TraCE: Trajectory Counterfactual Explanation Scores

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Clark, Jeffrey N., Small, Edward A., Keshtmand, Nawid, Wan, Michelle W. L., Mayoral, Elena Fillola, Werner, Enrico, Bourdeaux, Christopher P., Santos-Rodriguez, Raul
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917575044104192
author Clark, Jeffrey N.
Small, Edward A.
Keshtmand, Nawid
Wan, Michelle W. L.
Mayoral, Elena Fillola
Werner, Enrico
Bourdeaux, Christopher P.
Santos-Rodriguez, Raul
author_facet Clark, Jeffrey N.
Small, Edward A.
Keshtmand, Nawid
Wan, Michelle W. L.
Mayoral, Elena Fillola
Werner, Enrico
Bourdeaux, Christopher P.
Santos-Rodriguez, Raul
contents Counterfactual explanations, and their associated algorithmic recourse, are typically leveraged to understand, explain, and potentially alter a prediction coming from a black-box classifier. In this paper, we propose to extend the use of counterfactuals to evaluate progress in sequential decision making tasks. To this end, we introduce a model-agnostic modular framework, TraCE (Trajectory Counterfactual Explanation) scores, which is able to distill and condense progress in highly complex scenarios into a single value. We demonstrate TraCE's utility across domains by showcasing its main properties in two case studies spanning healthcare and climate change.
format Preprint
id arxiv_https___arxiv_org_abs_2309_15965
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle TraCE: Trajectory Counterfactual Explanation Scores
Clark, Jeffrey N.
Small, Edward A.
Keshtmand, Nawid
Wan, Michelle W. L.
Mayoral, Elena Fillola
Werner, Enrico
Bourdeaux, Christopher P.
Santos-Rodriguez, Raul
Machine Learning
Computers and Society
Metric Geometry
Counterfactual explanations, and their associated algorithmic recourse, are typically leveraged to understand, explain, and potentially alter a prediction coming from a black-box classifier. In this paper, we propose to extend the use of counterfactuals to evaluate progress in sequential decision making tasks. To this end, we introduce a model-agnostic modular framework, TraCE (Trajectory Counterfactual Explanation) scores, which is able to distill and condense progress in highly complex scenarios into a single value. We demonstrate TraCE's utility across domains by showcasing its main properties in two case studies spanning healthcare and climate change.
title TraCE: Trajectory Counterfactual Explanation Scores
topic Machine Learning
Computers and Society
Metric Geometry
url https://arxiv.org/abs/2309.15965