The Candide model: How narratives emerge where observations meet beliefs

Paul Van Eecke; Lara Verheyen; Tom Willaert; Katrien Beuls

doi:10.18653/v1/2023.wnu-1.7

The Candide model: How narratives emerge where observations meet beliefs

Paul Van Eecke, Lara Verheyen, Tom Willaert, Katrien Beuls

Résultats de recherche: Contribution dans un livre/un catalogue/un rapport/dans les actes d'une conférence › Article dans les actes d'une conférence/un colloque

93 Téléchargements (Pure)

Résumé

This paper presents the Candide model as a computational architecture for modelling human-like, narrative-based language understanding. The model starts from the idea that narratives emerge through the process of interpreting novel linguistic observations, such as utterances, paragraphs and texts, with respect to previously acquired knowledge and beliefs. Narratives are personal, as they are rooted in past experiences, and constitute perspectives on the world that might motivate different interpretations of the same observations. Concretely, the Candide model operationalises this idea by dynamically modelling the belief systems and background knowledge of individual agents, updating these as new linguistic observations come in, and exposing them to a logic reasoning engine that reveals the possible sources of divergent interpretations. Apart from introducing the foundational ideas, we also present a proof-of-concept implementation that demonstrates the approach through a number of illustrative examples.

langue originale	Anglais
titre	5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop
rédacteurs en chef	Nader Akoury, Elizabeth Clark, Mohit Iyyer, Snigdha Chaturvedi, Faeze Brahman, Khyathi Raghavi Chandu
Editeur	Association for Computational Linguistics
Pages	48-57
Nombre de pages	10
ISBN (Electronique)	9781959429920
Les DOIs	https://doi.org/10.18653/v1/2023.wnu-1.7
Etat de la publication	Publié - 2023
Evénement	The 61st Annual Meeting of the Association for Computational Linguistics - Toronto, Canada Durée: 9 juil. 2023 → 14 juil. 2023 Numéro de conférence: 61 https://2023.aclweb.org

Série de publications

Nom	Proceedings of the The 5th Workshop on Narrative Understanding

Une conférence

Une conférence	The 61st Annual Meeting of the Association for Computational Linguistics
Titre abrégé	ACL 2023
Pays/Territoire	Canada
La ville	Toronto
période	9/07/23 → 14/07/23
Adresse Internet	https://2023.aclweb.org

Financement

We are especially grateful to Remi van Trijp for his important contributions to the insightful discussions that led up to the development of the Candide model, as well as for his constructive feedback on a first version of this manuscript. The research reported on in this paper received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreements no. 951846 (MUHAI - Meaning and Understanding in Human-centric AI) and no. 101094752 (SoMe4Dem - Social Media for Democracy – understanding the causal mechanisms of digital citizenship), from the Research Foundation Flanders (FWO) through a postdoctoral grant awarded to Paul Van Eecke (grant no. 75929) and from the Flemish Government under the ‘Flanders AI Research Program’. The research reported on in this paper received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreements no. 951846 (MUHAI - Meaning and Understanding in Human-centric AI) and no. 101094752 (SoMe4Dem - Social Media for Democracy – understanding the causal mechanisms of digital citizenship), from the Research Foundation Flanders (FWO) through a postdoctoral grant awarded to Paul Van Eecke (grant no. 75929) and from the Flemish Government under the ‘Flanders AI Research Program’.

Bailleurs de fonds	Numéro du bailleur de fonds
Flanders AI Research Program
Vlaamse regering
European commission
Social Media for Democracy
Horizon 2020 Framework Programme	951846, 101094752
Fonds Wetenschappelijk Onderzoek - Vlaanderen (FWO)	75929

Accès au document

10.18653/v1/2023.wnu-1.7

2023.wnu-1.7Version finale publiée, 251 KBLicense: CC BY

https://aclanthology.org/2023.wnu-1.7/License: CC BY

Autres fichiers et liens

Lien vers la publication sur Scopus

Meaning and understanding in human-centric AI
Beuls, K. (Responsable du Projet)
1/10/20 → 30/09/24
Projet: Recherche

Contient cette citation

Van Eecke, P., Verheyen, L., Willaert, T., & Beuls, K. (2023). The Candide model: How narratives emerge where observations meet beliefs. Dans N. Akoury, E. Clark, M. Iyyer, S. Chaturvedi, F. Brahman, & K. R. Chandu (eds.), 5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop (p. 48-57). (Proceedings of the The 5th Workshop on Narrative Understanding). Association for Computational Linguistics. https://doi.org/10.18653/v1/2023.wnu-1.7

Van Eecke, Paul ; Verheyen, Lara ; Willaert, Tom et al. / The Candide model: How narratives emerge where observations meet beliefs. 5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop. Editeur / Nader Akoury ; Elizabeth Clark ; Mohit Iyyer ; Snigdha Chaturvedi ; Faeze Brahman ; Khyathi Raghavi Chandu. Association for Computational Linguistics, 2023. p. 48-57 (Proceedings of the The 5th Workshop on Narrative Understanding).

