AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Detecting Persuasive Atypicality by Modeling Contextual Compatibility

Meiqi Guo, Rebecca Hwa, Adriana Kovashka

ICCV 2021 · Detection and segmentation › Open-set and few-shot detection · approach: Visual relationship and scene graph detection · 13 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 Relative-Spatial attention instead of standard attentionunsupervised Transformer · Ads atypicality (micro avg)AUC (%): 62.86 -> 64.32Improved
use Relative-Spatial attention instead of standard attentionsupervised Transformer · Ads atypicality (micro avg)AUC (%): 65.58 -> 66.75Improved
use Relative-Spatial attention instead of standard attentionunsupervised Transformer · Ads category OMPAUC (%): 42.52 -> 55.18Hurt
use Relative-Spatial attention instead of standard attentionunsupervised Transformer · Ads category OIOAUC (%): 63.07 -> 67.31Improved
use Relative-Spatial attention instead of standard attentionunsupervised Transformer · Single-Object datasetAUC (%): 0.85 -> 0.78Hurt
increase encoder depth from 1 to 4 layers (ablation)Transformer · Ads atypicality (micro avg)AUC (%): 62.86 -> 64.14Improved
increase encoder depth from 1 to 4 layers (ablation)RS Transformer · Ads atypicality (micro avg)AUC (%): 64.32 -> 64.39No clear effect
add IoU as relative positional feature (ablation)RS Transformer (1 layer) · Ads atypicality (micro avg)AUC (%): 64.32 -> 64.99Improved
use RoI visual features instead of class labels (ablation)RS Transformer (unsup) · Ads atypicality (micro avg)AUC (%): 58.46 -> 64.32Improved
use RoI visual features instead of class labels (ablation)Transformer (unsup) · Ads atypicality (micro avg)AUC (%): 57.63 -> 62.86Improved
use supervised labels instead of masked reconstructionRS Transformer · Ads atypicality (micro avg)AUC (%): 64.32 -> 66.75No clear effect
use supervised labels instead of masked reconstructionTransformer · Single-Object datasetAUC (%): 0.85 -> 0.98Improved

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