AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

On the Complexity-Faithfulness Trade-off of Gradient-Based Explanations

Amir Mehrpanah, Matteo Gamba, Kevin Smith, Hossein Azizpour

ICCV 2025 · Learning theory and algorithms › Prediction and decision losses · approach: Provable post-hoc explanations · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace ReLU with smooth parameterization SP(beta=.9) in VanillaGrad (ablation)CNN · Imagenette (CNN)EF: .390 -> .202Improved
apply SmoothGrad post-hoc smoothing (ablation)ReLU CNN · Imagenette (CNN)EF: .390 -> .286Improved
apply SmoothGrad post-hoc smoothing (cost in explanation gap)ReLU CNN · Imagenette (CNN)dEF: 0 -> .104Hurt
apply GradCAM instead of VanillaGradReLU CNN · Imagenette (CNN)EF: .390 -> .293Improved
apply SmoothGrad on SP(beta=.9) network instead of ReLU (ablation)CNN · Imagenette (CNN)dEF: .104 -> .005Improved
apply GradCAM instead of VanillaGradResNet50 · ImageNet (pretrained)dEF: 0 -> .130Hurt
use ViT-B16 instead of ResNet50 for GradCAMGradCAM · ImageNet (pretrained)dEF: .130 -> .041Improved
use GELU instead of ReLU in ViT-B16 (ablation)VanillaGrad · Imagenette (ViT-B16 from scratch)EF: .239 -> .248Hurt
use GELU instead of ReLU in ViT-B16 (ablation)GradCAM · Imagenette (ViT-B16 from scratch)dEF: .051 -> .033Improved
increase smoothness-parameter beta toward ReLU (ablation)CNN · Fashion-MNIST, CIFAR10, ImagenetteEFHurt
add batch normalization (ablation)CNN · Imagenette 224x224spatial power spectrum tailHurt
add skip connections (ablation)CNN · Imagenette 224x224spatial power spectrum tailNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy