AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

A Dataset Generation Framework for Evaluating Megapixel Image Classifiers & Their Explanations

Gautam Machiraju, Sylvia Plevritis, Parag Mallick

ECCV 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Self-explaining and visual explanation networks

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 attention-based VGG-Att instead of VGG-19patch CNN · MISO-1 MD PPMsensitivity: 0.500 -> 0.721Improved
use VGG-Att instead of VGG-19patch CNN · MISO-1 DSP PPMsensitivity: 0.500 -> 0.999Improved
use VGG-Att instead of VGG-19patch CNN · MISO-1 EVSP PPMF-beta measure: 0.321 -> 0.819Improved
use VGG-Att instead of VGG-19patch CNN · MISO-1 FM PPMF-beta measure: 0.429 -> 0.924Improved
use VGG-Att instead of VGG-19patch CNN · MISO-1 EVP PPMsensitivity: 0.999 -> 0.999No clear effect
use VGG-Att saliency instead of VGG-19 saliencyexplanation maps · MISO-1 DSPF-beta measure: 0.000 -> 0.585Improved
use VGG-Att instead of VGG-19explanation maps · MISO-1 EVSPF-beta measure: 0.000 -> 0.691Improved
use VGG-Att instead of VGG-19explanation maps · MISO-1 GSPF-beta measure: 0.000 -> 0.350Improved
use VGG-Att instead of VGG-19explanation maps · MISO-2ScagDist: 1.000 -> 0.398Improved
use VGG-Att instead of VGG-19patch prediction maps · MISO-2sensitivity: 0.557 -> 0.400Hurt
use VGG-Att instead of VGG-19explanation maps · MISO-1 EVPF-beta measure: 0.475 -> 0.609Improved

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