Counterfactual Explanations for Machine Learning: Challenges Revisited
This work addresses challenges for practitioners in deploying interpretability techniques, but it is incremental as it revisits existing issues without new solutions.
The paper tackles the problem of slow adoption of counterfactual explanations in industry by identifying obstacles hindering their deployment, based on desirable properties and practical experience.
Counterfactual explanations (CFEs) are an emerging technique under the umbrella of interpretability of machine learning (ML) models. They provide ``what if'' feedback of the form ``if an input datapoint were $x'$ instead of $x$, then an ML model's output would be $y'$ instead of $y$.'' Counterfactual explainability for ML models has yet to see widespread adoption in industry. In this short paper, we posit reasons for this slow uptake. Leveraging recent work outlining desirable properties of CFEs and our experience running the ML wing of a model monitoring startup, we identify outstanding obstacles hindering CFE deployment in industry.