AI RESEARCH
Manifold-Aligned Guided Integrated Gradients for Reliable Feature Attribution
arXiv CS.LG
•
ArXi:2605.02167v1 Announce Type: new Feature attribution is central to diagnosing and trusting deep neural networks, and Integrated Gradients (IG) is widely used due to its axiomatic properties. However, IG can yield unreliable explanations when the integration path between a baseline and the input passes through regions with noisy gradients. While Guided Integrated Gradients reduces this sensitivity by adaptively updating low-gradient-magnitude features, input-space guidance still produces intermediate inputs that deviate from the data manifold.