AdViCE: Aggregated Visual Counterfactual Explanations for Machine Learning Model Validation

Gómez, Óscar; Holter, Steffen; Yuan, Jun; Bertini, Enrico

doi:10.1109/vis49827.2021.9623271

Cited by 25 publications

(27 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For instance, early work Lash et al [2017] under the name of inverse classification maintained sparsity by partitioning the features into immutable and mutable features, and imposing budgetary constraints on allowable changes to the mutable features. Likewise Laugel et al Laugel et al [2018] advocated a sampling approach with their growing spheres method to traverse the input space where Gomez et al Gomez et al [2020] submitted a heuristic search method supported with visualisations to the FICO data challenge FICO [2018]. Their considerations of data distribution and feature importance, along with counterfactual explanations are in line with our current work.…”

Section: Background and Previous Worksupporting

confidence: 73%

Counterfactual Explanations via Latent Space Projection and Interpolation

Barr¹,

R.²,

Sharpe³

et al. 2021

Preprint

View full text Add to dashboard Cite

Counterfactual explanations represent the minimal change to a data sample that alters its predicted classification, typically from an unfavorable initial class to a desired target class. Counterfactuals help answer questions such as "what needs to change for this application to get accepted for a loan?". A number of recently proposed approaches to counterfactual generation give varying definitions of "plausible" counterfactuals and methods to generate them. However, many of these methods are computationally intensive and provide unconvincing explanations. Here we introduce SharpShooter, a method for binary classification that starts by creating a projected version of the input that classifies as the target class. Counterfactual candidates are then generated in latent space on the interpolation line between the input and its projection. We then demonstrate that our framework translates core characteristics of a sample to its counterfactual through the use of learned representations. Furthermore, we show that SharpShooter is competitive across common quality metrics on tabular and image datasets while being three orders of magnitude faster than two comparable methods and excels at measures of realism, making it well-suited for high velocity machine learning applications which require timely explanations. * Equal contribution.Preprint. Under review.

show abstract

Section: Background and Previous Worksupporting

confidence: 73%

Counterfactual Explanations via Latent Space Projection and Interpolation

Barr¹,

R.²,

Sharpe³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Bayesian Rule List [4]) and an instance-based explanation uses an instance to explain the behaviour of a black-box model (e.g. Visual Counterfactual Explanations (ViCE) [5]). Rudin [6] posited that inherently interpretable models (i.e.…”

Section: Introductionmentioning

confidence: 99%

Model-Agnostic Counterfactual Explanations in Credit Scoring

2022

View full text Add to dashboard Cite

The past decade has shown a surge in the use and application of machine learning and deep learning models across different domains. One such domain is the credit scoring domain, where applicants are scored to assess their credit-worthiness for loan applications. During the scoring process, it is key to assure that there are no biases or discriminations that are incurred. Despite the proliferation of machine learning and deep learning models (referred to as black-box models in the literature) in credit scoring, there is still a need to explain how each prediction is made by the black-box models. Most of the machine learning and deep learning models are likely to be prone to unintended bias and discrimination that may occur in the datasets. To avoid the element of model bias and discrimination, it is imperative to explain each prediction during the scoring process. Our study proposes a novel optimisation formulation that generates a sparse counterfactual via a custom genetic algorithm to explain a black-box model's prediction. This study uses publicly available credit scoring datasets. Furthermore, we validated the generated counterfactual explanations by comparing them to the counterfactual explanations from credit scoring experts. The proposed explanation technique does not only explains rejected applications, it can also be used to explain approved loans.

show abstract

“…Figura 2.1 Interface do sistema ViCE (Gomez et al, 2020) Figura 2.2 Interface do sistema VIDA Lab (Piesanen and Kerrigan, 2020) Figura 2.3 Interface de preenchimento de dados do sistema VIDA Lab (Piesanen and Kerrigan, 2020) Figura 2.4 Interface do sistema What If (Code, 2021) Figura 2.5 Interface do sistema AdViCE (Gomez et al, 2021) Figura 3.1 Componente usado para features binárias Figura 3.2 Componente usado para features categóricas Figura 3.3 Componente usado para features numéricas Figura 3.4 Dataset COMPAS Figura 3.5 Interface de geração instantânea com um dataset selecionado Figura 3.6 Interface de geração instantânea com resultado Figura 3.7 Interface de geração iterativa primeira versão em estado inicial Figura 3.8 Interface de geração iterativa primeira versão ponto selecionado Figura 3.9 Interface de geração iterativa primeira versão cenário 1 Figura 3.10 Interface de geração iterativa primeira versão gráfico do cenário Figura 3.11 Interface de geração iterativa primeira versão gráfico clicado Figura 3.12 Interface de geração iterativa primeira versão cenário final Figura 3.13 OceanUI em estado inicial Figura 3.14 OceanUI dataset selecionado Figura 3.15 OceanUI ponto de entrada preenchido e classe calculada Figura 3.16 OceanUI erro por excesso de restrições Figura 3.17 OceanUI novo cenário instanciado Figura 3.18 OceanUI cenário editado Figura 3.19 OceanUI restrição contraditória ao valor editado Figura 3.20 OceanUI cenário seguinte instanciado com contrafactual encontrado Figura 3.21 OceanUI cenário final para comparação…”

Section: Lista De Figurasunclassified

“…Ao final do presente trabalho nos deparamos com a continuidade de ViCE (Gomez et al, 2020) que é o AdViCE (Gomez et al, 2021) (Figura 2.5), neste projeto observamos a abordagem utilizada para a manipulação de features numéricas e pudemos ver que a forma que desenvolvemos o nosso componente é promissora uma vez que é similar à feita por estes autores, além de embasar a necessidade e importância de imposição de restrições para personalização das explicações contrafactuais. E similar ao ViCE o AdViCE também permite Figura 2.4: Interface do sistema What If (Code, 2021) a geração de contrafactuais e as exibe de forma gráfica, além de permitir aos usuários impor algumas restrições e filtros sobre o modelo.…”

Section: Interfaces Para Contrafactuaisunclassified

See 1 more Smart Citation

Oceanui: Interface Para Geração De Explicações Contrafactuais

PEREIRA¹

View full text Add to dashboard Cite

show abstract

AdViCE: Aggregated Visual Counterfactual Explanations for Machine Learning Model Validation

Cited by 25 publications

References 17 publications

Counterfactual Explanations via Latent Space Projection and Interpolation

Counterfactual Explanations via Latent Space Projection and Interpolation

Model-Agnostic Counterfactual Explanations in Credit Scoring

Oceanui: Interface Para Geração De Explicações Contrafactuais

Contact Info

Product

Resources

About