Timezone: »

Feasible and Desirable Counterfactual Generation by Preserving Human Defined Constraints
Homayun Afrabandpey · Michael Spranger

We present a human-in-the-loop approach to generate counterfactual (CF) explanations that preserve global and local feasibility constraints. Global feasibility constraints refer to the causal constraints necessary for generating actionable CF explanation. Assuming a domain expert with knowledge on unary and binary causal constraints, our approach efficiently employs this knowledge to generate CF explanation by rejecting gradient steps that violate these constraints. Local feasibility constraints are user-level constraints necessary for generating desirable CF explanation. We extract these constraints from the end-user of the model and exploit them during CF generation via user-defined distance metric. Through user studies, we demonstrate that incorporating causal constraints during CF generation results in significantly better explanations in terms of feasibility and desirability for participants. Adopting local and global feasibility constraints simultaneously, although improves user satisfaction, does not significantly improves desirability of the participants compared to only incorporating global constraints.

Author Information

Homayun Afrabandpey (Nokia Technologies)

I’m a senior scientist at Nokia Technologies working on deep neural network compression and explainable AI. I completed my PhD in 2019. My thesis research lied in the space of human-in-the-loop machine learning with focus on knowledge elicitation, probabilistic modeling, and active learning.

Michael Spranger (Sony)

More from the Same Authors