AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Making Sense of Dependence: Efficient Black-box Explanations Using Dependence Measure

Paul Novello, Thomas FEL, David Vigouroux

NeurIPS 2022 · Probabilistic and scientific modeling › Molecules and proteins · approach: Prediction-powered and conformal inference · 7 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
use HSIC attribution instead of RISEblack-box attribution · ImageNet val, ResNet50Deletion: 0.114 -> 0.106Improved
use HSIC attribution instead of RISEblack-box attribution · ImageNet val, ResNet50Insertion: 0.554 -> 0.470Hurt
use accurate variant (p=1536) instead of efficient (p=764)HSIC attribution · ImageNet val, ResNet50Insertion: 0.470 -> 0.481Improved
use HSIC eff. instead of RISE for execution timeblack-box attribution · ImageNet, ResNet50Exec. time (s): 8.427 -> 0.956Improved
use HSIC eff. instead of Sobol for execution timeblack-box attribution · ImageNet, ResNet50Exec. time (s): 5.254 -> 0.956Improved
use HSIC instead of D-RISE on object detectionYOLOv4 · COCO valInsertion: 0.634 -> 0.658Improved
use HSIC instead of D-RISE on object detectionYOLOv4 · COCO valmuFidelity: 0.442 -> 0.568Improved
use HSIC instead of Kernel Shap on object detectionYOLOv4 · COCO valDeletion: 0.070 -> 0.088Hurt
use HSIC instead of D-RISE for runtimeYOLOv4 · COCO valExec. time (s): 155 -> 34Improved
use HSIC instead of Kernel Shap for runtimeYOLOv4 · COCO valExec. time (s): 192 -> 34Improved
use HSIC eff. instead of Integrated Gradients (white-box) for Deletionattribution · ImageNet val, ResNet50Deletion: 0.138 -> 0.106Improved
use HSIC eff. instead of GradCAM++ (white-box) for Insertionattribution · ImageNet val, ResNet50Insertion: 0.497 -> 0.470Hurt

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