An Analysis of LIME for Text Data

Dina Mardaoui, Damien Garreau
Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, PMLR 130:3493-3501, 2021.

Abstract

Text data are increasingly handled in an automated fashion by machine learning algorithms. But the models handling these data are not always well-understood due to their complexity and are more and more often referred to as “black-boxes.” Interpretability methods aim to explain how these models operate. Among them, LIME has become one of the most popular in recent years. However, it comes without theoretical guarantees: even for simple models, we are not sure that LIME behaves accurately. In this paper, we provide a first theoretical analysis of LIME for text data. As a consequence of our theoretical findings, we show that LIME indeed provides meaningful explanations for simple models, namely decision trees and linear models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v130-mardaoui21a, title = { An Analysis of LIME for Text Data }, author = {Mardaoui, Dina and Garreau, Damien}, booktitle = {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics}, pages = {3493--3501}, year = {2021}, editor = {Banerjee, Arindam and Fukumizu, Kenji}, volume = {130}, series = {Proceedings of Machine Learning Research}, month = {13--15 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v130/mardaoui21a/mardaoui21a.pdf}, url = {https://proceedings.mlr.press/v130/mardaoui21a.html}, abstract = { Text data are increasingly handled in an automated fashion by machine learning algorithms. But the models handling these data are not always well-understood due to their complexity and are more and more often referred to as “black-boxes.” Interpretability methods aim to explain how these models operate. Among them, LIME has become one of the most popular in recent years. However, it comes without theoretical guarantees: even for simple models, we are not sure that LIME behaves accurately. In this paper, we provide a first theoretical analysis of LIME for text data. As a consequence of our theoretical findings, we show that LIME indeed provides meaningful explanations for simple models, namely decision trees and linear models. } }
Endnote
%0 Conference Paper %T An Analysis of LIME for Text Data %A Dina Mardaoui %A Damien Garreau %B Proceedings of The 24th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2021 %E Arindam Banerjee %E Kenji Fukumizu %F pmlr-v130-mardaoui21a %I PMLR %P 3493--3501 %U https://proceedings.mlr.press/v130/mardaoui21a.html %V 130 %X Text data are increasingly handled in an automated fashion by machine learning algorithms. But the models handling these data are not always well-understood due to their complexity and are more and more often referred to as “black-boxes.” Interpretability methods aim to explain how these models operate. Among them, LIME has become one of the most popular in recent years. However, it comes without theoretical guarantees: even for simple models, we are not sure that LIME behaves accurately. In this paper, we provide a first theoretical analysis of LIME for text data. As a consequence of our theoretical findings, we show that LIME indeed provides meaningful explanations for simple models, namely decision trees and linear models.
APA
Mardaoui, D. & Garreau, D.. (2021). An Analysis of LIME for Text Data . Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 130:3493-3501 Available from https://proceedings.mlr.press/v130/mardaoui21a.html.

Related Material