The 8088 The 8088 ← All news
arXiv cs.AI AI Research Apr 27

H-Sets: Hessian-Guided Discovery of Set-Level Feature Interactions in Image Classifiers

★★★★★ significance 2/5

Researchers introduce H-Sets, a new framework for discovering and attributing higher-order feature interactions in image classifiers. The method uses input Hessians and a two-stage process to create more faithful and sparse saliency maps compared to existing attribution methods.

Why it matters Improved attribution fidelity is essential for debugging the opaque, higher-order feature interactions driving modern computer vision models.
Read the original at arXiv cs.AI

Tags

#feature attribution #interpretability #image classification #hessian-guided #deep learning

Related coverage