Title The metric-aware kernel-width choice for LIME
Authors Barrera-Vicen A. , PALUZO HIDALGO, EDUARDO, Gutiérrez-Naranjo M.A.
External publication No
Means CEUR Workshop Proc.
Scope Conference Paper
Nature Científica
SJR Impact 0.191
Web https://www.scopus.com/inward/record.uri?eid=2-s2.0-85178587954&partnerID=40&md5=8f3f390d5ef84eb8f523b1325acdbd39
Publication date 01/01/2023
Scopus Id 2-s2.0-85178587954
Abstract Local Interpretable Model-Agnostic Explanations (LIME) are a well-known approach to provide local interpretability to Machine Learning models. LIME uses an exponential smoothing kernel based on the kernel width value, which defines the width of the local neighbourhood. In this paper, we study the influence of the distances for these local explanations, and we explore the choice of kernel width to guarantee a fair performance comparison between the distances. © 2023 CEUR-WS. All rights reserved.
Keywords Explainability; Exponential smoothing; Interpretability; Kernel width; Local interpretable model-agnostic explanation; Local neighborhoods; Machine learning models; Performance comparison; Smoothing kernels; XAI; Lime
Universidad Loyola members