← Search

Eslam Zaher

2 accepted papers

2026

Counterfactual Explanations on Robust Perceptual Geodesics

ICLR 2026poster

Latent-space optimization methods for counterfactual explanations—framed as minimal semantic perturbations that change model predictions—inherit the ambiguity of Wachter et al.’s objective: the choice of distance metric dictates whether perturbations are meaningful or adversarial. Existing approache…

Cited by 0SourceScholar
2024

Manifold Integrated Gradients: Riemannian Geometry for Feature Attribution

ICML 2024poster

In this paper, we dive into the reliability concerns of Integrated Gradients (IG), a prevalent feature attribution method for black-box deep learning models. We particularly address two predominant challenges associated with IG: the generation of noisy feature visualizations for vision models and th…