DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación
Otros (Other)
Can Interpretability Layouts Influence Human Perception of Offensive Sentences?
6th International Workshop, EXTRAAMAS 2024, Auckland, New Zealand, May 6–10, 2024, Revised Selected Papers.- Conference proceedings., This paper conducts a user study to assess whether three machine learning (ML) interpretability layouts can influence participants’ views when evaluating sentences containing hate speech, focusing on the “Misogyny” and “Racism” classes. Given the existence of divergent conclusions in the literature, we provide statistical and qualitative analyses of questionnaire responses using the Generalized Additive Model to estimate participants’ ratings, incorporating within-subject and between-subject designs. While our statistical analysis indicates that none of the interpretability layouts significantly influences participants’ views, our qualitative analysis demonstrates the advantages of ML interpretability: 1) triggering participants to provide corrective feedback in case of discrepancies between their views and the model, and 2) providing insights to evaluate a model’s behavior beyond traditional performance metrics., This research was supported by the EU funded VALAWAI (# 101070930) and WeNet (# 823783) projects, the Spanish funded VAE (# TED2021-131295BC31) and Rhymas (# PID2020-113594RB-100) projects, and the Generalitat de Catalunya funded Ajuts a grups de recerca de Catalunya (# 2021 SGR 00754) project., Peer reviewed
Proyecto:
EC / HE 101070930
,
EC / H2020 823783
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020 PID2020-113594RB-I00
,
AEI TED2021-131295BC31
DOI: http://hdl.handle.net/10261/378119, https://api.elsevier.com/content/abstract/scopus_id/85206096954
HANDLE: http://hdl.handle.net/10261/378119, https://api.elsevier.com/content/abstract/scopus_id/85206096954