Natural language explanations of classifier behavior

Rodrigo Monteiro De Aquino, Fabio Cozman

Research output: Contribution to conferenceConference Paper

Abstract

© 2019 IEEE. Tools that enhance interpretability of classifiers tend to focus on the knowledgeable data scientist. Here we propose techniques that generate textual explanations of the internal behavior of a given classifier, aiming at less technically proficient users of machine learning resources. Our approach has been positively evaluated by a group of users who received its textual output.
Original languageAmerican English
Pages239-242
Number of pages4
DOIs
StatePublished - 1 Jun 2019
Externally publishedYes
EventProceedings - IEEE 2nd International Conference on Artificial Intelligence and Knowledge Engineering, AIKE 2019 -
Duration: 1 Jun 2019 → …

Conference

ConferenceProceedings - IEEE 2nd International Conference on Artificial Intelligence and Knowledge Engineering, AIKE 2019
Period1/06/19 → …

Fingerprint

Dive into the research topics of 'Natural language explanations of classifier behavior'. Together they form a unique fingerprint.

Cite this