., Aydoğan, R., & Omicini, A. (2023). A General-Purpose Protocol for Multi-agent Based Explanations. In D. Calvaresi, A. Najjar, A. Omicini, R. Carli, G. Ciatto, R. Aydogan, Y. Mualla, & K. Främling (Eds.), Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers (pp. 38-58). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14127 LNAI). Springer. https://doi.org/10.1007/978-3-031-40878-6_3

Ciatto, Giovanni ; Magnini, Matteo ; Buzcu, Berk et al. / A General-Purpose Protocol for Multi-agent Based Explanations. Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers. editor / Davide Calvaresi ; Amro Najjar ; Andrea Omicini ; Rachele Carli ; Giovanni Ciatto ; Reyhan Aydogan ; Yazan Mualla ; Kary Främling. Springer, 2023. pp. 38-58 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{dbab64875652479d86d8aa6087c39400,

title = "A General-Purpose Protocol for Multi-agent Based Explanations",

abstract = "Building on prior works on explanation negotiation protocols, this paper proposes a general-purpose protocol for multi-agent systems where recommender agents may need to provide explanations for their recommendations. The protocol specifies the roles and responsibilities of the explainee and the explainer agent and the types of information that should be exchanged between them to ensure a clear and effective explanation. However, it does not prescribe any particular sort of recommendation or explanation, hence remaining agnostic w.r.t. such notions. Novelty lays in the extended support for both ordinary and contrastive explanations, as well as for the situation where no explanation is needed as none is requested by the explainee. Accordingly, we formally present and analyse the protocol, motivating its design and discussing its generality. We also discuss the reification of the protocol into a re-usable software library, namely PyXMas, which is meant to support developers willing to build explainable MAS leveraging our protocol. Finally, we discuss how custom notions of recommendation and explanation can be easily plugged into PyXMas.",

keywords = "explanation protocols, multi-agent systems, PyXMas, recommender systems, Spade, XAI",

author = "Giovanni Ciatto and Matteo Magnini and Berk Buzcu and Reyhan Aydoğan and Andrea Omicini",

note = "Green Open Access added to TU Delft Institutional Repository {\textquoteleft}You share, we take care!{\textquoteright} – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public. ; Proceedings of the 5th International Workshop on EXTRAAMAS 2023 ; Conference date: 29-05-2023 Through 29-05-2023",

year = "2023",

doi = "10.1007/978-3-031-40878-6_3",

language = "English",

isbn = "9783031408779",

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

publisher = "Springer",

pages = "38--58",

editor = "Davide Calvaresi and Amro Najjar and Andrea Omicini and Rachele Carli and Giovanni Ciatto and Reyhan Aydogan and Yazan Mualla and Kary Fr{\"a}mling",

booktitle = "Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers",

}

Ciatto, G, Magnini, M, Buzcu, B, Aydoğan, R & Omicini, A 2023, A General-Purpose Protocol for Multi-agent Based Explanations. in D Calvaresi, A Najjar, A Omicini, R Carli, G Ciatto, R Aydogan, Y Mualla & K Främling (eds), Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14127 LNAI, Springer, pp. 38-58, Proceedings of the 5th International Workshop on EXTRAAMAS 2023, London, United Kingdom, 29/05/23. https://doi.org/10.1007/978-3-031-40878-6_3

A General-Purpose Protocol for Multi-agent Based Explanations. / Ciatto, Giovanni; Magnini, Matteo; Buzcu, Berk et al.
Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers. ed. / Davide Calvaresi; Amro Najjar; Andrea Omicini; Rachele Carli; Giovanni Ciatto; Reyhan Aydogan; Yazan Mualla; Kary Främling. Springer, 2023. p. 38-58 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14127 LNAI).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - A General-Purpose Protocol for Multi-agent Based Explanations

AU - Ciatto, Giovanni

AU - Magnini, Matteo

AU - Buzcu, Berk

AU - Aydoğan, Reyhan

AU - Omicini, Andrea

N1 - Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2023

Y1 - 2023

N2 - Building on prior works on explanation negotiation protocols, this paper proposes a general-purpose protocol for multi-agent systems where recommender agents may need to provide explanations for their recommendations. The protocol specifies the roles and responsibilities of the explainee and the explainer agent and the types of information that should be exchanged between them to ensure a clear and effective explanation. However, it does not prescribe any particular sort of recommendation or explanation, hence remaining agnostic w.r.t. such notions. Novelty lays in the extended support for both ordinary and contrastive explanations, as well as for the situation where no explanation is needed as none is requested by the explainee. Accordingly, we formally present and analyse the protocol, motivating its design and discussing its generality. We also discuss the reification of the protocol into a re-usable software library, namely PyXMas, which is meant to support developers willing to build explainable MAS leveraging our protocol. Finally, we discuss how custom notions of recommendation and explanation can be easily plugged into PyXMas.

AB - Building on prior works on explanation negotiation protocols, this paper proposes a general-purpose protocol for multi-agent systems where recommender agents may need to provide explanations for their recommendations. The protocol specifies the roles and responsibilities of the explainee and the explainer agent and the types of information that should be exchanged between them to ensure a clear and effective explanation. However, it does not prescribe any particular sort of recommendation or explanation, hence remaining agnostic w.r.t. such notions. Novelty lays in the extended support for both ordinary and contrastive explanations, as well as for the situation where no explanation is needed as none is requested by the explainee. Accordingly, we formally present and analyse the protocol, motivating its design and discussing its generality. We also discuss the reification of the protocol into a re-usable software library, namely PyXMas, which is meant to support developers willing to build explainable MAS leveraging our protocol. Finally, we discuss how custom notions of recommendation and explanation can be easily plugged into PyXMas.

KW - explanation protocols

KW - multi-agent systems

KW - PyXMas

KW - recommender systems

KW - Spade

KW - XAI

UR - http://www.scopus.com/inward/record.url?scp=85172214167&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-40878-6_3

DO - 10.1007/978-3-031-40878-6_3

M3 - Conference contribution

AN - SCOPUS:85172214167

SN - 9783031408779

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

SP - 38

EP - 58

BT - Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers

A2 - Calvaresi, Davide

A2 - Najjar, Amro

A2 - Omicini, Andrea

A2 - Carli, Rachele

A2 - Ciatto, Giovanni

A2 - Aydogan, Reyhan

A2 - Mualla, Yazan

A2 - Främling, Kary

PB - Springer

T2 - Proceedings of the 5th International Workshop on EXTRAAMAS 2023

Y2 - 29 May 2023 through 29 May 2023

ER -

Ciatto G, Magnini M, Buzcu B, Aydoğan R, Omicini A. A General-Purpose Protocol for Multi-agent Based Explanations. In Calvaresi D, Najjar A, Omicini A, Carli R, Ciatto G, Aydogan R, Mualla Y, Främling K, editors, Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers. Springer. 2023. p. 38-58. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-40878-6_3