Workshop on Explainable Logic-Based Knowledge Representation (XLoKR) 2024
Abstract
Bibtex
@inproceedings{bienvenuFLxlokr25,
author = {Bienvenu, Meghyn and Figueira, Diego and Lafourcade, Pierre},
title = {Shapley Revisited: Tractable Responsibility Measures for Query Answers (Extended Abstract)},
booktitle = {The Sixth Workshop on Explainable Logic-Based Knowledge Representation},
date = {2025-11-13},
url = {https://sites.google.com/view/xlokr2025/},
abstract = {This extended abstract provides context for and a summary of our PODS’25 paper, which revisits the question of what consti- tutes a reasonable responsibility measure for database queries. While recent work has focused on one particular measure (which we term the ‘drastic Shapley value’) defined using the Shapley value, we show that there are in fact multiple ways of applying the Shapley value to obtain a responsibility measure for Boolean monotone queries. This leads us to define a new family of Shapely-based responsibility measures – weighted sums of minimal supports (WSMS) – which satisfy intuitive properties and enjoy much better computational properties than the drastic Shapley value. These definitions and results laid key foundations for our KR’25 paper on Shapley-based responsibility measures in ontology-mediated query answering.},
}