Learning Interpretability for Visualizations using Adapted Cox Models through a User Experiment

Adrien Bibal; Benoît Frénay

Learning Interpretability for Visualizations using Adapted Cox Models through a User Experiment

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

43 Téléchargements (Pure)

Résumé

In order to be useful, visualizations need to be interpretable. This paper uses a user-based approach to combine and assess quality measures in order to better model user preferences. Results show that cluster separability measures are outperformed by a neighborhood conservation measure, even though the former are usually considered as intuitively representative of user motives. Moreover, combining measures, as opposed to using a single measure, further improves prediction performances.

langue originale	Anglais
titre	NIPS Workshop on Interpretable Machine Learning in Complex Systems
Lieu de publication	Barcelona
Etat de la publication	Publié - 2016
Evénement	Thirtieth Conference on Neural Information Processing Systems - Sagrada Familia, Barcelonne, Espagne Durée: 5 déc. 2016 → 10 déc. 2016

Colloque

Colloque	Thirtieth Conference on Neural Information Processing Systems
Titre abrégé	NIPS 2016 (NeurIPS)
Pays/Territoire	Espagne
La ville	Barcelonne
période	5/12/16 → 10/12/16

Accès au document

2016-Bibal-Learning Interpretability for Visualizations using Adapted Cox Models through a User ExperimentAccepted author manuscript, 281 KB

Interpretability and Explainability in Machine Learning and their Application to Nonlinear Dimensionality Reduction
Bibal, A. (Auteur), FRENAY, B. (Promoteur), VANHOOF, W. (Président), Cleve, A. (Jury), Dumas, B. (Jury), Lee, J. A. (Jury) & Galarraga, L. (Jury), 16 nov. 2020
Student thesis: Doc types › Docteur en Sciences

Fichier

Contient cette citation

@inproceedings{30b04314ee184247bbb75eb4dc56551a,

title = "Learning Interpretability for Visualizations using Adapted Cox Models through a User Experiment",

abstract = "In order to be useful, visualizations need to be interpretable. This paper uses a user-based approach to combine and assess quality measures in order to better model user preferences. Results show that cluster separability measures are outperformed by a neighborhood conservation measure, even though the former are usually considered as intuitively representative of user motives. Moreover, combining measures, as opposed to using a single measure, further improves prediction performances.",

author = "Adrien Bibal and Beno{\^i}t Fr{\'e}nay",

year = "2016",

language = "English",

booktitle = "NIPS Workshop on Interpretable Machine Learning in Complex Systems",

note = "Thirtieth Conference on Neural Information Processing Systems, NIPS 2016 (NeurIPS) ; Conference date: 05-12-2016 Through 10-12-2016",

}

Learning Interpretability for Visualizations using Adapted Cox Models through a User Experiment. / Bibal, Adrien ; Frénay, Benoît.
NIPS Workshop on Interpretable Machine Learning in Complex Systems. Barcelona, 2016.

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 - Learning Interpretability for Visualizations using Adapted Cox Models through a User Experiment

AU - Bibal, Adrien

AU - Frénay, Benoît

PY - 2016

Y1 - 2016

N2 - In order to be useful, visualizations need to be interpretable. This paper uses a user-based approach to combine and assess quality measures in order to better model user preferences. Results show that cluster separability measures are outperformed by a neighborhood conservation measure, even though the former are usually considered as intuitively representative of user motives. Moreover, combining measures, as opposed to using a single measure, further improves prediction performances.

AB - In order to be useful, visualizations need to be interpretable. This paper uses a user-based approach to combine and assess quality measures in order to better model user preferences. Results show that cluster separability measures are outperformed by a neighborhood conservation measure, even though the former are usually considered as intuitively representative of user motives. Moreover, combining measures, as opposed to using a single measure, further improves prediction performances.

M3 - Conference contribution

BT - NIPS Workshop on Interpretable Machine Learning in Complex Systems

CY - Barcelona

T2 - Thirtieth Conference on Neural Information Processing Systems

Y2 - 5 December 2016 through 10 December 2016

ER -

Learning Interpretability for Visualizations using Adapted Cox Models through a User Experiment

Résumé

Colloque

Accès au document

Empreinte digitale

Thèses de l'étudiant

Interpretability and Explainability in Machine Learning and their Application to Nonlinear Dimensionality Reduction

Contient cette citation