English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In this paper, we propose leveraging causal generative learning as aninterpretable tool for explaining image classifiers. Specifically, we present agenerative counterfactual inference approach to study the influence of visualfeatures (i.e., pixels) as well as causal factors through generative learning.To this end, we first uncover the most influential pixels on a classifier'sdecision by varying the value of a causal attribute via counterfactualinference and computing both Shapely and contrastive explanations forcounterfactual images with these different attribute values. We then establisha Monte-Carlo mechanism using the generator of a causal generative model inorder to adapt Shapley explainers to produce feature importances for thehuman-interpretable attributes of a causal dataset in the case where aclassifier has been trained exclusively on the images of the dataset. Finally,we present optimization methods for creating counterfactual explanations ofclassifiers by means of counterfactual inference, proposing straightforwardapproaches for both differentiable and arbitrary classifiers. We exploit theMorpho-MNIST causal dataset as a case study for exploring our proposed methodsfor generating counterfacutl explantions. We employ visual explanation methodsfrom OmnixAI open source toolkit to compare them with our proposed methods. Byemploying quantitative metrics to measure the interpretability ofcounterfactual explanations, we find that our proposed methods ofcounterfactual explanation offer more interpretable explanations compared tothose generated from OmnixAI. This finding suggests that our methods arewell-suited for generating highly interpretable counterfactual explanations oncausal datasets.

Causal Generative Explainers using Counterfactual Inference: A Case Study on the Morpho-MNIST Dataset