@inproceedings{9cbc6213e4524fa4b321a7aec9a7a65a,

title = "The Candide model: How narratives emerge where observations meet beliefs",

abstract = "This paper presents the Candide model as a computational architecture for modelling human-like, narrative-based language understanding. The model starts from the idea that narratives emerge through the process of interpreting novel linguistic observations, such as utterances, paragraphs and texts, with respect to previously acquired knowledge and beliefs. Narratives are personal, as they are rooted in past experiences, and constitute perspectives on the world that might motivate different interpretations of the same observations. Concretely, the Candide model operationalises this idea by dynamically modelling the belief systems and background knowledge of individual agents, updating these as new linguistic observations come in, and exposing them to a logic reasoning engine that reveals the possible sources of divergent interpretations. Apart from introducing the foundational ideas, we also present a proof-of-concept implementation that demonstrates the approach through a number of illustrative examples.",

keywords = "language understanding, reasoning, personal dynamic memory",

author = "\{Van Eecke\}, Paul and Lara Verheyen and Tom Willaert and Katrien Beuls",

note = "Publisher Copyright: {\textcopyright} 2023 Association for Computational Linguistics.; The 61st Annual Meeting of the Association for Computational Linguistics, ACL 2023 ; Conference date: 09-07-2023 Through 14-07-2023",

year = "2023",

doi = "10.18653/v1/2023.wnu-1.7",

language = "English",

series = "Proceedings of the The 5th Workshop on Narrative Understanding",

publisher = "Association for Computational Linguistics",

pages = "48--57",

editor = "Nader Akoury and Elizabeth Clark and Mohit Iyyer and Snigdha Chaturvedi and Faeze Brahman and Chandu, \{Khyathi Raghavi\}",

booktitle = "5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop",

url = "https://2023.aclweb.org",

}

Van Eecke, P, Verheyen, L, Willaert, T & Beuls, K 2023, The Candide model: How narratives emerge where observations meet beliefs. Dans N Akoury, E Clark, M Iyyer, S Chaturvedi, F Brahman & KR Chandu (eds), 5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop. Proceedings of the The 5th Workshop on Narrative Understanding, Association for Computational Linguistics, p. 48-57, The 61st Annual Meeting of the Association for Computational Linguistics, Toronto, Canada, 9/07/23. https://doi.org/10.18653/v1/2023.wnu-1.7

The Candide model: How narratives emerge where observations meet beliefs. / Van Eecke, Paul; Verheyen, Lara; Willaert, Tom et al.
5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop. Ed. / Nader Akoury; Elizabeth Clark; Mohit Iyyer; Snigdha Chaturvedi; Faeze Brahman; Khyathi Raghavi Chandu. Association for Computational Linguistics, 2023. p. 48-57 (Proceedings of the The 5th Workshop on Narrative Understanding).

Résultats de recherche: Contribution dans un livre/un catalogue/un rapport/dans les actes d'une conférence › Article dans les actes d'une conférence/un colloque

TY - GEN

T1 - The Candide model: How narratives emerge where observations meet beliefs

AU - Van Eecke, Paul

AU - Verheyen, Lara

AU - Willaert, Tom

AU - Beuls, Katrien

N1 - Conference code: 61

PY - 2023

Y1 - 2023

N2 - This paper presents the Candide model as a computational architecture for modelling human-like, narrative-based language understanding. The model starts from the idea that narratives emerge through the process of interpreting novel linguistic observations, such as utterances, paragraphs and texts, with respect to previously acquired knowledge and beliefs. Narratives are personal, as they are rooted in past experiences, and constitute perspectives on the world that might motivate different interpretations of the same observations. Concretely, the Candide model operationalises this idea by dynamically modelling the belief systems and background knowledge of individual agents, updating these as new linguistic observations come in, and exposing them to a logic reasoning engine that reveals the possible sources of divergent interpretations. Apart from introducing the foundational ideas, we also present a proof-of-concept implementation that demonstrates the approach through a number of illustrative examples.

AB - This paper presents the Candide model as a computational architecture for modelling human-like, narrative-based language understanding. The model starts from the idea that narratives emerge through the process of interpreting novel linguistic observations, such as utterances, paragraphs and texts, with respect to previously acquired knowledge and beliefs. Narratives are personal, as they are rooted in past experiences, and constitute perspectives on the world that might motivate different interpretations of the same observations. Concretely, the Candide model operationalises this idea by dynamically modelling the belief systems and background knowledge of individual agents, updating these as new linguistic observations come in, and exposing them to a logic reasoning engine that reveals the possible sources of divergent interpretations. Apart from introducing the foundational ideas, we also present a proof-of-concept implementation that demonstrates the approach through a number of illustrative examples.

KW - language understanding

KW - reasoning

KW - personal dynamic memory

UR - https://www.scopus.com/pages/publications/85175657027

U2 - 10.18653/v1/2023.wnu-1.7

DO - 10.18653/v1/2023.wnu-1.7

M3 - Conference contribution

T3 - Proceedings of the The 5th Workshop on Narrative Understanding

SP - 48

EP - 57

BT - 5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop

A2 - Akoury, Nader

A2 - Clark, Elizabeth

A2 - Iyyer, Mohit

A2 - Chaturvedi, Snigdha

A2 - Brahman, Faeze

A2 - Chandu, Khyathi Raghavi

PB - Association for Computational Linguistics

T2 - The 61st Annual Meeting of the Association for Computational Linguistics

Y2 - 9 July 2023 through 14 July 2023

ER -

Van Eecke P, Verheyen L, Willaert T, Beuls K. The Candide model: How narratives emerge where observations meet beliefs. Dans Akoury N, Clark E, Iyyer M, Chaturvedi S, Brahman F, Chandu KR, rédacteurs en chef, 5th Workshop on Narrative Understanding, WNU 2023 - Proceedings of the Workshop. Association for Computational Linguistics. 2023. p. 48-57. (Proceedings of the The 5th Workshop on Narrative Understanding). doi: 10.18653/v1/2023.wnu-1.7

The Candide model: How narratives emerge where observations meet beliefs

Résumé

Série de publications

Une conférence

Financement

Accès au document

Autres fichiers et liens

Empreinte digitale

Projets

Meaning and understanding in human-centric AI

Contient cette citation