Short Boolean Formulas as Explanations in Practice
Jaakkola, Reijo; Janhunen, Tomi; Kuusisto, Antti; Feyzbakhsh Rankooh, Masood; Vilander, Miikka (2023-09-24)
Jaakkola, Reijo
Janhunen, Tomi
Kuusisto, Antti
Feyzbakhsh Rankooh, Masood
Vilander, Miikka
Teoksen toimittaja(t)
Gaggl, Sarah
Martinez, Maria Vanina
Ortiz, Magdalena
Springer
24.09.2023
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202310058655
https://urn.fi/URN:NBN:fi:tuni-202310058655
Kuvaus
Peer reviewed
Tiivistelmä
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.
Kokoelmat
- TUNICRIS-julkaisut [19214]