Explaining Black-Box Models through Counterfactuals

P. Altmeyer; C.C.S. Liem; A. van Deursen

doi:10.21105/jcon.00130

Explaining Black-Box Models through Counterfactuals

P. Altmeyer^*, C.C.S. Liem, A. van Deursen

^*Corresponding author for this work

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

172 Downloads (Pure)

Abstract

We present CounterfactualExplanations.jl: a package for generating Counterfactual Explanations (CE) and Algorithmic Recourse (AR) for black-box models in Julia. CE explain how inputs into a model need to change to yield specific model predictions. Explanations that involve realistic and actionable changes can be used to provide AR: a set of proposed actions for individuals to change an undesirable outcome for the better. In this article, we discuss the usefulness of CE for Explainable Artificial Intelligence and demonstrate the functionality of our package. The package is straightforward to use and designed with a focus on customization and extensibility. We envision it to one day be the go-to place for explaining arbitrary predictive models in Julia through a diverse suite of counterfactual generators.

Original language	English
Title of host publication	The Proceedings of the JuliaCon Conferences (JCON)
Number of pages	10
DOIs	https://doi.org/10.21105/jcon.00130
Publication status	Published - 2023

Keywords

Julia
Explainable AI
Counterfactual Explanations
Algorithmic Recourse

Access to Document

10.21105/jcon.00130Licence: CC BY

10.21105.jcon.00130Final published version, 760 KBLicence: CC BY

Software and data underlying the publication: Explaining Black-Box Models through Counterfactuals
Altmeyer, P. (Creator), van Deursen, A. (Creator) & Liem, C. C. S. (Creator), TU Delft - 4TU.ResearchData, 15 Jul 2025
DOI: 10.4121/975D2C39-F78E-45D8-A46A-A61E441B1D53
Dataset/Software: Software

Cite this

@inproceedings{446dc87927824f899e25120e912448ae,

title = "Explaining Black-Box Models through Counterfactuals",

abstract = "We present CounterfactualExplanations.jl: a package for generating Counterfactual Explanations (CE) and Algorithmic Recourse (AR) for black-box models in Julia. CE explain how inputs into a model need to change to yield specific model predictions. Explanations that involve realistic and actionable changes can be used to provide AR: a set of proposed actions for individuals to change an undesirable outcome for the better. In this article, we discuss the usefulness of CE for Explainable Artificial Intelligence and demonstrate the functionality of our package. The package is straightforward to use and designed with a focus on customization and extensibility. We envision it to one day be the go-to place for explaining arbitrary predictive models in Julia through a diverse suite of counterfactual generators. ",

keywords = "Julia, Explainable AI, Counterfactual Explanations, Algorithmic Recourse",

author = "P. Altmeyer and C.C.S. Liem and {van Deursen}, A.",

year = "2023",

doi = "10.21105/jcon.00130",

language = "English",

booktitle = "The Proceedings of the JuliaCon Conferences (JCON)",

}

TY - GEN

T1 - Explaining Black-Box Models through Counterfactuals

AU - Altmeyer, P.

AU - Liem, C.C.S.

AU - van Deursen, A.

PY - 2023

Y1 - 2023

N2 - We present CounterfactualExplanations.jl: a package for generating Counterfactual Explanations (CE) and Algorithmic Recourse (AR) for black-box models in Julia. CE explain how inputs into a model need to change to yield specific model predictions. Explanations that involve realistic and actionable changes can be used to provide AR: a set of proposed actions for individuals to change an undesirable outcome for the better. In this article, we discuss the usefulness of CE for Explainable Artificial Intelligence and demonstrate the functionality of our package. The package is straightforward to use and designed with a focus on customization and extensibility. We envision it to one day be the go-to place for explaining arbitrary predictive models in Julia through a diverse suite of counterfactual generators.

AB - We present CounterfactualExplanations.jl: a package for generating Counterfactual Explanations (CE) and Algorithmic Recourse (AR) for black-box models in Julia. CE explain how inputs into a model need to change to yield specific model predictions. Explanations that involve realistic and actionable changes can be used to provide AR: a set of proposed actions for individuals to change an undesirable outcome for the better. In this article, we discuss the usefulness of CE for Explainable Artificial Intelligence and demonstrate the functionality of our package. The package is straightforward to use and designed with a focus on customization and extensibility. We envision it to one day be the go-to place for explaining arbitrary predictive models in Julia through a diverse suite of counterfactual generators.

KW - Julia

KW - Explainable AI

KW - Counterfactual Explanations

KW - Algorithmic Recourse

U2 - 10.21105/jcon.00130

DO - 10.21105/jcon.00130

M3 - Conference contribution

BT - The Proceedings of the JuliaCon Conferences (JCON)

ER -

Explaining Black-Box Models through Counterfactuals

Abstract

Keywords

Access to Document

Fingerprint

Datasets

Software and data underlying the publication: Explaining Black-Box Models through Counterfactuals

Cite this