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

Model interpretability plays a central role in human-AI decision-makingsystems. Ideally, explanations should be expressed using human-interpretablesemantic concepts. Moreover, the causal relations between these concepts shouldbe captured by the explainer to allow for reasoning about the explanations.Lastly, explanation methods should be efficient and not compromise theperformance of the predictive task. Despite the rapid advances in AIexplainability in recent years, as far as we know to date, no method fulfillsthese three properties. Indeed, mainstream methods for local conceptexplainability do not produce causal explanations and incur a trade-off betweenexplainability and prediction performance. We present DiConStruct, anexplanation method that is both concept-based and causal, with the goal ofcreating more interpretable local explanations in the form of structural causalmodels and concept attributions. Our explainer works as a distillation model toany black-box machine learning model by approximating its predictions whileproducing the respective explanations. Because of this, DiConStruct generatesexplanations efficiently while not impacting the black-box prediction task. Wevalidate our method on an image dataset and a tabular dataset, showing thatDiConStruct approximates the black-box models with higher fidelity than otherconcept explainability baselines, while providing explanations that include thecausal relations between the concepts.

DiConStruct: Causal Concept-based Explanations through Black-Box Distillation