Short Boolean Formulas as Explanations in Practice

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

1 Citation (Scopus)
17 Downloads (Pure)

Abstract

We investigate explainability via short Boolean formulas in the data model based on unary relations. As an explanation of length k, we take a Boolean formula of length k that minimizes the error with respect to the target attribute to be explained. We first provide novel quantitative bounds for the expected error in this scenario. We then also demonstrate how the setting works in practice by studying three concrete data sets. In each case, we calculate explanation formulas of different lengths using an encoding in Answer Set Programming. The most accurate formulas we obtain achieve errors similar to other methods on the same data sets. However, due to overfitting, these formulas are not necessarily ideal explanations, so we use cross validation to identify a suitable length for explanations. By limiting to shorter formulas, we obtain explanations that avoid overfitting but are still reasonably accurate and also, importantly, human interpretable.
Original languageEnglish
Title of host publicationLogics in Artificial Intelligence
EditorsSarah Gaggl, Maria Vanina Martinez, Magdalena Ortiz
PublisherSpringer
Pages90-105
Number of pages16
Volume14281
ISBN (Electronic)978-3-031-43619-2
ISBN (Print)978-3-031-43618-5
DOIs
Publication statusPublished - 24 Sept 2023
Publication typeA4 Article in conference proceedings
EventEuropean Conference on Logics in Artificial Intelligence - Dresden, Germany
Duration: 20 Sept 202322 Sept 2023

Publication series

NameLecture Notes in Computer Science
Volume14281
ISSN (Electronic)1611-3349

Conference

ConferenceEuropean Conference on Logics in Artificial Intelligence
Country/TerritoryGermany
CityDresden
Period20/09/2322/09/23

Publication forum classification

  • Publication forum level 1

Fingerprint

Dive into the research topics of 'Short Boolean Formulas as Explanations in Practice'. Together they form a unique fingerprint.

Cite this